The target segment on the campaign brief is SMB New Logos. SMB is short for small and medium-sized business, and new logos are companies that haven’t bought from you before. SMB New Logos looks like a segment, but it only tells you that the companies are small or medium-sized and that they aren’t customers yet. It doesn’t tell you what the buyers in those companies want.

You then have to write one message for every small and medium-sized company that isn’t a customer yet. A 20-person accounting practice and a 300-person freight company both qualify. They buy for different reasons, through different people, on different budgets. A message written for both is too general to persuade either, so part of the media budget is spent on buyers the message doesn’t fit. You also can’t say how many of those companies want the product you’re advertising.

The business stakeholder had a sensible reason to write SMB New Logos. Sales territories are divided by company size, and the revenue plan separates new customers from existing customers, so the business already reports its revenue under names like SMB New Logos.

A segment is a group of buyers who want the same problem fixed. Dividing a market into segments is called market segmentation, and nobody did it before the business stakeholder wrote this campaign brief. With a market segmentation, you know which segments are in the market, what problem each segment’s buyers want fixed and how many companies are in each segment. You can draft a segmentation with 12 interviews and your sales team’s last 50 deals, and you can run a survey later to count the companies in each segment. You write the result of each step on one worksheet, which is attached at the end of this guide.

1. Define the whole market before you divide it

Write a market definition: one sentence that says which companies could buy this kind of product. Kotler and Armstrong define market segmentation as dividing a market into smaller groups of buyers with distinct needs, characteristics or behaviours who might require separate products or marketing mixes. The market you divide is every company that could buy: your customers, your competitors’ customers and the companies that haven’t bought from anyone.

To help illustrate, take a software company that sells scheduling software to companies that send employees out to customer sites: trades, equipment servicing, home care and facilities maintenance. The campaign brief from the business stakeholder says SMB New Logos, 10 to 500 employees, Australia and New Zealand. The scheduling company has 900 customers, and most of them are plumbing and electrical contractors.

The marketing team defines the market as every company in Australia and New Zealand that sends employees out to customer sites. The size limit of 10 to 500 employees from the campaign brief is left out, because a company with 5 employees or 800 employees could also buy scheduling software.

Next, work out the market size, which here means the number of companies that match your market definition. Use free public data. The Australian Bureau of Statistics (ABS) publishes counts of businesses by industry, and Stats NZ publishes the same for New Zealand. List the industries whose companies match your market definition, then add up the number of businesses in each industry. Where only some companies in an industry match, estimate the share. For the scheduling company, the market size is about 40,500 companies.

2. Interview customers, competitors’ customers and companies that bought nothing

Start with 12 interviews: 4 with customers, 4 with competitors’ customers and 4 with companies that bought nothing. The number 12 comes from research on interviewing. In a 2006 study of 60 in-depth interviews, almost all the topics the researchers recorded had come up by interview 12. Most of the common topics had come up by interview 6. The authors, Greg Guest, Arwen Bunce and Laura Johnson, interviewed a group of similar people for a health study. They warn that 12 probably won’t be enough when the people interviewed differ from each other.

Your buyers do differ from each other. Some chose you, some chose a competitor and some bought nothing. Use the first 12 interviews to draft the segmentation, then interview about 12 of each kind of buyer to finish it. If you sell several products or sell in several countries, your buyers vary more, so you need more interviews. In this example, the marketing team starts with 12 interviews and later completes 36 in total.

To find competitors’ customers and companies that bought nothing, ask the sales team for a list of the deals lost in the last 12 months. The list shows which companies chose a competitor and which companies looked at your product and bought nothing. For companies that have never contacted you, use the member list of a trade association.

Interview the person who makes the buying decision. The scheduling company interviews owners and operations managers. Tell the buyer that the interview takes 20 minutes and that you aren’t selling anything. Ask someone outside the sales team to run the interview, so that the buyer doesn’t expect a sales pitch. Some companies will say no, so contact more companies than you need.

3. Ask every buyer the same questions

Ask about the task your product does and how the buyer decided, and leave your product out of the questions. Asking every buyer the same questions lets you compare the answers. The scheduling company asks about scheduling. In your own questions, replace the word scheduling with the task your buyers use your product for, such as payroll or invoicing.

• How do you handle scheduling today, and what goes wrong?

• What made you look for something better, or what would make you look?

• What happens if you change nothing?

• What are you measured on in your job?

• Who else has a say in a decision like this?

• What else did you consider, or would you consider, instead of buying software?

Write one page of notes for each interview, in the buyer’s own words. Record the interview if the buyer agrees.

An operations manager at an electrical contractor tells the scheduling company’s interviewer that the contractor missed 3 appointments in one month and lost a contract. Instead of buying software, the operations manager would have hired a second dispatcher. The owner of another electrical contractor of the same size describes a different problem: jobs finished on a Tuesday were invoiced 3 weeks later. Instead of buying software, the owner would have hired a part-time bookkeeper.

4. Sort buyers by the problem they wanted fixed

The 2 electrical contractors are in the same industry and are the same size, and their buyers wanted different problems fixed. A 200-person home care provider is in a different industry and is a different size. Its buyer described the same problem as the operations manager at the electrical contractor: missed appointments. Sort by the problem that made each buyer start looking, because industry and company size would put these companies in the wrong groups.

When a buyer names 2 problems, use the problem the buyer gave as the reason for starting to look. When 2 or more buyers name the same problem, you have a draft segment.

Then test your draft segments against Kotler and Armstrong’s definition, which says the buyers in a segment might require separate products or marketing mixes. If 2 draft segments would buy the same product, for the same reason, in response to the same message, merge them into one segment.

Name each segment after what its buyers want, using the buyers’ words, and give it a row on the worksheet. In that row, write the problem, what the buyer would do instead of buying, and the job title of the person who makes the buying decision. After 12 interviews, the scheduling company has these draft segments:

• Stop losing jobs. A missed or late appointment costs the company a job and brings a complaint.

• Get paid faster. Invoices go out weeks after the job is done.

• Pass the audit. The company’s own customers demand proof of every visit.

• Run a national fleet. Each region schedules in its own way, and head office staff can’t see every region’s jobs in one place.

Some companies have no problem to fix, because phone calls and a whiteboard still work for them. Put those companies in a last row called Not buying, separate from the segments, so that the companies on the worksheet add up to the market size. A company that looked at products and bought nothing still had a problem, so it belongs in a segment.

Small and medium-sized companies that haven’t bought from you are in every row. The business stakeholder wrote them down as one segment.

5. Work out how many companies are in each segment

A survey is the only way to count the companies in each segment. Until you have a survey budget, rank the segments by number of deals, which you can do this week with data you already have. Go through the last 50 deals, won and lost, with the sales team, and put each deal in a segment. Salespeople often don’t record in the customer relationship management (CRM) system why the company was looking, so ask the salesperson who handled each deal. Then count the deals in each segment and rank the segments from most deals to fewest.

The scheduling company counts 26 deals in Stop losing jobs, 18 in Get paid faster, 4 in Run a national fleet and 2 in Pass the audit. No deals are in Not buying, because a company with no problem to fix doesn’t contact a sales team. Leave Not buying unranked.

Your sales team mostly deals with buyers who already found your company, so you can’t use the ranking to count the companies in each segment. Show the ranking to the business stakeholder as an estimate of which segments are larger, and explain that you need a survey to find out how many companies are in each segment.

When you have the budget, run the survey with a sample of companies from the whole market. A research agency can find the companies for you. Ask one question that lists each segment’s problem in the buyers’ words, and ask each company to choose the problem that would make it look for a product like yours. Add 2 more answers, “a different problem” and “no problem”. If many companies choose “a different problem”, interview some of them, because you may have missed a segment. Ask for the company’s industry, its number of employees and the job title of the person answering.

The scheduling company surveys 600 companies, and 35 per cent of them choose the missed appointments problem. The marketing team multiplies the market size of 40,500 by that percentage and estimates about 14,000 companies in Stop losing jobs. Very few companies choose “a different problem”, so the marketing team keeps the 4 segments.

After the survey, the order is different. Pass the audit had 2 of the 50 deals and ranked last. The marketing team now estimates about 6,000 companies in it, so it is the third-largest segment. It had only 2 deals because the sales team had rarely sold to fire safety, medical equipment servicing and utilities contracting companies, where customers ask for proof of every visit.

6. Add industry and company size to each segment

Industry and company size are called firmographics. Add them to the worksheet now. The media team and the sales team need the industry, the company size and the job title to reach the buyers in each segment.

For a draft segmentation, take the industries and company sizes from your interviews and the 50 deals. After the survey, check which industries and company sizes appear more often among the companies in a segment than among all the companies surveyed, and write them down. For the scheduling company, trades companies with 20 to 200 employees appear more often in Stop losing jobs.

The problem buyers want fixed defines the segment, and firmographics tell you where to find it.

7. Test the segmentation before anyone uses it

Kotler and Armstrong set out requirements for effective segmentation. Turn each requirement into a check on your worksheet.

• Measurable. You have counted the companies in every segment. If you only have a ranking, write “not yet counted” and ask the business stakeholder for a survey budget. Keep using the draft segmentation in the meantime.

• Accessible. You can say where to reach each segment: which industries, which company sizes and which job title.

• Substantial. Each segment has enough companies to be worth its own campaign.

• Differentiable. You can place every company in one segment by the problem that made its buyer start looking, and no 2 segments would respond to the same message.

• Actionable. You can describe the product, the message and the sales approach each segment would need.

Then check whether the buyers in each segment are customers, competitors’ customers or companies that bought nothing. If every buyer in a segment is your own customer, you don’t yet know whether competitors’ customers have the same problem. Interview more competitors’ customers before you use the segment.

When a segment fails a check, go back to the interviews. Merge the segment with another segment, split it, or interview more buyers.

8. Keep every segment on the worksheet, including the ones you won’t serve

The scheduling company can’t yet serve companies that schedule across several regions. The marketing team is tempted to delete Run a national fleet. If the marketing team deletes the segment now, the business stakeholder never gets to decide whether to serve those 1,500 companies. Keep Not buying on the worksheet as well, because some of those companies will later have a scheduling problem and move into a segment.

Choosing which segments to serve is called targeting. Kotler and Armstrong define market targeting as evaluating each market segment’s attractiveness and selecting one or more segments to enter. The business stakeholder makes that decision with the marketing team after the segmentation is finished.

Take the worksheet to the business stakeholder who wrote SMB New Logos on the campaign brief. The business stakeholder can now see what the buyers in each segment want and how many companies are in each segment. Then the business stakeholder chooses the segments for the campaign and writes them on the campaign brief.

Use AI for the slow parts, and check its work

Record each interview if the buyer agrees, and tell the buyer you’ll use an AI tool to transcribe it. With a transcript, you have the buyer’s exact words. Ask an AI assistant to pull the buyer’s answers out of each transcript and list them under the 6 interview questions. Then ask it to group the buyers by the problem that made each buyer start looking. It can also write each segment’s problem as an answer for the survey question.

Check its work against the transcripts. For each buyer, confirm that the problem the AI assistant chose is the problem the buyer said made them start looking, and that every quote appears in the transcript. Check your company’s policy before you put customer conversations into an AI tool.

Next steps

• Find the target segment on your next campaign brief, and ask the business stakeholder what problem those buyers want fixed.

• Write the market definition and work out the market size from public data.

• Ask the sales team for the deals lost in the last 12 months and book 12 interviews.

• Sort the interviews by problem, and rank the draft segments using the last 50 deals.

• Add industry and company size, run the checks, and ask for a survey budget.

• Take the worksheet to the business stakeholder who wrote the campaign brief.

You can start even if you’re a junior marketer and the campaign brief isn’t yours. Book 3 of the 12 interviews this week, with companies that chose a competitor. If the 3 buyers describe different problems, show your manager the interview notes and ask for time to do the other 9 interviews.

Evidence informs. Judgement decides.

The Segmentation Build

The Segmentation Build is the one-page worksheet used in this guide. On the worksheet you write the market definition and the market size, fill in one row for each segment and run the checks. The interview questions and a worked example are printed on it, so you can fill it in without this guide open.

The Segmentation Build

118KB ∙ PDF file

Download

Use this worksheet to help determine your own market segments

Download

The argument for why segmentation comes before targeting is in the essay Segmentation Is How You Understand a Market. Targeting Is How You Choose. The In Practice guide on targeting will be published after this guide.

New essays land here first: subscribe and the next one will find you.

And if you want to build the judgement, not just the output, Back2Marketing’s B2B Marketing Fundamentals course teaches marketing from first principles.

Subscribe now

Sources

• Philip Kotler and Gary Armstrong, Principles of Marketing. Definitions of market segmentation and market targeting, and the requirements for effective segmentation (measurable, accessible, substantial, differentiable and actionable).

• Greg Guest, Arwen Bunce and Laura Johnson, How Many Interviews Are Enough? An Experiment with Data Saturation and Variability, Field Methods, vol. 18, no. 1, February 2006. 60 in-depth interviews with women in 2 West African countries, from a health research study. Saturation occurred within the first 12 interviews. After 12 interviews the authors had created 92 per cent of the codes developed for the first 30 transcripts, and 94 per cent of the high-prevalence codes had been identified within the first 6. The authors caution that 12 interviews will likely not be enough if the group is relatively heterogeneous. Primary source

• Australian Bureau of Statistics, Counts of Australian Businesses, including Entries and Exits, July 2022 to June 2026, released 18 August 2026. Free counts of actively trading businesses by industry and by employment size range. Primary source

• Stats NZ, New Zealand Business Demography Statistics, at February 2025, released 30 October 2025. Counts of enterprises by industry and by employee count.

A strategy built on the audience slide is a guess. Research what buyers want and you can defend every decision.

Most marketing plans open with an audience slide. It lists company size, industry, region and a few job titles, and sometimes it adds a persona with a stock photograph and an invented name. The slide usually gets called segmentation. Most of the time it’s a description of the customers the company already has, sorted by whichever fields the customer relationship management (CRM) system stores.

The company then picks its target customers from that slide and builds its positioning and campaigns on the pick. If the slide describes the wrong thing, the budget and the sales team’s time go to the wrong buyers.

Segmentation is a separate piece of work. It has to come before any of those decisions, because it tells you what each group of buyers wants and how large each group is.

The audience slide is useful, and it’s easy to see why companies stop there

Industry, size and location tell you where to find buyers and how to reach them, and no media plan works without them. The data is already in the CRM and costs nothing to use. Research takes time and money, and the marketing plan is usually due first.

What the slide can’t tell you is what any of those buyers want, and a strategy ought to be built on what buyers want.

Segmentation turns a market into something you can decide about

A market of 40,000 companies is too large and too varied for a marketing team to act on. Sorting it by industry and headcount doesn’t fix that. A 400-person manufacturer and a 400-person law firm sit in the same size band. The manufacturer is buying to stop a production line going down, and the law firm is buying to get partners out of administrative work. One proposition can’t serve both, and one media plan won’t reach both sets of decision makers.

Kotler and Armstrong define market segmentation as dividing a market into smaller groups of buyers with distinct needs, characteristics or behaviours who might require separate products or marketing mixes. Divided into a small number of groups, each wanting a different outcome for a different reason, the market becomes a set of options a company can compare.

Kotler and Armstrong also set out the requirements a segmentation has to meet before it can be used to choose target segments. A segmentation that fails them is only a description of the market.

Measurable. You can say how many companies sit in each group and roughly what the category is worth to them. Without a number on each group, one segment can’t be compared with another, and targeting becomes a matter of taste.

Accessible. You can reach and serve the group. A segment defined only by an attitude tends to fail here, because there’s no way to find the people who hold it.

Substantial. The group is large enough to be worth serving. In B2B this catches the segment built around three flagship accounts, which is a sales plan rather than a segment.

Differentiable. The groups respond differently to different marketing. If two segments would respond the same way to the same campaign, one segment has been split in two for no reason.

Actionable. The company can build programmes to attract and serve the group. If the product can’t serve it and the sales team can’t support it, it’s a finding about the market rather than an option the company can take.

A usable segmentation gives the business a list of segments to choose between, with the size and value of each one known.

Market segmentation describes the whole market, including buyers you’ll never serve

A segmentation covers everyone who buys the category: the companies buying from competitors, the companies that haven’t bought from anyone yet, and the groups the company has no intention of selling to.

Including buyers you won’t serve looks like wasted effort, but a segmentation drawn only around existing customers can’t show which groups you’re absent from or how large they are. Nor can it show whether the segment you’re winning in is the one that’s growing.

A segmentation that only contains your own customers describes your customer base, not the market.

A company that knows which groups it’s absent from, and which groups are growing, can write a strategy. Without that knowledge it can only plan to do more of what already worked.

Segmentation belongs to the market rather than to the company, because the groups exist before anyone researches them. Two competitors who each research the category properly should find the same groups. What each company does with the groups comes later, at targeting.

A segment holds together around what its buyers are trying to achieve

Groups drawn on industry and headcount tend to contain buyers who want different things. The most common example is treating small and medium-sized business (SMB) as one segment. A 30-person accounting firm buys to get invoices out without hiring an administrator, and a 300-person logistics company buys to keep its depots running. Both count as SMB.

The two firms have different budgets, different buyers and different problems, and the only thing they share is that neither is an enterprise. SMB is a size band with a set of assumed needs attached. It doesn’t meet the requirements for a segment.

What holds a segment together is what its members are trying to achieve, what they believe about the category, and how they buy. In B2B that means asking what triggers the purchase, what the buyer is measured on, what happens if they do nothing, who else must approve it, and what would make them leave their current supplier. Two companies that answer those questions the same way will respond to the same proposition, even if one runs hospitals and one moves freight.

Firmographics, the facts about a company such as its industry and size, come back at the end. Once the groups exist, you look at what each group’s members have in common from the outside: the industries they cluster in, the sizes they tend to be, the roles that lead the buying. Industry, size and role are what make a segment findable. Without them you can’t buy media against the segment or tell the sales team who to call.

What buyers need defines the segment, and firmographics tell you where to find it.

The description comes from research, not from the database

What buyers are trying to achieve isn’t in the CRM. The CRM holds what the company recorded about the customers it won. It has nothing about the buyers who chose someone else, and nothing about why anyone bought.

You can’t discover what buyers want from a database of what they bought.

Finding out why buyers bought takes research. Interviews come first, with buyers who chose you, buyers who chose a competitor and buyers who did nothing. The interviews tell you what the issues are, in the words buyers use rather than the words the category uses. They’re where you find out that two industries you treat separately are solving the same problem, or that one group’s real alternative is hiring someone rather than buying software.

Then comes a survey with a representative sample, to find out how common each issue is and how many companies sit in each group. The survey puts a number on each group, and that’s what lets you compare one segment with another.

Interviews on their own give you a detailed description you can’t size. A survey on its own can ask the wrong questions, because nobody checked what buyers were trying to do before writing the questionnaire. So the interviews come before the survey, and the questionnaire is written from what the interviews found.

Cluster the survey responses on needs and buying behaviour first, then look at where each cluster differs from the market average on industry, size and role. Stop at the smallest number of groups that keeps buyers who want different things apart.

Research is the part companies skip, and targeting and positioning both depend on it. A segmentation assembled from internal data describes the company’s past, while one built from research describes the market as it is now.

Once each segment has a description and a number, the company has a real choice to make.

Segmentation makes targeting possible, and targeting is where the company commits

Entering a segment means building the product for those buyers, making the proof they ask for, teaching the sales team how they buy, and pricing for the size of company involved. Targeting is the first point in a marketing strategy where the company gives something up.

Kotler and Armstrong define market targeting as evaluating each market segment’s attractiveness and selecting one or more segments to enter. They also set out what to weigh. Size comes first, meaning how large the segment is and whether it’s growing. Competition comes next: who else wants these buyers, how strong those rivals are, how easily buyers can switch, and how much power buyers hold over price. Fit comes last, and it asks whether serving the segment suits the company’s own objectives and resources.

Fit is the test that rules out most segments that look attractive on size alone. Enterprise buyers in regulated industries are a large segment with real budget. Serving them takes certifications, security review capacity and an implementation team that a smaller company doesn’t have.

In B2B the choice isn’t optional. B2B vendors have no mass market to fall back on, so they must target defined segments, as the Marketer’s Toolkit from Harvard Business School Press puts it. Every B2B company ends up with target segments, chosen deliberately or by default.

The essence of strategy is choosing what not to do. Michael Porter wrote that in 1996, and targeting is where marketing makes that choice. Naming the segments the company won’t pursue frees the budget, the roadmap and the sales capacity for the segments it will.

A company can’t evaluate the attractiveness of groups it has never defined, so it ends up choosing between the customers it already has and the customers a competitor just won.

Segmentation has done its job when someone can say no to a segment and explain why.

Segmentation makes positioning possible, because a position exists in the minds of a particular group

One position stretched across a whole market says very little, because different groups buy the same product for different reasons. Kotler and Armstrong define positioning as arranging for a product to occupy a clear, distinctive and desirable place, relative to competing products, in the minds of target consumers.

Every part of the definition depends on knowing the segment. The minds belong to a particular group of buyers. The competing products are the ones that group compares you with, and different groups compare you with different things. Desirable means desirable to that group of buyers, and the only way to know what they want is to have asked them.

You can’t hold a position in the minds of buyers you haven’t researched.

The competing products are where segmentation matters most, and where it’s most often skipped. A buyer’s frame of reference is the set of options they think they’re choosing between, and it’s rarely the category the company uses to describe itself. One group is comparing two suppliers. Another group is deciding whether to buy anything at all, or to hire two more people and keep working the way they do now.

A company that knows what each segment is comparing it with can position against those alternatives. Working from one audience slide, a company positions against the competitors it chose to name, which may not be the alternatives buyers are weighing up.

A company can hold one broad brand position with a sharper position for each target segment underneath it. The company can only write the sharper positions if the segments are real.

What a good segmentation looks like

To help illustrate, take a software company that sells scheduling tools to businesses that send staff out to customer sites: trades, equipment servicing, home care, facilities maintenance. Its audience slide says SMB field service companies, 10 to 500 staff, Australia and New Zealand. Its CRM holds 900 customers, most of them plumbing and electrical contractors, because that’s who the founders knew.

The company interviews 36 buyers: current customers, companies that chose a competitor, and companies that looked and bought nothing. Then it surveys 600 companies across the whole market. The groups that come out are defined by what the buyer is trying to fix, not by trade or headcount.

Stop losing jobs. For these businesses a missed or late appointment means lost revenue and a complaint. The operations manager buys, and the trigger is a run of complaints. The alternative is hiring another dispatcher. About 14,000 companies sit in this group, mostly trades with 20 to 200 field staff.

Get paid faster. The job is done on Tuesday and the invoice goes out three weeks later. The owner or the finance lead buys when cash is squeezed, and the alternative is a part-time bookkeeper. Some 9,000 companies are in this position, most of them small.

Pass the audit. About 6,000 companies in fire safety, medical equipment servicing and utilities contracting have customers who demand proof of every visit. The compliance or quality manager buys after a failed audit or a new contract clause. The alternative is paper forms and admin staff.

Run a national fleet. Multi-region operators with in-house IT want scheduling wired into their enterprise resource planning (ERP) system and their reporting. The IT or operations director buys, slowly. The alternative is building it themselves. They number about 1,500 and have the largest budgets in the market.

Not buying. These businesses have a few field staff and run on phone calls and a whiteboard. There are about 10,000 of them. They’re in the segmentation because the market isn’t complete without them, and because some of them grow into the groups above.

Every company in the market sits in exactly one group, including the companies that buy from competitors. Trade and headcount appear only at the end, as the way to find each group.

Targeting is now a decision the company can defend. It chooses stop losing jobs and get paid faster: the product already does both, the buyers are reachable through trade associations, and the sales team knows them. It says no to run a national fleet because it has no integration team, and no to pass the audit until the compliance features ship next year.

Positioning follows. To the operations manager, the product is the alternative to another dispatcher. To the owner, it’s the alternative to a bookkeeper. The company has one product and two positions, both written from what buyers said.

A good segmentation names what each group wants, puts a number on it, and includes the buyers you won’t serve.

The company has now chosen its segments. It still has to decide who should see its advertising, and the common answer is wrong.

Choosing a market isn’t the same as narrowing your reach

Segmentation and targeting decide where the company competes. How widely to advertise is a separate decision, and marketers often merge it with the choice of segments.

Targeting has a cost. The Marketer’s Toolkit from Harvard Business School Press warns that focusing on narrow segments reduces the number of buyers for whom the offer represents value. A narrow focus also ties the company’s future to the fortunes of the segments it chose. Every time you narrow the audience, you lose buyers. Narrow it once, when you choose the target segments, and use that choice for the product, the sales team and lead generation. Keep brand advertising broad.

Many marketers do the opposite, buying audience data and advertising only to the job titles that match the target segment. 68 per cent of B2B marketers believe hyper-targeting is more effective than broad targeting, according to the LinkedIn B2B Institute’s 2030 B2B Trends report. The evidence doesn’t support them.

Ad-tech vendors sell audience segments labelled IT decision makers. Nico Neumann, Catherine Tucker, Kumar Subramanyam and John Marshall tested those segments in a 2023 study published in Quantitative Marketing and Economics. Between 9.4 and 18.2 per cent of the people in the segments were IT decision makers, depending on the vendor. Showing the same ads to a random audience with no targeting at all reached IT decision makers 16 per cent of the time. Paying for the segments bought nothing.

The same study found what does work. Ads placed on business content reached IT decision makers 41.8 per cent of the time. The placement used the publisher’s own data about what its readers look at. Ad targeting works when it uses the publisher’s own data, and it fails when it relies on bought third-party segments.

Campaigns that address customers and non-customers together do best on business results. These reach strategies scored 1.6 on Les Binet and Peter Field’s measure of very large business effects, against 1.0 for acquisition strategies and zero for loyalty strategies. The scores come from Binet and Field’s analysis of B2B cases in the IPA Databank.

Up to 95 per cent of business buyers aren’t in the market for a category at any one time, according to John Dawes at the Ehrenberg-Bass Institute. The ratio is a rule of thumb rather than a constant. Most of the buyers who will eventually choose a supplier aren’t comparing anything today. A campaign aimed only at buyers who look ready now misses them.

Binet and Field separate brand building from sales activation. Sales activation, which includes lead generation, asks a buyer to respond now, and they write that it should be tightly targeted on buyers who are ready to respond. Brand building works on buyers long before they come to the market, and it needs much broader reach. So lead generation goes to the target segments. Brand advertising should reach every company the business serves now or could serve later, as far as the budget allows.

The target segments decide where the product, the sales team and lead generation go, and brand advertising should still reach every buyer you could serve.

Before you accept a segmentation, put it through these checks

A segmentation usually arrives as a slide in a planning meeting, ready to be approved. Run these checks before anyone approves it.

1. Find your competitors’ customers in it. If every group is populated by companies you already sell to, the work described your customer base. It can’t tell you where growth is.

2. Place a company in exactly one group. If an account could sit in two groups, the dividing lines aren’t doing anything, and two teams will market to the same company in two different ways.

3. Name each group for what its buyers want, not for how much you want them. A segment called Mid-Market Manufacturing tells you where to find people. Name it for the outcome its buyers are chasing and it tells you what to say to them. Names like Tier One or Priority Accounts describe a targeting decision, not a segment, and they’re the names that get used once the sales manager stops reading the segment profiles.

4. Put a number on each group. Record how many companies it holds, what they spend on the category, and whether that spend is rising. Segments without numbers can’t be compared, and a targeting decision made without comparison is a preference.

5. Ask the sales team to name real accounts in each group. If the people talking to buyers every day can’t place accounts, the segmentation came out of a dataset rather than out of the market, and nobody will use it.

If a segmentation fails these checks, it isn’t ready to choose from, and any targeting decision made on top of it is a guess with numbers attached. If it passes, the hard part is over. What’s left is deciding which groups to serve, and writing down which groups you won’t.

Segmentation comes first because everything else depends on it

Segmentation gets treated as preparation, the slide you build before the planning starts. It’s the first step in planning. A company that understands its market in detail can make a targeting choice it can defend, and hold a position that means something to the buyers it chose. A company working from an audience slide is guessing, and the errors compound. If the segmentation is wrong, the targeting is wrong, and the positioning and every tactical decision after it are wrong too.

All of this can be done without new technology. It takes asking buyers what they’re trying to achieve, counting how many of them there are, and being willing to write down which groups the company won’t serve this year. A marketer at any level can open the current audience slide and ask which of its groups came from research and which came from the CRM.

We believe marketing is a commercial discipline, not a promotional service.

We believe the decisions that matter most are made before anyone writes a brief, and that they belong to the business as much as to marketing.

We believe judgement is what turns evidence into a decision, and that no amount of data removes the need to choose.

Segmentation is how you understand a market. Targeting is how you choose.

Back2Marketing exists to help marketers make better decisions, by combining the best available evidence with commercial judgement and practical experience.

Evidence informs. Judgement decides.

New essays land here first: subscribe and the next one will find you.

And if you want to build the judgement, not just the output, Back2Marketing’s B2B Marketing Fundamentals course teaches marketing from first principles.

Subscribe now

Sources

• Philip Kotler and Gary Armstrong, Principles of Marketing. Definitions of market segmentation, market targeting and positioning, the requirements for effective segmentation, and the factors to weigh when choosing segments. Edition and year to confirm.

• Harvard Business School Press, Market Customization: Segmentation, Targeting, and Positioning, chapter 4 of Marketer’s Toolkit: The 10 Strategies You Need to Succeed, 2006. B2B vendors must target defined segments, and a narrow focus reduces the number of buyers and ties the company to the segments it chose.

• Michael E. Porter, What Is Strategy?, Harvard Business Review, November and December 1996. “The essence of strategy is choosing what not to do.”

• LinkedIn B2B Institute, 2030 B2B Trends: Contrarian Ideas for the Next Decade, 2020, Trend 03, The Death of Hyper-Targeting. 68 per cent of B2B marketers believe hyper-targeting is more effective than broad targeting.

• Nico Neumann, Catherine E. Tucker, Kumar Subramanyam and John Marshall, Is first- or third-party audience data more effective for reaching the ‘right’ customers? The case of IT decision-makers, Quantitative Marketing and Economics, vol. 21, no. 4, 2023. Bought IT decision maker segments were 9.4 to 18.2 per cent accurate, a random audience reached 16 per cent, and placement using publisher data reached 41.8 per cent.

• Les Binet and Peter Field, The 5 Principles of Growth in B2B Marketing: Empirical Observations on B2B Effectiveness, LinkedIn B2B Institute, 2019, drawing on B2B cases in the IPA Databank. Reach strategies scored 1.6 on very large business effects, acquisition strategies 1.0 and loyalty strategies zero. Activation should be tightly targeted on hot prospects, and brand building requires much broader reach.

• John Dawes, Ehrenberg-Bass Institute with the LinkedIn B2B Institute, Advertising Effectiveness and the 95-5 Rule: Most B2B Buyers Are Not in the Market Right Now, May 2021. Up to 95 per cent of business buyers are not in the market for a category at any one time.

The request usually arrives already answered. We need a social strategy. We should be doing more video. We need an AI content engine. Put more into paid search. Someone has picked the tactic, and marketing is being asked to build a case for it.

Nine decisions sit in front of every one of those requests. Until they are made, there is no way to tell whether more video is the right answer or an expensive distraction.

Everyone Can Now Build. Few Know How to Market. argued that as execution gets cheaper, the value moves to the decisions upstream of it. These are those decisions, in the order they have to be made, with what each one produces and how to tell whether you have actually made it.

Work through them in order. Each one narrows the next.

1. Diagnose the situation

Before deciding what marketing should do, establish what is happening.

Three views:

1. The market: how large is it, is it growing or contracting, where is growth coming from, what are competitors doing, what has changed structurally in the last two years.

2. The business: where revenue comes from today by product, segment and region, which parts are growing, where margins are strongest, where performance is deteriorating.

3. The customer: what they value, how they buy, what they consider instead, what stops them buying, and what they currently think of you.

What it produces: one page naming the single most important problem or opportunity marketing needs to address, with the evidence for it.

How to tell you have done it: you can state the problem in one sentence, and point to the data behind it. A pile of research with no conclusion is not a diagnosis.

2. Segment the market

Split the market into groups of customers who behave differently enough to need a different response from you.

Choose the basis deliberately. Demographics and firmographics are convenient rather than useful. Needs, buying situation, category involvement, attitudes, occasions and value to the business usually explain more about why people buy.

In B2B, company size, industry and geography are how the database is organised. That is account classification, not segmentation.

What it produces: three to six named segments, each with its size, its growth rate, its value to the business and what makes its buying behaviour different.

How to tell you have done it: the segments change a decision. If your segmentation does not alter who you target, what you offer, how you position or where you invest, it is a description of the market and nothing more.

3. Choose where you will compete

Targeting is the point at which strategy requires sacrifice. Some segments are growing fast and fiercely contested. Others are profitable but small. Some look attractive until you check whether the company can actually serve them.

Score each segment on three things: how attractive it is commercially, how winnable it is given your capabilities and brand, and how well it fits the objective the business has set.

You can target more than one segment. You still have to rank them, because a budget spread evenly across every possible audience is usually too thin to work anywhere.

What it produces: a ranked list of segments with a share of investment against each, and a written list of who you are choosing not to pursue this year.

How to tell you have done it: the list of what you are not doing exists, and someone senior has agreed to it.

4. Decide your position

Positioning is the place you want to occupy in the customer’s mind relative to the alternatives they could choose, including doing nothing.

Write it for the segments you ranked in step 3. What should make you relevant and distinctive to them, and why should they pick you over the other options in front of them.

Then check it is credible. The product has to support the position, and so do the price, the customer experience and the route to market. Communications can reinforce a position. They cannot hold one up while the rest of the business contradicts it.

What it produces: a positioning statement per priority segment, plus the proof points that make each one believable.

How to tell you have done it: you can name what a competitor would have to stop doing for the position to be theirs instead of yours.

5. Set the marketing objectives

Business objectives and marketing objectives are not the same thing. “Grow revenue by 15 per cent” is a business objective. Marketing’s job is to say what has to change in the market for that growth to become possible.

That might be brand awareness, consideration, penetration, trial, retention, price perception, or demand in one specific segment. It comes from the diagnosis in step 1: if awareness is already 90 per cent, raising awareness is not the useful goal.

What it produces: three to five objectives, each with a measure, a starting number, a target and a date.

How to tell you have done it: each objective traces back to a line in the diagnosis, and you can say which business objective it serves.

6. Check the product supports the strategy

Product is part of marketing, even where marketing does not own it.

Does the product solve the need the research identified? Does it deliver the value the positioning promises? Are there features customers do not value, or something important missing? Does the portfolio make sense to a buyer, or only to the organisation that built it?

Sometimes the most valuable marketing decision is a change to the offer rather than a new campaign. Promotion will not carry a product the market does not want.

What it produces: a short list of product changes the strategy depends on, with an owner for each and a date by which the answer is needed.

How to tell you have done it: the product owner has seen the list and either committed to it or told you it is not happening, so you can plan around the answer.

7. Check the price supports the strategy

Price sets demand, signals value, reinforces or undermines the positioning, and decides the economics of acquiring a customer.

A premium position with bargain pricing creates one problem. A mass-market proposition priced at a premium creates another. The right price reflects customer value, competitive alternatives, willingness to pay and the commercial objective, rather than cost plus a margin or the highest number finance thinks the market will accept.

This is also where the unit economics get checked. Revenue growth that destroys margin is not good growth, and a customer who costs more to acquire than they will ever return is not a marketing win because a dashboard counted the conversion.

What it produces: a written view on whether current pricing supports or undermines the position, and the cost to acquire and serve a customer in each priority segment.

How to tell you have done it: you know the payback period on a new customer in your top segment.

8. Build the communications strategy

Only now do the channel questions become answerable, because you know who you need to reach, what they currently think, what you want them to think instead and what has to change.

Decide in this order: the message, the creative idea that makes it memorable, how much reach and how often over what period, the split between brand building and activation, and only then the channels.

Binet and Field’s analysis of the IPA Effectiveness Databank found that campaigns weighted too far toward short-term activation underperform over time, with roughly a 60:40 split of brand to activation as the long-run balance in the cases they studied. Their later work with business marketers put the B2B balance nearer an even split.

This is also where AI belongs, on research, analysis, ideation, production, personalisation and optimisation. It makes execution faster. It does not make a decision you have not taken.

What it produces: a communications plan with a message, a reach and frequency target per segment, a brand and activation split, and a channel list that follows from all three.

How to tell you have done it: you can explain why each channel is on the list by pointing at a segment and an objective.

9. Make sure customers can actually buy

Distribution gets the least attention and loses the most revenue. In B2B that means direct sales, partners, resellers, marketplaces, self-service, account teams, or some combination.

The route has to match the targeting and the position. A high-touch enterprise proposition and a low-cost self-service offer need different routes, and a premium brand may deliberately restrict availability because being everywhere would undermine the position.

What it produces: the route to market for each priority segment, and a list of the places a customer currently cannot buy from you but should be able to.

How to tell you have done it: you have asked a customer to describe how they bought, rather than assuming you know.

Use the sequence backwards when a plan is not working

When communications are generating attention and sales are not moving, do not start with the media plan. Go back up the list.

Can customers buy easily? Is the price right? Does the product deliver what the positioning promises? Is the position relevant to the people you targeted? Did you target the right segments? Was the diagnosis correct?

The first question you cannot answer is where the plan actually broke. It is almost never the last thing you changed.

The same applies to the tactical request. Before answering “should we be doing more video”, ask which objective it serves, which segment it reaches, what has to change for them, and what the diagnosis said. If those answers are not there, the conversation about video is premature.

Next steps

• Write the one-page diagnosis, with the market, business and customer views and a single named problem

• Draw up three to six segments and check that each one would change a decision

• Rank the segments, allocate a share of investment to each, and write down who you are not pursuing

• Write a positioning statement per priority segment with its proof points

• Set three to five marketing objectives, each with a measure, a starting number, a target and a date

• Take the product and price questions to whoever owns them and get an answer you can plan around

• Build the communications plan last, message first and channels last

• Map the route to market for each priority segment

Only then is the tactical conversation worth having.

Evidence informs. Judgement decides.

B2B Marketing Strategy

The Strategy Sequence Check

A one-page diagnostic listing all nine decisions, what each one should have produced, and a column to tick whether it exists in writing and where it lives. The first decision you cannot tick is where your strategy stops and your assumptions start. Useful before a planning cycle, and useful when an existing plan is not working.

The Strategy Sequence Check

108KB ∙ PDF file

Download

One page, nine decisions that come before any tactic. Tick only what exists in writing. The first one you cannot tick is where your strategy stops

Download

If you want to work through the whole sequence properly, the B2B Marketing Fundamentals course covers it end to end and finishes with your own One-Page Growth Strategy.

Sources

• Theodore Levitt, Marketing Myopia, Harvard Business Review, 1960. Businesses defined by the products they make rather than the customer needs they serve become vulnerable when the market moves. Primary source

• Philip Kotler, Marketing Management. The planning sequence used here: research and analysis, then segmentation, targeting and positioning, then the marketing mix, then implementation and control.

• Les Binet and Peter Field, The Long and the Short of It: Balancing Short and Long-Term Marketing Strategies, IPA, 2013. Analysis of the IPA Effectiveness Databank. Campaigns weighted too far toward short-term activation underperform over time; roughly 60:40 brand to activation as the long-run balance in the cases studied. Their later work with business marketers indicates a B2B balance nearer an even split, and the authors describe that figure as tentative. Primary source

I have sat in planning meetings that ran like this. Finance opens with the number the business has to hit: grow revenue fifteen per cent. Four slides later, the second instruction lands. Take five per cent out of every cost line, flat, across the board. No guidance on which parts of the plan matter more than others.

Marketing then spends three weeks rebuilding last year’s plan to fit a number it had no hand in setting.

The Budget Gets Set Before the Strategy argued why that order is wrong. Here are the seven steps I use instead, and a worksheet to run them on.

1. Split the revenue target by source of growth

“Grow revenue by fifteen per cent” tells you the size of the gap. It does not tell you what has to happen for the business to close it. Splitting the target means writing down which parts of the business produce those fifteen points, and how many points each one produces.

There are a limited number of places growth comes from. Work through the list:

• more customers in markets you already sell to

• existing customers buying more, buying more often, or buying a second product

• a price rise or a shift in product mix

• a new product

• a new segment or a new geography

• fewer customers lost

Get the split from finance first. The annual plan or the board pack usually carries revenue by product, by region, by segment, and by new versus existing customer. That is most of the work already done. Where it does not exist, build a draft yourself from last year’s actuals plus the growth assumptions each business owner is carrying. Then send that draft to your finance business partner and the sales leader and ask them to mark it up, rather than asking them to produce one from scratch.

Then test it. The points add to fifteen. Every line has a named owner. Every line is plausible against what that part of the business delivered last year. If cross-sell grew two points last year and the plan needs five, someone has to say what changes to make up the difference.

The marketing requirement for each source of growth is different. Six points from a segment the company does not sell to today needs awareness, credibility and demand creation from a standing start, and that is expensive. Five points of cross-sell to customers who already buy needs customer marketing and sales enablement, which costs far less. A single blended target gives you no way to tell those apart, and an even five per cent cut treats them as if they were the same.

Ask finance for the revenue plan split by product, segment and new versus existing customer. That document is where the marketing budget starts.

2. For each source of growth, name what is blocking it

For each source of growth, write what has to become true among customers. Six points from mid-market means buyers in that segment have to know the company exists, believe it serves businesses their size, and have a reason to put it on a shortlist.

Then name the single biggest obstacle standing in the way: awareness, consideration, the proposition, price, the route to market, sales capacity, or product fit.

This is what decides where the money goes. If mid-market buyers already know the company but do not believe it serves businesses their size, more media buys more of the wrong impression. The work is proposition and proof: segment-specific messaging, customer evidence, pricing and packaging. If they have never heard of the company at all, reach is the job and media is the answer.

Write the obstacle next to each source of growth and keep it there. When someone proposes a new campaign or channel in March, you have a written test for it: which of these obstacles does it remove, and for which source of growth?

Fund the thing blocking growth, not the activities that ran last year.

3. Convert marketing’s contribution into outcomes with owners

Write what marketing will deliver as measurable outcomes rather than as a list of campaigns.

Not six campaigns, three events and a content programme. Instead: unaided awareness in the mid-market segment from 8 to 20 per cent by the end of Q3. Consideration from 12 to 20 per cent. Four hundred qualified opportunities into the mid-market sales team at an average deal size of $40,000.

Each outcome needs a measure, a baseline, a target, a date and a named owner. Written that way they can be tracked through the year, defended in a review, and renegotiated when the budget moves.

Write the dependencies on other teams in the same place. If four hundred opportunities only produce the revenue when sales converts at 22 per cent, and sales has converted at 15 per cent for two years, put both numbers in the plan and name who owns closing that gap. Without it, marketing carries the revenue number while another team controls half of what produces it.

An outcome with a baseline, a target, a date and an owner survives a budget negotiation. An activity list does not.

4. Cost what the strategy requires, then take it to finance

Now build the cost, starting from the work each outcome needs: research, brand development, creative, content, media, events, partner programmes, product marketing, customer marketing, technology, agency support, headcount.

Mark Ritson made this argument for zero-based budgeting in 2016: the team researches, builds the plan and finishes with a request and a promised return, rather than inheriting a percentage of projected sales.

Cost the full period the outcomes need. Awareness built in Q3 does not convert in Q3. If the plan needs twelve months of sustained investment and the budget funds three, the plan will not deliver what it says it will.

Put the full cost in front of finance before you take anything out of it. A number you have already trimmed to what you think will get approved leaves the business with no view of what the growth target actually costs.

Present the full cost of the plan. Let the business decide what to remove, with the consequence written next to it.

5. Separate committed, core and strategic money

Three categories, each with a dollar figure against it.

• Committed: contracts, technology platforms, sponsorships and agency retainers already signed

• Core: salaries, tools and infrastructure that run the function

• Strategic: media, research, creative development, new market activity, testing

Take a $10 million budget with $6 million in committed and core costs. Four million funds growth. A ten per cent cut to the total is $1 million, and because the committed costs are contracted, it comes out of the strategic line. That is a 25 per cent cut to the growth money.

Show all three figures every time the total is discussed, so the conversation about a cut happens against the $4 million rather than the $10 million.

Show the total, the committed and core costs, and the money actually available to fund growth.

6. Cost three scenarios and state what each one delivers

Minimum viable: the lowest investment that still delivers at least one source of growth in full. Do fewer things properly rather than fund all three sources at a level too low to work anywhere.

Recommended: the investment that funds every outcome in step 3, with the evidence and assumptions attached.

Accelerated: what more money buys, and how much more it costs.

Against each scenario, write out what the business gets and what it gives up:

• which sources of growth are funded, part funded or dropped

• reach and frequency in each target segment

• which geographies are included

• what happens to the launch date

• what happens to brand investment against lead generation

Then put the revenue consequence next to each one, in the same units finance used. If the minimum viable scenario drops the mid-market segment, the plan delivers nine points of growth rather than fifteen. Write that on the slide.

If the plan goes to the board, present the marketing strategy to the board in the same terms: the growth each scenario delivers and what it costs.

Every scenario carries the growth it delivers. A cheaper budget with the same revenue target is not a scenario.

7. Check the recommendation against both benchmarks, last

Two checks, answering two different questions.

Spend as a percentage of revenue. Gartner’s 2026 CMO Spend Survey of 401 marketing leaders puts marketing budgets at 7.8 per cent of company revenue. The CMO Survey, covering 308 US marketing leaders, puts it at 9.0 per cent. Grace Kite’s analysis of three datasets, reported by Mark Ritson, supports 5 to 10 per cent of turnover on advertising, with 10 per cent as a practical starting point. Those three broadly agree, so treat 5 to 10 per cent as what a typical established business spends to hold its position.

Share of voice against share of market. This is the more useful test, because it measures you against your competitors rather than against an average of everyone. Binet and Field’s analysis of the IPA Effectiveness Databank found that brands spending a larger share of category advertising than their share of market tend to gain share, at roughly half a point of market share for every ten points of that excess. It varies by category, brand size, channel and pricing power, so use it as a direction rather than a formula.

Run both, then explain where you sit.

• Above 10 per cent of revenue. This is what buying share looks like. Taking share from a competitor means outspending your current share of market, and that shows up as a higher percentage of revenue. Entering a new market, launching a product, and recovering from several years of underinvestment all do the same thing. Say which one applies and for how long.

• Inside the range, but losing share. The total looks normal while competitors outspend you in the places that matter. Check share of voice inside the segments from step 1 rather than across the whole category.

• Below 5 per cent of revenue. Check whether the plan was costed for the growth it promises, or whether outcomes were dropped to make the number fit.

Use the benchmarks to test the number you built. Do not use them to produce it.

Next steps

• Get the revenue plan from finance, split by product, segment and new versus existing customer

• Write the growth split: source, points of the target, owner. Check the points add up and each line is plausible against last year’s performance

• Name the biggest obstacle blocking each source of growth

• Convert marketing’s contribution into outcomes with a baseline, target, date and owner, and write down the dependencies on sales and product

• Split the current budget into committed, core and strategic, and find out how much is actually available

• Cost the recommended scenario in full, then build the minimum viable and accelerated versions with the growth each one delivers

• Check the recommendation against both benchmarks, spend as a percentage of revenue and share of voice against share of market, and write down why you sit where you sit

Build the number from the growth plan, then defend it.

Evidence informs. Judgement decides.

The Budget Build

A one page worksheet for all seven steps: the growth split table, the constraint test, the outcomes table, the committed, core and strategic breakdown, and the three scenario grid. The three published benchmarks are printed on it, so the pressure test happens on the page instead of sending you looking for numbers. Fill it in and the whole budget conversation sits on one sheet.

The Budget Build Worksheet

117KB ∙ PDF file

Download

One page, seven steps. From the revenue target to three costed scenarios and a number you can defend. Worked examples and the benchmarks are printed on the sheet.

Download

If you want the planning sequence behind this, from diagnosis through segmentation, targeting, positioning and objectives, the B2B Marketing Fundamentals course covers it end to end and finishes with your own One-Page Growth Strategy.

Sources

• Gartner, 2026 CMO Spend Survey, published May 2026. 401 CMOs and marketing leaders in North America, the UK and Europe, surveyed January to March 2026. Marketing budgets average 7.8 per cent of company revenue, around 18 per cent below the mean four years earlier. Primary source

• The CMO Survey, 35th edition, 2026. 308 US marketing leaders, 97 per cent VP or above. Marketing budgets at 9.0 per cent of company revenue. Primary source

• Mark Ritson, Why Unilever is right to adopt zero-based budgeting, Marketing Week, January 2016. The team researches, builds the plan and ends with a budget request and a promised return, rather than accepting an arbitrary percentage of projected sales. Primary source

• Les Binet and Peter Field, The Long and the Short of It: Balancing Short and Long-Term Marketing Strategies, IPA, 2013. Analysis of the IPA Effectiveness Databank. Brands whose share of category advertising exceeds their share of market tend to gain share, at roughly half a point of market share for every ten points of excess share of voice; the relationship varies by category, brand size, channel and pricing power. Primary source

• Mark Ritson, ‘Triple-Cooked Marketing Budgets’: The full recipe, Marketing Week, October 2022. Grace Kite’s analysis of three datasets supports 5 to 10 per cent of turnover on advertising as a rule of thumb; Ritson proposes 10 per cent as a practical starting point. Primary source

I have been through enough annual planning cycles to know how marketing budgets get set. There is a number from last year. Finance has a view on what the business can afford this year. Revenue is expected to grow, so marketing gets a small increase. Margins are under pressure, so every function is asked to find savings. Some years the instruction is simply to hold the budget flat and deliver more. By the time the marketing team starts planning, one of the biggest strategic decisions has already been made.

Think about the sequence. The amount available to marketing has been fixed before the business has worked through what it needs marketing to achieve. The company might be entering a new market, launching a product, facing a far more aggressive competitor, trying to establish a category, or reversing a decline in consideration. The growth ambition might be materially higher than last year’s. Whatever has changed in the market, the starting point is usually the same: here is the number, now build a plan that fits inside it.

This is so normal inside large organisations that we rarely question the logic. When marketing does push back, a familiar question follows: “Well, what should we be spending?” That is when the benchmarks come out. Gartner says marketing budgets averaged 7.8 per cent of company revenue in 2026. The latest CMO Survey puts the figure at 9.0 per cent. Both are credible pieces of research. Neither can tell you what your marketing budget should be. Neither can last year’s spreadsheet.

This is not an essay about how much to spend. It is an essay about the order in which the decision gets made, because the order is where the damage is done.

When the budget is set before the strategy, the envelope has already made the strategic choices.

We have become very comfortable starting with the number

Marketing budget conversations have an anchor. Sometimes it is last year’s spend, sometimes a percentage of revenue, sometimes an industry benchmark. Often finance has already sized the pool and marketing is asked to work within it. The logic is understandable. Companies have finite resources, and marketing competes with product, sales, operations and technology for investment. A CMO cannot arrive at the CFO’s office with a wish list and expect it to be funded. Financial discipline matters, and marketers should be accountable for how they invest the company’s money.

The problem starts when the constraint substitutes for the strategy. Last year’s budget tells you what the business spent last year. A benchmark tells you what other companies spend. Neither tells you what needs to happen in your market this year. Yet both carry enormous influence because they feel objective. A budget set at 8 per cent of revenue has a reassuring logic. It can be compared with an external study, placed on a board slide and defended as broadly in line with the market.

The average, however, has no idea what you are trying to achieve. A category leader defending its position does not face the same task as a challenger trying to double its share. A business entering Japan for the first time is not in the position of one that has traded there for thirty years. A company with high awareness and weak conversion has a different problem from a company nobody has heard of. Two B2B software firms of similar size can have completely different brand strength, market maturity, sales coverage, pricing power, existing demand and competitive intensity. Put them all into a benchmark and that context disappears. What remains is a precise-looking average that may have very little to do with the decision in front of you.

There is a further reason to handle the number carefully. Gartner’s CMO Spend Survey represents one cohort of companies and marketing leaders. The CMO Survey represents another. Their methodologies, company mixes and definitions differ, which is one reason their headline figures do. That does not make either benchmark wrong. It is a reminder that a benchmark describes the organisations in the study. It does not prescribe what every company should spend.

A benchmark describes other people’s decisions. It cannot make yours.

Last year’s budget is a benchmark too

External benchmarks get the attention, but in my experience the stronger anchor is internal. We spent $20 million last year. Revenue is forecast to grow 5 per cent, so perhaps marketing gets another 5 per cent. The company needs to protect margin, so the budget stays flat. Everyone has an efficiency target, so marketing takes out 10 per cent. Once that number is on the table, it becomes the frame for the whole planning conversation.

There is good reason to use history. It tells us what things cost, what we are already committed to, and what worked, what did not and where money was wasted. The danger is when historical spend hardens into an assumption about future need. Last year may have included a major launch, or should have included one and did not. The company may have been overinvesting, or chronically underinvesting and has now built that underinvestment into its planning. Competitors may have stepped up, customer behaviour may have shifted, media costs may have moved, or the business may have set a growth target that bears no relationship to the resources behind it.

Once the previous number is the anchor, enormous organisational energy goes into debating small movements around it. Marketing argues for a few per cent more. Finance asks where savings can be found. Teams cut activities to fit the envelope. All of that can happen without anyone revisiting the more important question of what the business needs marketing to accomplish.

The current data suggests the tension is getting sharper. Gartner reports that marketing budgets have been on a plateau since 2022, with the 2026 average at 7.8 per cent of company revenue, 18 per cent below the mean allocation four years earlier. At the same time, 73 per cent of CMOs describe the growth expectations placed on them as high, very high or overly ambitious, and in the same 2026 research 56 per cent said their marketing organisation lacked the budget to deliver its strategy. There may be good reasons to constrain investment. But there is an obvious tension when growth expectations rise without any discussion of what level of investment that growth requires.

Starting from zero

There is a well-established idea in budgeting that challenges the habit of starting with last year’s number: zero-based budgeting. Mark Ritson has written about it for years. In a 2016 Marketing Week column on Unilever’s adoption of the approach, he challenged the assumption that it was simply a cost-cutting exercise. A proper zero-based approach, he argued, starts with the research and the marketing plan, determines what the plan requires and builds the budget from there, rather than accepting an arbitrary percentage of forecast sales and working backwards.

I like the principle more than the terminology. No sensible marketing leader walks into planning pretending the previous year never happened. Historical performance matters. Existing commitments matter. What the business can afford matters. We should know whether activity delivered a return and whether money was wasted. The useful part of zero-based thinking is removing the assumption that the previous allocation automatically deserves to become the next one.

The sequence changes when you approach it this way. You begin with the business problem and the growth ambition, understand what is happening in the market, decide where marketing can make a difference, build the strategy, and then determine what the strategy requires. The eventual budget still has to survive commercial scrutiny. Leadership may decide the company cannot afford the plan. Marketing may need to make trade-offs. The growth ambition itself may need to change. At least those trade-offs are made consciously, instead of being hidden inside an inherited number.

Ritson revisited budgeting in 2022 with his “Triple-Cooked Marketing Budgets” approach, and importantly he did not claim there was a perfect calculation for every brand. Drawing on econometrician Grace Kite’s analysis of three effectiveness datasets, he proposed broad rules of thumb as a practical starting point, then adjustment for the context of the business. The answer is not to replace last year’s arbitrary number with a different arbitrary number.

The value is in knowing what the evidence can tell you, and where judgement has to take over.

The budgeting problem is also a marketing problem

There is a deeper reason marketing budgets so often begin with cost. Marketing itself has spent years becoming defined by efficiency. Digital gave us unprecedented visibility into activity and performance. We could see clicks, leads, conversions, acquisition costs, return on ad spend and pipeline. That was progress, but it also changed how marketing talked about its value and, eventually, how organisations thought about the function.

What could be attributed became easier to defend. Activity with an immediate, measurable return became easier to fund. Marketing became very good at answering questions about what a lead cost, whether acquisition costs were rising, what media returned, whether conversion could be improved and whether a process could be automated. Agencies were pushed to deliver more efficiently, technology promised more productivity, and the growing supply of performance data made short-term results far easier to scrutinise than long-term market effects.

These are useful disciplines. The problem is the imbalance that develops when they dominate. The 2026 CMO Survey found that marketing leaders spend 68 per cent of their time managing the present and 32 per cent preparing for the future, a split the survey says has held every year since 2019. Under pressure from CEOs, boards and CFOs, 70.6 per cent said they were focusing more on short-term impact over long-run gains. The same research found that marketing spending decisions remain more reactive than strategic, shaped heavily by financial pressure and executive response.

This changes what marketing becomes inside an organisation. If most conversations with the C-suite concern cost, productivity, attribution and near-term return, marketing is gradually understood through those lenses.

A function that is mostly asked to demonstrate efficiency will eventually be managed as an efficiency function.

That is when the annual planning conversation becomes one about delivering the same outcome with a little less money.

When marketing looks like an expensive department

If marketing is understood mainly as the department that produces campaigns, buys media, generates leads, creates content and runs channels, it looks expensive. Salaries, agencies, technology, production, events, media and research all sit as visible costs in the P&L, and it is easy for someone to look at the total and ask what could be removed.

Gartner research from 2024 shows how widespread that perception is. In a survey of 395 CMOs, 47 per cent said marketing was viewed inside their organisation as an expense rather than a strategic investment. A separate Gartner survey of 378 senior marketing leaders the same year found that only 52 per cent believed they were successfully proving marketing’s value and receiving credit for its contribution to business outcomes. That is not only a communications problem for CMOs. It says something about how the role of marketing is understood inside many organisations.

The role has become too narrow. Marketing should help the business understand markets and customers, find growth, shape propositions, create and capture demand, build brands, support pricing power, increase penetration and make the company easier to choose. Sometimes it creates demand that sales can convert today. Sometimes it builds the conditions that make growth easier next year. And sometimes the most valuable thing a marketing team can do is recognise that communications are not the problem at all. The constraint may be the product, the price, distribution, the proposition or the customer experience.

This broader commercial orientation is what distinguishes the strongest marketing functions I have worked in and observed. They know where the company expects growth to come from and understand the market well enough to have a view on whether the ambition is realistic. They are close enough to customers to see emerging opportunities and commercially confident enough to challenge assumptions. They can make a case for investment without reducing every decision to a short-term ROI calculation, because the organisation understands that marketing is one of the ways the company creates growth.

There is research behind this. Gartner describes a group it calls “market-shaper CMOs”: leaders who use deep customer and market insight to identify unmet needs and influence business direction. In a Gartner survey of 125 CEOs and CFOs, companies where the CMO was effective at market shaping were 2.6 times more likely to exceed revenue and profit goals, and CMOs who were very effective market shapers were eight times more likely to exceed performance expectations than those who were not. The distinction Gartner draws is the useful part. These leaders do more than translate an existing enterprise strategy into marketing activity. They help shape where the business goes next.

AI is making the choice more urgent

For all the talk of AI transforming marketing, most of the conversation still begins with efficiency. How much time can we save, how much content can we automate, can we cut production costs, can a smaller team do more, can we reduce our dependence on agencies? These are legitimate questions. Marketing should eliminate unnecessary work and use technology to improve how it operates. Nobody benefits from protecting inefficient processes.

The risk is that efficiency becomes the limit of our ambition for AI. Gartner’s 2026 research found marketing leaders allocating an average of 15.3 per cent of their budgets to AI initiatives, with 70 per cent of CMOs calling AI leadership a critical goal, yet only 30 per cent reporting mature AI readiness. Gartner’s advice to marketing leaders is telling: reposition AI as a growth engine and move beyond viewing it only as an efficiency tool.

The productivity gains will be real. The more interesting opportunity sits beyond producing the same marketing for less money. AI can improve how marketers understand customers, interrogate data, spot changes in demand, explore segments, test propositions and learn from experiments. It can make research and insight more accessible, improve coordination between sales and marketing, and give marketers more capacity for the problems where judgement and commercial understanding matter.

There will be situations where the efficiency dividend should come out of the cost base. Businesses have to stay competitive and marketing has no special right to preserve waste. But there is no rule that every dollar or hour AI saves must disappear from marketing. That capacity can be reinvested in customer research, creative quality, brand building, experimentation, capability or market expansion, wherever it creates more commercial value.

If our biggest ambition for AI is making marketing cheaper, we should not be surprised when organisations conclude they need less marketing.

Efficiency matters, but it is not the strategy

Good marketing leaders should care deeply about efficiency. Money spent on ineffective activity should be redirected. Processes that consume time without improving outcomes should go. Technology nobody uses should go. Agencies should be accountable for the value they create. Every dollar has an opportunity cost, and treating marketing as a growth investment makes financial discipline more important, not less.

The danger is when efficiency becomes the organising idea. A business can become extremely efficient at activities that contribute very little to growth. If AI or process improvement releases 10 per cent of a marketing budget, there is a strategic decision to make about where that capital creates the most value. The company may urgently need the margin. Marketing may have a higher-return opportunity that has been underfunded for years. Money might move from production into media, from a saturated acquisition channel into retention, or toward the longer-term demand creation that keeps being sacrificed because every investment is judged against this quarter’s pipeline.

This is where the long-running evidence on advertising investment earns its place. Nielsen’s review of share-of-voice research summarises the established relationship between share of voice and market share: Binet and Field, analysing 171 campaigns, found that market share typically grew by around 0.5 percentage points for every 10 points of excess share of voice. Nielsen is careful about the nuances. The relationship varies by category, brand size, channel and competitive circumstance. That makes share of voice useful evidence for budget planning, and not another universal formula to apply without context.

The point is not that marketing should always spend more. Sometimes the right answer is to spend less. Sometimes there are better uses for the capital elsewhere in the business. A growth-oriented view of marketing does not mean protecting every budget line or assuming more investment produces more growth. It means making those decisions in the context of the commercial opportunity, rather than treating cost reduction as evidence of a better marketing strategy.

So how much should you spend?

This is the point where most articles about marketing budgets produce a number. There are useful reference points. Gartner reports an average of 7.8 per cent of revenue in its 2026 survey; the CMO Survey reports 9.0 per cent. Ritson’s 2022 work, drawing on Grace Kite’s analysis, suggests that spending between 5 and 10 per cent of turnover on advertising is a sound rule of thumb, with Ritson proposing 10 per cent as the practical starting point. All of these are valuable, provided we remember what they are.

I would want to know those numbers. I would want my team to know them. I would use benchmarks, share of voice, previous spend, unit economics, margins, market data and effectiveness research to challenge the plan. What I would not do is outsource the budget decision to any one of them.

The CMO Survey also splits its figure by company type: 7.0 per cent of revenue at B2B product companies and 10.1 per cent at B2B services companies. The full set of B2B marketing budget benchmarks is on the Back2Marketing blog.

The starting point is the growth ambition and the market situation. Where is growth expected to come from, and what is happening in the category? What needs to change among customers for the ambition to be realised? How strong is the brand, what are competitors doing, and where can marketing actually influence the outcome? What would need to be done well enough, for long enough, to give the strategy a credible chance?

Only then can we sensibly ask what that strategy costs. At that point last year’s spend becomes evidence rather than destiny. Benchmarks tell us whether our number is unusual. Share of voice tells us something about competitive investment. Unit economics and margins tell us what the business can sustain, and finance imposes the real constraints every strategy has to live within.

If the plan says marketing needs 15 per cent of revenue while comparable companies spend 7, the gap deserves scrutiny. The proposal may be extravagant. The benchmark may be inappropriate. The company may have set a growth ambition it cannot afford, or it may be facing years of accumulated underinvestment.

The purpose of the benchmark is to expose the question, not to answer it.

Five checks before the budget is set

The argument compresses into five checks, worth running before the first number goes on the slide:

1. Start with the growth ambition, not the envelope. Write down where the business expects growth to come from and how much, before anyone quotes last year’s spend.

2. Diagnose the market before you cost the plan. Name what has to change among customers, what competitors are doing and where marketing can actually influence the outcome.

3. Cost the strategy, then compare. Build the budget from what the plan requires, and only then set it against last year and the benchmarks.

4. Treat benchmarks and share of voice as pressure tests. A number that is unusual needs an explanation. It does not need to be changed until it has one.

5. Make the trade-offs visible. If the business cannot afford the plan, say which part of the ambition is being given up, rather than hiding the cut inside a flat budget.

If you cannot answer these before the number is set, the number is setting your strategy. If you can, the number becomes what it should have been all along: evidence.

What the budget says about marketing

There is a reason I care about this beyond the annual irritation of budget season. How a company sets its marketing budget tells you how it sees marketing. When the conversation begins with last year’s cost base, it moves naturally toward increases, reductions and efficiencies. When it begins with the growth ambition, the questions get broader and more commercially useful. Where does the company want to win? What would need to change for that to happen? What is stopping it now? Where can marketing make a real difference? What is the business prepared to invest, and what should it reasonably expect in return?

High-performing marketing functions should welcome that scrutiny. Marketing does not deserve investment simply because it is marketing, and a growth engine still has to produce growth. But the scrutiny has to work both ways. If a business wants aggressive growth while holding marketing investment flat, that assumption deserves to be examined. If it expects marketing to build future demand while measuring the function entirely on the current quarter, the consequences of that choice should be visible. If it wants AI transformation but frames the opportunity almost entirely around reducing headcount and production costs, it should be clear about what kind of marketing function it is creating.

Most importantly, if the budget has been decided before anyone has agreed the marketing strategy, we should be honest about what has happened. The investment envelope has constrained the strategy before the strategic choices were made. That is the marketing budget trap.

Back2Marketing exists to help marketers plan the other way round.

We believe marketing is a commercial discipline, not a promotional service. Its budget is a consequence of what the business needs marketing to achieve, not the starting point of the plan.

We believe benchmarks, last year’s spend and effectiveness research belong in the conversation as evidence, constraints and pressure tests. They are not substitutes for strategic thinking, and a precise-looking average is not a strategy.

Above all, we believe the sequence matters more than the number. A better planning process starts with where the business wants to grow, what needs to change in the market and what role marketing can realistically play. The numbers still matter, perhaps more than ever. They just come second.

The budget gets set before the strategy. That is the marketing budget trap, and it is a choice, not a law of finance.

Evidence informs. Judgement decides.

New essays land here first: subscribe and the next one will find you.

And if you want to build the judgement, not just the output, Back2Marketing’s B2B Marketing Fundamentals course teaches marketing from first principles.

Sources

• Gartner, CMO Spend in 2026: Redefining Marketing Investment Under Constraint, June 2026. Marketing budgets have plateaued since 2022; the 2026 average is 7.8 per cent of company revenue, 18 per cent below the mean allocation four years earlier; 73 per cent of CMOs describe growth expectations as high, very high or overly ambitious.

• Gartner, 2026 CMO Spend Survey, May 2026. 401 CMOs and marketing leaders surveyed January to March 2026; 56 per cent say their marketing organisation lacks the budget to deliver its 2026 strategy; AI at 15.3 per cent of marketing budgets; 70 per cent name AI leadership a critical goal; 30 per cent report mature AI readiness.

• The CMO Survey, 35th edition, 2026. 308 marketing leaders at US for-profit companies, 97 per cent at VP level or above; marketing budgets at 9.0 per cent of company revenue; 68 per cent of time managing the present against 32 per cent preparing for the future, every year since 2019; 70.6 per cent shifting toward short-term impact over long-run gains.

• Gartner, Only 52% of Senior Marketing Leaders Can Prove Marketing’s Value, press release, September 2024. Two surveys: 378 senior marketing leaders (April to May 2024), of whom 52 per cent say they successfully prove marketing’s value; and 395 CMOs (February to March 2024), of whom 47 per cent say marketing is viewed as an expense rather than a strategic investment.

• Gartner, Top Three Priorities for CMOs to Deliver Marketing Excellence in 2025, December 2024, and Market-Shaping CMOs Are Eight Times More Likely to Succeed in Their Role, May 2025. Both report one survey of 125 CEOs and CFOs conducted August to September 2024: companies where the CMO is effective at market shaping are 2.6 times more likely to exceed revenue and profit goals; 88 per cent of CMOs who are very effective market shapers exceed performance expectations, which Gartner reports as eight times more likely.

• Mark Ritson, Why Unilever is right to adopt zero-based budgeting, Marketing Week, January 2016. The marketing team researches, builds the plan and concludes it with a budget request and a promised return, rather than accepting an arbitrary percentage of projected sales.

• Mark Ritson, Triple-Cooked Marketing Budgets: The full recipe, Marketing Week, October 2022, and Marketing Week’s summary Ritson’s ‘foolproof’ system for marketing budgets, October 2022. Grace Kite’s analysis of three datasets (the ARC database, Paul Dyson’s budget optimisations and Nielsen’s 2022 ROI study) supports a rule of thumb of 5 to 10 per cent of turnover on advertising; Ritson proposes 10 per cent as the practical starting point.

• Nielsen, Need to know: What is share of voice?, 2025. Summarises Binet and Field’s analysis of 171 campaigns (1980 to 2010): market share typically gained 0.5 points for every 10 points of excess share of voice, with the relationship varying by category, brand size, channel and pricing power.

Every marketing advantage dies by becoming abundant. AI is the fastest cycle yet. It leaves judgement and taste scarce.

AI did not make marketing mediocre. Marketing was already mediocre. AI made mediocre marketing almost free.

That distinction decides what you do next. If AI caused the problem, the fix is a policy about AI. If AI removed the cost of a problem the industry already had, the fix is a harder conversation about what your marketing was ever good at.

Start with the figures. Eighty-seven per cent of B2B marketers say AI has made them more productive. Fifty-eight per cent say their content quality has improved. Fewer than two in five say their content is performing better. All three come from the Content Marketing Institute and MarketingProfs, surveying 1,015 B2B marketers in mid 2025. Quality is up, output is up, productivity is up, and results are flat.

This is not an essay about whether AI is making marketing worse. It is an essay about what happens to advantage when competence becomes free, and what that leaves for a marketer to be paid for.

The flat results are not a paradox. They are what happens when everyone gets better at the same time. Improvement that is universally available stops being improvement and becomes the new floor.

Advantage is not a property of a capability. It is a property of that capability’s scarcity.Advantage is not a property of a capability. It is a property of that capability’s scarcity.

Every marketing advantage dies the same way

Think about what has delivered real advantage in marketing over the last thirty years, and when each one stopped.

Having a website. Understanding how search worked. Buying media efficiently. Owning a marketing automation stack. Publishing consistently. Running paid social before your category noticed it existed.

Every one of those was a real edge. Every one is now hygiene. And each died the same death, which had nothing to do with the capability getting worse. It got easy.

The industry keeps missing this because it keeps mistaking the capability for the advantage, then acting surprised when the returns evaporate. Media buying threw off outsized returns while it required specialist knowledge and relationships. Programmatic made competent buying available to anyone with a budget, and the advantage moved on. Search threw off outsized returns while most companies did not understand it, and then everyone hired for it. Marketing automation was a moat until the moat came in a subscription. Nothing that can be installed is a moat.

Notice too that the half-life keeps shortening. Websites took years to travel from advantage to table stakes. Marketing automation moved faster. Content marketing faster again. Generative AI has compressed the cycle to a couple of years.

By the time the Content Marketing Institute fielded its survey in mid 2025, 89 per cent of the B2B marketers using AI were already using it to generate or optimise copy. A capability at that level of adoption is not an advantage. It is a baseline. If your team is still writing the business case for AI adoption, the thing you are building the case for stopped differentiating anybody some time ago.

This is the honest answer to “how do we get an edge from AI”, and it is not the one people want. You don’t, not from adoption. Adoption is the cost of staying in the game. The edge has already moved to whatever AI left scarce, and the rest of this essay is about finding it.

Adoption is now the cost of entry. The advantage is in what AI cannot make abundant.

The mediocrity came first

Before blaming the tools, look at what marketing was producing when the tools were expensive.

System1, working with LinkedIn’s B2B Institute, tested 1,600 B2B ads with around six million people worldwide over four years. Seventy-five per cent scored one star or less on System1’s emotional response measure, the level that contributes effectively nothing to long-term market share growth. Not one of the 1,600 reached five stars.

That was 2021. That was work with budget behind it, agencies attached and senior sign-off. It still landed as noise.

System1’s later research with eatbigfish and Peter Field, The Extraordinary Cost of Dull, put a figure on the flatness. In the UK, 60 per cent of responses to B2B television advertising registered as emotionally neutral, against 52 per cent for advertising overall. B2B is the duller category by measurement, not just by reputation. The same research, looking at US advertising generally, found that at the same level of excess share of voice non-dull advertising delivered around 1.3 points of annualised share gain while extremely dull advertising delivered around 0.1. An order of magnitude of growth, decided by whether the work registered at all.

Peter Field’s work for the IPA traced the same decline across almost 600 case studies. Creatively awarded campaigns were roughly twelve times as efficient as non-awarded campaigns in the twelve years to 2008. By 2018 that advantage had fallen below four times. Over the same period the share of awarded campaigns that were short term rose from around 5 per cent to almost 40 per cent.

None of that required a language model. The industry manufactured its own mediocrity through short-termism, measurement pressure and the steady substitution of activation for brand, which is the argument the previous essay in this series makes in full.

What generative AI changed is the price. Producing forgettable work used to cost a budget, a team and a quarter. Now it costs a prompt.

AI lowered the cost of being forgettable, and volume responded accordingly.

Average is now infinitely scalable

There is a mechanism behind the convergence, and it is better documented than most of the commentary suggests.

Anil Doshi at UCL and Oliver Hauser at the University of Exeter randomised 293 writers across three conditions: no AI, one AI-generated idea, and five. Six hundred independent evaluators assessed the results, published in Science Advances in July 2024. Individual work got better. Novelty rose 5.4 per cent with one AI idea and 8.1 per cent with five; usefulness rose 3.7 and 9.0 per cent. Collective similarity between the stories rose too, by 10.7 per cent of the measured range with a single AI idea.

Every writer improved. The body of work became more alike. For one marketer that reads as a win, because your draft beats the one you would have written alone. For a category it reads as convergence, because every competitor’s draft improved in the same direction at the same time.

There is a complication worth knowing. A University of Michigan study with more than 800 participants across 40 countries found the opposite effect: high AI exposure increased collective idea diversity. The designs differ in one respect. Doshi and Hauser gave people AI output to work from. Michigan used AI to surface ideas participants would not otherwise have reached. Substitution converges. Expansion diverges. Which one you get is a decision about process, not a property of the technology, and most teams have never made that decision consciously.

The volume side is now measurable. Graphite, sampling 55,400 URLs from Common Crawl, estimates that roughly half of newly published web articles are primarily AI-generated, and has been reporting a figure near that mark for five consecutive quarters. Its detectors have published error rates, so treat the exact percentage cautiously. The important finding is not whether AI has technically crossed 50 per cent. It is that human and AI-generated articles are now being produced in roughly equal volume.

What happens to that content is the more interesting finding. Across 31,493 keywords, 86 per cent of the articles ranking in Google results were human-written. Only 7 per cent of top-position results were AI-generated, half the baseline rate. Citations inside AI answer engines split roughly 82 per cent human to 18 per cent AI. Production share and visibility share have come apart.

Meanwhile the surface being competed for is shrinking. Pew Research Center tracked 68,879 Google searches from a panel of 900 US adults. When an AI summary appeared, 8 per cent of visits produced a click on a traditional result, against 15 per cent when no summary appeared. One per cent clicked a link inside the summary itself. Ahrefs, comparing 150,000 keywords with AI Overviews against 150,000 without, reported in February 2026 a 58 per cent average drop in click-through rate for top-ranking pages, up from 34.5 per cent in its study the previous April. SparkToro, using Similarweb clickstream data, put the share of US Google searches ending without any click at 68 per cent for the first four months of 2026.

More competent content, produced faster, competing for a materially smaller surface.

And the audience is not neutral about it. The Nuremberg Institute for Market Decisions ran controlled experiments alongside a 3,000-person survey across the US, UK and Germany. Identical advertising rated lower on emotional impact when labelled AI-generated than when labelled human-made, and only a quarter of respondents thought they could recognise AI content in the first place. Which makes disclosure the wrong thing to be anxious about.

Undifferentiated work fails on its own terms, long before anyone forms a view about how it was made.

Differentiation was never the mechanism

Here the industry’s own framing gets in the way, and it is worth clearing out.

Almost every article about B2B sameness quotes a statistic along the lines of “86 per cent of B2B buyers see no real difference between suppliers”, usually credited to CEB and Google. We went looking for the primary source. No study reports that figure. The nearest origin is a 2013 CEB and Google paper which found that only 14 per cent of business decision makers saw a real difference between suppliers and valued it enough to pay for it. Somewhere along the way the second half of the question was dropped and the remainder inverted, and the result has circulated through agency blogs and conference decks citing each other for a decade.

Here is what is documented. Gartner surveyed more than 1,100 B2B customers in late 2020 and found 64 per cent could not distinguish one brand’s digital experience from a competitor’s. Gartner also names perceptions of difference between supplier offerings as one of three drivers of buyer confidence, and its earlier work found that buyers overwhelmed by high-quality information were 153 per cent more likely to settle for a smaller, less ambitious purchase than they originally planned. The information was good. Volume alone was enough to shrink the deal.

Now the deeper point, which cuts against how most B2B teams think about this entirely. In a 2007 paper in the Australasian Marketing Journal, Jenni Romaniuk, Byron Sharp and Andrew Ehrenberg showed a low level of perceived differentiation between competing brands across many categories and two countries. Buyers largely did not see meaningful differences. They bought anyway. The authors’ conclusion was to stop putting perceived differentiation at the centre of brand strategy and put distinctiveness there instead: unique associations that make a brand easily identifiable.

If distinctiveness rather than differentiation is the engine, then “everything sounds the same” describes a memory failure, not a positioning one. The question is identifiability. Does anything about your work make it retrievable later?

Retrievable when is the part that gets skipped. John Dawes’s work at Ehrenberg-Bass with LinkedIn’s B2B Institute is the useful frame here, shorthanded as the 95:5 rule: up to 95 per cent of business buyers are not in market at any given moment. Dawes presents that as a heuristic rather than a measured constant, so use it as a way of thinking and not as a number in a board deck.

The measured version comes from 6sense’s B2B Buyer Experience Report for 2025, drawing on nearly 4,000 responses. Ninety-five per cent of winning vendors were already on the buyer’s day-one shortlist. Around four in five deals went to the vendor preferred before any seller contact. Buyers reached first contact roughly 61 per cent of the way through their process, earlier than the 69 per cent recorded the year before. Companion research put 85 per cent of buyers as having prior experience with the vendor they selected. That is a statement about winners rather than about your odds, so do not turn it into a conversion rate. What it establishes is enough: by the time you reach a competitive evaluation, most of the outcome is already sitting in the buyer’s memory of you.

Which is where Romaniuk’s category entry points come in. Buyers retrieve brands by situation: the problem, the moment, the trigger. Attach yourself to a handful of the situations that matter and you get recalled when one occurs. Publish competent, unmemorable content attached to nothing in particular and you do not.

The real cost of sameness lands before the comparison. You were never in the set to be compared.

What abundance leaves scarce

Back to the question the first section left open. Every capability that becomes abundant stops conferring advantage. So what has survived every one of these cycles without becoming abundant?

Two things, and they are not the same thing, though the industry uses them interchangeably.

Judgement determines what should be done. Taste determines how it should be expressed.

Judgement is the commercial call: which market, which problem, which trade-off, what to stop doing. The expressive call comes later and answers a different question, about what makes the work land, what to cut, and what deserves to exist at all. Neither has ever been available in a subscription. You cannot install taste, and no prompt can make the final judgement for you.

The evidence for the judgement premium is strong and getting stronger. Fabrizio Dell’Acqua, Ethan Mollick, Karim Lakhani and colleagues ran an experiment with 758 consultants at Boston Consulting Group, since published in Organization Science. On tasks inside AI’s capability frontier, consultants with AI access completed 12.2 per cent more tasks, 25.1 per cent faster, at higher quality. On tasks outside that frontier, consultants with AI access were 19 percentage points less likely to produce a correct solution than those working without it.

Same tool, same people. The variable was whether the task sat inside the model’s competence or outside it, and nobody hands you a map of where that line falls. Finding it is a judgement call, and the study prices what getting it wrong costs. AI does not reduce the need for judgement. It raises the price of not having any, because a confident wrong answer produced in four seconds travels a great deal further than a slow one.

The labour market is already repricing this. Stanford’s Digital Economy Lab, using ADP payroll data covering millions of US workers, finds employment among 22 to 25 year olds in the most AI-exposed occupations now sits 19 per cent below where it would be had it kept pace with less exposed peers. Experienced workers show no comparable gap. When this series covered that research in the second essay, the figure was 16 per cent. In the August 2026 version it is 19, and the researchers note the divergence has widened steadily since they first documented it. The finding underneath the headline is the one to carry: declines concentrate where AI substitutes for human tasks. Where it complements, employment is flat or rising, particularly for experienced workers.

Now taste, which the industry treats as a soft word for something with a very hard price tag.

Go back to the Cost of Dull research. In its UK B2B sample, of roughly £103 million in advertising spend, more than £53 million sat in the extremely dull quartile. More than half of the money went into the worst quarter of the work. That is not a budget problem. Every one of those campaigns was funded, approved and produced to a professional standard. Somebody signed it off. What was missing was the discernment to look at competent work and say it is not good enough to be remembered, and the standing to say it before the money went out the door.

Taste is expensive to lack and impossible to buy. Anu Atluru put it plainly in a 2024 essay: “In a world of scarcity, we treasure tools. In a world of abundance, we treasure taste.” Marketing is now firmly in the second world, and most of its processes were built for the first.

There is a complication the profession has barely started to face. Judgement and taste are built by doing the work: running the campaign, defending the budget, watching a decision fail and understanding why. David Duncan argued in Harvard Business Review in February 2026 that AI is absorbing the repetitive early-career tasks which historically produced professional discernment. Put that next to the Stanford data and the shape is uncomfortable. Fewer entry-level roles, and less formative work inside the roles that remain, squeezes the pipeline that produces the one capability the market is now paying a premium for.

The scarce asset is getting scarcer. That is good news for anyone who already has it, and a serious problem for the profession.

Five decisions worth making now

Diagnosis without a decision is commentary. Here is what changes if you accept the argument:

1. Stop looking for advantage in adoption. A capability your competitors can buy this quarter is not, by itself, a strategy. Ask of anything you are about to invest in: how long before this is available to everyone in my category, and what happens to my position then? If the answer is under two years, you are buying hygiene, and should budget for it as such rather than pitching it as an edge.

2. Measure retrieval, not output. Volume metrics reward the one thing AI made free. If your reporting is built on pieces published, keywords ranked and sessions delivered, you have instrumented the part of the system that no longer differentiates you, in an environment where the click surface is contracting. Add a memory-side measure: whether your brand is retrieved for the buying situations that matter, tracked over time. It moves slowly and is harder to defend in a monthly review than a traffic chart, so make the case before you need it, not during a budget cut.

   The same test applies to B2B thought leadership: judge it by whether buyers remember and trust you, not by how much you publish.

3. Decide your category entry points before you brief anything. Identify the buying situations you want to be retrieved for and keep the list short enough to be memorable. Filter each candidate on the three tests Romaniuk sets out: is the association credible for you, how much competition is there for it, and is it common enough among buyers to be worth owning. Everything you publish attaches to one of them or it does not get made. If your content calendar is currently a list of topics, converting it to a list of situations is the highest-leverage change available to you this quarter.

4. Move AI upstream of the draft. The homogenisation pattern comes from taking AI output as a starting point and polishing it. The diversity pattern comes from using AI to reach options you would not have reached alone. Same tools, different placement in the process. Use it to widen the option set, pressure-test a position, argue against your own case and find the evidence that contradicts you. Then reject the first plausible output as a matter of routine. This is slower per asset, and it only works if someone in the room has the standing to say no. Both of those are the point.

5. Protect the apprenticeship in your own team. If the Stanford pattern reaches marketing and entry-level work keeps thinning, an individual team develops a judgement gap long before the profession does. Hand juniors decisions rather than tasks. Let them own a call, defend it, and be wrong occasionally with support. The cost is short-term output. Weigh it against a team that can operate the tools and cannot tell when the tools are wrong.

If you cannot say which of these your team has already decided, adoption is still doing your strategy for you. If you can, you are competing on the two things AI has not made free.

What this comes down to

Every marketing advantage of the last thirty years died by becoming abundant. Websites, search, media buying, automation, content. AI is running the same cycle faster than any of them, and adoption crossed into hygiene while most teams were still building the business case.

Competent execution is no longer where sustainable advantage lives. It cannot be, because competent execution is getting cheaper and more widely available by the day.

What is left is the judgement to decide what deserves to exist, and the taste to make it worth remembering. Neither can be bought, installed, subscribed to or prompted. Both are built slowly, by people doing the work, which is why the supply is tightening at the exact moment the price is going up.

We believe marketing is a commercial discipline, not a promotional service. Its job is to be retrieved when a buyer’s situation arises, and no volume of competent content does that on its own.

We believe every capability that can be bought will be bought by everyone, and that advantage lives in what remains scarce afterwards: the judgement to choose and the taste to make it land.

Above all, we believe those two capabilities are built by doing the work, and that a profession which automates away its own apprenticeship will pay for it later.

Marketing was already mediocre. AI just made mediocre free. That is an uncomfortable place for the profession, and an extraordinary opportunity for anyone willing to build the thing that does not scale.

Back2Marketing exists to help marketers build that thing.

Evidence informs. Judgement decides.

New essays land here first: subscribe and the next one will find you.

And if you want to build the judgement, not just the output, Back2Marketing’s B2B Marketing Fundamentals course teaches marketing from first principles.

Sources

• Content Marketing Institute and MarketingProfs, B2B Content and Marketing Trends: Insights for 2026, October 2025. 1,015 B2B marketers surveyed June to August 2025. 87 per cent report improved productivity; 58 per cent improved content quality; 39 per cent improved content performance; among marketers using AI, 89 per cent use it to generate or optimise written copy. The productivity and performance figures are separate items and are not presented as a single computed gap.

• System1 and LinkedIn’s B2B Institute, research on B2B creative effectiveness, reported in Marketing Week, January 2021. 1,600 B2B ads tested with approximately six million people worldwide over four years; 75 per cent scored one star or less on System1’s measure; none reached five stars.

• System1, eatbigfish and Peter Field, The Extraordinary Cost of Dull, July 2024. 60 per cent of responses to UK B2B television advertising neutral against 52 per cent for UK television advertising overall; of roughly £103 million in UK B2B advertising spend analysed, more than £53 million sat in the extremely dull quartile; for US advertising generally, annualised share gain of around 1.3 points for non-dull advertising against around 0.1 for extremely dull at equivalent excess share of voice.

• Peter Field, The Crisis in Creative Effectiveness, IPA, June 2019. Almost 600 IPA case studies, 1996 to 2018. Creatively awarded campaigns around twelve times as efficient as non-awarded in the period to 2008, falling below four times by 2018; share of awarded campaigns that were short term rose from around 5 per cent to almost 40 per cent.

• Anil R. Doshi and Oliver P. Hauser, Generative AI enhances individual creativity but reduces the collective diversity of novel content, Science Advances, July 2024. 293 writers randomised across three conditions, 600 evaluators. Novelty up 5.4 per cent with one AI idea and 8.1 per cent with five; usefulness up 3.7 and 9.0 per cent; collective similarity up 10.7 per cent of the measured range with one AI idea.

• Joshua Ashkinaze, Julia Mendelsohn, Li Qiwei, Ceren Budak and Eric Gilbert (University of Michigan), How AI Ideas Affect the Creativity, Diversity, and Evolution of Human Ideas, ACM Collective Intelligence Conference, 2025. More than 800 participants across 40 countries; high AI exposure increased collective idea diversity without affecting individual creativity.

• Graphite, AI Now Writes as Many Online Articles as Humans Do, May 2026. 55,400 URLs randomly sampled from Common Crawl, classified by three detectors with published false-positive rates of 1.4 to 1.8 per cent and false-negative rates under 2 per cent. Primarily AI-generated articles at 50.9 per cent in Q4 2025 and 49.9 per cent in Q1 2026, near 50 per cent for five consecutive quarters. The essay reports comparable volumes rather than treating 50 per cent as a milestone.

• Graphite, AI Content in Search and LLMs, October 2025. 31,493 keywords; 86 per cent of ranking articles human-written; 7 per cent of position-one results AI-generated; approximately 82 per cent of AI answer-engine citations human-written. Correlation only.

• Ahrefs, AI Overviews Reduce Clicks by 58%, February 2026. 150,000 keywords with an AI Overview against 150,000 informational keywords without; 58 per cent lower average click-through rate for top-ranking pages as of December 2025, up from 34.5 per cent in the April 2025 study. Vendor research on its own data.

• Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results, July 2025. 68,879 searches tracked across a panel of 900 US adults, March 2025. 8 per cent click rate on traditional results with an AI summary present against 15 per cent without; 1 per cent on links inside the summary. General consumer search rather than B2B specific.

• SparkToro with Similarweb clickstream data, In 2026, Less Than One-Third of Google Searches Still Send a Click, June 2026. 68 per cent of US Google searches ended without a click, January to April 2026. General consumer search rather than B2B specific.

• Fabian Buder and Matthias Unfried, Nuremberg Institute for Market Decisions, Transparency Without Trust, NIM Insights, 2024. 3,000 respondents across the US, UK and Germany plus two controlled experiments. Identical advertising rated lower, especially on emotional aspects, when presented as AI-made; 25 per cent believed they could recognise AI-generated content, a confidence measure rather than a test of ability.

• Gartner, press release, May 2021. Survey of more than 1,100 B2B customers, December 2020. 64 per cent cannot tell the difference between one B2B brand’s digital experience and another’s; perceptions of difference between supplier offerings named as one of three statistically significant drivers of customer confidence.

• Gartner, press release, July 2019. Survey of more than 1,000 B2B customers. Customers facing conflicting high-quality information are 153 per cent more likely to settle for a course of action smaller and less disruptive than originally planned.

• CEB Marketing Leadership Council with Google and Motista, From Promotion to Emotion, 2013. Approximately 3,000 B2B purchasers, 36 brands. Only 14 per cent of business decision makers saw a real difference between suppliers and valued it enough to pay for it; the origin of the widely misquoted “86 per cent” statistic.

• Jenni Romaniuk, Byron Sharp and Andrew Ehrenberg, Evidence concerning the importance of perceived brand differentiation, Australasian Marketing Journal, 2007. Low perceived differentiation across competing brands with purchase continuing regardless; distinctiveness proposed in place of differentiation. Not conducted in B2B categories.

• Jenni Romaniuk, Ehrenberg-Bass Institute with LinkedIn’s B2B Institute, Category Entry Points in a B2B World, 2022. Buyers retrieve brands by buying situation; candidate entry points are screened on credibility, competition and commonness.

• John Dawes, Ehrenberg-Bass Institute with LinkedIn’s B2B Institute, The 95:5 Rule, 2021. Presented by the author as a heuristic rather than a measured constant.

• 6sense, B2B Buyer Experience Report 2025, nearly 4,000 responses, and In It (Multiple Times) To Win It, November 2025. 95 per cent of winning vendors on the day-one shortlist; the vendor ranked first wins about 80 per cent of the time; first contact at 61 per cent of the journey, down from 69 per cent; 85 per cent of buyers had prior experience with the vendor selected. These describe winners rather than conversion odds.

• Fabrizio Dell’Acqua, Edward McFowland III, Ethan Mollick, Hila Lifshitz-Assaf, Katherine C. Kellogg, Saran Rajendran, Lisa Krayer, François Candelon and Karim R. Lakhani, Navigating the Jagged Technological Frontier, Harvard Business School working paper 2023, published in Organization Science, 2026. 758 BCG consultants, AI access randomised. Inside the frontier, 12.2 per cent more tasks completed, 25.1 per cent faster; outside it, 19 percentage points less likely to produce a correct solution.

• Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, Stanford Digital Economy Lab, Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence, August 2026 version. ADP payroll data. Employment of 22 to 25 year olds in the most AI-exposed occupations 19 per cent below the counterfactual, up from 16 per cent in the November 2025 version; no comparable gap for experienced workers; declines concentrated where AI substitutes rather than complements. Covers AI-exposed occupations across the economy rather than marketing specifically.

• David S. Duncan, How Do Workers Develop Good Judgment in the AI Era?, Harvard Business Review, February 2026. Argument rather than original research.

• Anu Atluru, Taste is Eating Silicon Valley, Working Theorys, September 2024. Argument rather than research; quotation as given.

Digital marketing was one of the most important advances in the history of the profession. I spent years inside it, and I would not give any of it back.

It gave businesses direct access to their customers. It made testing faster and cheaper. It let a small company reach a market that used to need a national media budget. And it let marketers watch what customers did instead of guessing.

It also had a consequence nobody planned. As the channels became measurable, running the channels became the centre of how marketing was organised, judged and understood. Marketers became specialists in search, social, automation, analytics, content and conversion. The larger decisions, which customers to serve, what to offer them, at what price and through which route to market, moved to other parts of the business.

The result is a contradiction I have watched in rooms I was sitting in. Marketing has more data, more technology and more specialist skill than at any point in its history, and less say over the decisions that decide whether the business grows. Forty-seven per cent of CMOs say their company views marketing as an expense rather than a strategic investment, according to a Gartner survey of 395 CMOs in early 2024. Only 52 per cent of senior marketing leaders say they can prove marketing’s value and get credit for it, according to a separate Gartner survey of 378 leaders the same year.

This is not an essay against digital marketing. It is an essay about what we traded away while we were getting good at it, because AI is about to offer us the same trade again.

Digital gave marketing more to measure and less to decide.

In the first edition, Everyone Can Now Build. Few Know How to Market., I argued that AI is creating far more builders than marketers. In the last, AI Isn’t Replacing Marketers. It’s Changing What Marketers Are Paid to Do., I argued that execution is becoming cheap and judgement is what employers will keep paying for. This edition is about how that judgement left the marketing department in the first place. It was not taken from us. We gave it away one trade at a time, and each trade looked sensible when we made it.

Digital was a real advance

Search let a business appear at the exact moment a customer expressed a need. Websites let customers research, compare and buy on their own terms. Email gave every company a direct, cheap channel to the people who had already bought. Social platforms let customers talk back. And a campaign could be launched, watched and adjusted in days rather than months.

Digital also gave marketing a way to prove its worth. The profession had always struggled to connect its spending to the company’s results. Digital platforms supplied a chain of measures that appeared to link exposure, behaviour and revenue: impressions counted, clicks tracked, leads attributed, conversion paths observed, and the cost of winning a customer set beside the revenue that customer produced.

Many of those measures are still useful. A conversion rate tells you whether a landing page works. Cost per lead tells you which of two tactics is cheaper. The problem was never that marketing became measurable.

The problem was that measurable activity, clicks, leads and conversions, became the definition of marketing.

The channel replaced the discipline

“Digital marketing” originally described a set of channels: search, websites, email and social. Then it became a profession of its own, with its own roles, teams, agencies and qualifications. Marketing departments were split into paid search, paid social, content, lifecycle, automation, ecommerce, conversion optimisation and analytics.

Specialisation was the right response. A search specialist has to understand auction dynamics and keyword intent. An automation practitioner has to understand data architecture and journey logic. An analyst has to make sense of enormous volumes of behavioural data. None of that is learned in a month, and doing it badly wastes real money.

The cost was that no single team was responsible for the customer’s full experience of the company. A customer does not experience a business as a set of channel teams. They meet the product, the website, the salesperson, the service desk, the advertising and the price as one relationship, and each of those contacts affects whether they buy again. Inside the company, each of those contacts is owned by a different team with its own targets. The search team optimises search. The social team optimises engagement. The website team lifts conversion. The media team hits its campaign numbers. Every team does its job well, and nobody is responsible for the decisions that determine growth: which customers to prioritise, what the business offers them that a competitor does not, and how much of this year’s budget should go to buyers who will not purchase until next year.

When a marketing department is organised by channel, the channel becomes the starting point for every conversation. The planning meeting opens with what to do on social, in search, in email and on the website. It should open with the customer and the commercial problem.

Organise marketing by channel and the channels will decide the strategy.

The measurable replaced the valuable

Digital did not only multiply the number of measures. It changed how fast they arrived. Traditional marketing effects built slowly and were hard to isolate. Digital platforms reported within hours, in numbers that could sit on a dashboard and be discussed at the next performance meeting. That made digital activity far easier to manage, and it handed short-term measures a decisive advantage inside the company.

A click can be counted today. The effect of brand advertising on memory can be measured too, through brand tracking, brand lift studies and marketing mix modelling, but those take longer, cost more and arrive as estimates rather than counts. A lead can be entered in a system today. Familiarity among buyers who will not purchase for two years shows up in next year’s tracking study. When a manager has to show results by the end of the quarter, the click wins, because it is on the dashboard now and the brand tracker is not due until next year.

I have sat in the meeting where this happens. A modest brand investment is on the table beside a retargeting budget. The retargeting budget arrives with a chart showing cost per conversion to two decimal places. The brand investment arrives with a theory about buyers who are not in the market this year. The budget follows the chart. Everyone in the room behaves reasonably, and the brand budget shrinks again.

Les Binet and Peter Field documented this in The Long and the Short of It, using the IPA databank to separate short-term activation, which converts demand that already exists, from long-term brand building, which creates demand that does not exist yet. Their warning was specific: using very short-term online metrics as the main measure of performance damages long-term results. A campaign can have a strong click-through rate and change nobody’s behaviour. A channel can report a handsome return while taking credit for customers who would have bought anyway. A team can beat its lead target and create very little revenue.

The problem deepens when a measure stops being an observation and becomes a target. Once a metric is a target, people organise their work around moving it. That is the idea usually associated with Goodhart’s law. Reward a team for lead volume and it will find the audiences that produce cheap leads, whether or not they ever become customers. Let cost per acquisition run the budget and the money goes to the people easiest to convert, not the ones worth the most. Treat return on ad spend as the measure of effectiveness and the channels that capture existing intent will always beat the channels that created it.

A team measured on lead volume will learn to produce leads. It can hit its lead target every quarter for a year while revenue falls, because the leads it found were never going to buy. Nobody in that team has done anything wrong. They have done what the target asked.

A measurement framework is not an admin choice. It decides where the money goes.

Attribution counted the last click and called it the cause

Attribution was digital marketing’s biggest promise. For the first time, marketers appeared able to follow a customer from exposure to purchase and assign value to every step, as if marketing spend could be audited like a ledger.

The difficulty is that seeing an event before a purchase does not prove it caused the purchase. A customer who searches for a company by name, clicks the ad and buys may have been shaped by years of recommendations, brand advertising and product reputation. The final click is visible because it happened close to the sale, not because it created the demand.

Observational attribution methods “often fail to accurately measure the true effect of advertising”, even after controlling for extensive demographic and behavioural data. That is the conclusion of Brett Gordon and colleagues in Marketing Science in 2019, after comparing 15 large randomised advertising experiments on Facebook with the attribution methods the industry normally relies on. Randall Lewis and Justin Rao found something harder still, in the Quarterly Journal of Economics in 2015. Across 25 large field experiments with major US retailers and brokerages, most reaching millions of customers, individual purchasing was so volatile relative to the effect of advertising that the median confidence interval on return on investment was more than 100 percentage points wide. Even experiments at that scale could not pin the answer down.

None of this means marketing cannot be measured. It means the answer is rarely as simple as a platform dashboard makes it look. Les Binet has written that in the early years of digital “not every click translated to a genuine, incremental conversion”, and his advice is to combine attribution, experiments and marketing mix modelling, because attribution shows which touchpoints came before a sale, experiments show whether the advertising caused it, and mix modelling shows how spending across channels relates to sales over time. Digital gave us far more information about what customers did. We treated it as a full account of why.

The last click gets the credit because it is closest to the sale, not because it caused the sale.

Optimisation replaced diagnosis

Optimisation is one of digital marketing’s real strengths. Audiences can be refined, creative tested, friction removed and money moved toward whatever is working. But optimisation always starts with a system that already exists. It improves a chosen activity against a chosen objective. It cannot tell you whether the activity was the right one to choose.

A team can lift the conversion rate of a landing page without asking whether the offer is strong enough. It can cut the cost of a lead without asking whether the company is chasing the right market. It can improve email open rates without asking whether anyone wants the emails.

Diagnosis starts earlier. It looks at the market, the customer, the category, the product and the price before deciding what to do. Digital tilted companies toward optimisation because improvement could be watched every day: the dashboard showed what moved since yesterday, the platform recommended the next change, and the next test was always ready to run.

Diagnosis is slower, and it carries a risk optimisation never does. It may find that the problem does not belong to the marketing team at all. The offer may be no different from the competitor’s. The price may not match what customers think the product is worth. The product may be wrong for the segment. The company may simply be hard to buy from. Those findings cross into other departments and ask them to change decisions they have already made.

It is always easier to adjust the campaign than to reopen a decision the company has already made.

Capture replaced creation

Digital channels are very good at finding people who already want to buy. Search catches a customer who is looking. Retargeting reaches someone who has already visited. Automation nurtures a known prospect. Review sites influence a buyer who is comparing options. All of that is worth doing, because a business should make it easy for an interested customer to find it, understand it and choose it. The problem starts when capturing demand is mistaken for the whole of marketing.

Up to 95 per cent of business buyers are not in the market for a category at any given time, according to John Dawes’s research with the Ehrenberg-Bass Institute and the LinkedIn B2B Institute. Dawes presents the ratio as a rule of thumb rather than a law, and the point holds across most categories: most of your future customers cannot be converted today, however precisely a platform can target them, because they already have a supplier, or have no need this quarter, or are under contract without budget. What marketing can do is shape what they remember when the need arrives. Dawes argues that advertising works mainly by creating and refreshing the memories a buyer draws on when the buying situation comes around.

Digital performance systems favour the people closest to purchase, because their behaviour is easy to see and their response arrives fast. So the money accumulates around search, retargeting, lead generation and conversion, because those activities can show they are close to revenue.

Being close to the revenue is not the same as having created it.

Capture and creation are both necessary. The danger is that one of them looks more accountable because its effects are easier to count.

Promotion replaced marketing

The narrowing of marketing was never only a measurement problem. Over time it changed where marketing sat in the company.

Marketing in its full sense covers the creation, communication, delivery and exchange of value. That is the American Marketing Association’s definition, and communication is one of its four verbs. Philip Kotler’s planning sequence in Marketing Management starts with research and moves through segmentation, targeting and positioning before it reaches the marketing mix and implementation. The discipline was designed as a series of decisions about the market and the offer, with execution near the end.

In many companies today, marketing enters near the end instead. Product decides what gets built. Leadership picks the growth market. Finance sets the price. Sales chooses the priority accounts. Customer experience sits in a function of its own. Marketing receives the finished offer and is asked to create awareness, generate leads and speed up revenue. It was not in the room when the product, the price or the target segment was decided, and running the campaigns well does not earn it a place in that room.

How marketing leaders spend their time points the same way. Marketing leaders spend 68 per cent of their time managing the present and 32 per cent preparing for the future, a split that has held every year since 2019, according to the 2026 CMO Survey of 308 US marketing leaders.

When growth falls short in a company organised this way, marketing is asked to increase demand for an offer it did not help shape, among customers it did not choose, at a price it did not set. Better communication sometimes helps. More often the problem is the offer, the price or the segment, and no campaign can fix those.

A marketing team that only enters after the offer, the price and the segment are fixed is a promotions department, whatever it is called.

The problem was never digital

The purpose of marketing did not change when customers moved online. A business still has to understand its market, decide which customers to serve, build an offer those customers value, price it, distribute it and make it easy to buy. It still has to create demand among people who are not ready to buy today, which means building familiarity and trust over years, while being ready to convert that demand when it arrives.

Digital tools can serve every part of that work. Search data shows how customers describe their own problems. Website behaviour shows where they get stuck. Customer systems show how a deal actually moves. Online communities show what a category argues about. Experiments test an offer before serious money is spent. But a dashboard cannot decide which market to enter. A platform cannot decide what the company should stand for. An attribution model cannot decide how much of this year’s revenue to trade for next year’s.

Responsibility for the split between the tools and the decisions is widely shared. Companies promoted the people who could show a number every month. Quarterly targets made short-term response more attractive. Org charts broke responsibility for the customer into pieces. Technology companies sold measurement systems that made their own platforms look accountable. And marketers played a part. It was easier to build a career on channel expertise, which showed up on a dashboard every month, than on commercial judgement, which had to be argued for.

The industry also invented a false split between “traditional” and “digital” marketing, as if customer value, positioning and brand building belonged to an earlier era and data, automation and performance were the future. Customers never experienced that split. They met advertising, recommendations, salespeople, websites, products and service as one relationship. The answer is not to go back to how marketing worked before the internet.

The answer is to stop treating digital marketing as a substitute for marketing.

Start with a diagnosis of the growth problem before choosing a channel. Design measurement around the commercial objective rather than inheriting it from the platform, and keep the metrics that manage activity separate from the measures that show whether the business grew. And put marketing back into the decisions about the offer, the price and the segment before the company commits to them and asks a campaign to rescue them.

Specialist skills stay valuable. They are worth more, not less, when the specialist understands the market their channel operates in. The strongest search strategist understands demand beyond search. The strongest automation practitioner understands customers beyond the database. The strongest performance marketer understands that most of the effect of advertising arrives after the conversion window closes.

Before your next campaign plan

Run these before the brief is written.

1. Diagnose before you optimise. Name the growth constraint, whether awareness, relevance, distinctiveness, availability, price, product or conversion, before choosing what to do about it. A campaign without a diagnosis is a guess.

2. Design measurement around the objective, not the platform. Decide what business outcome the activity has to change, then choose measures that can detect it.

3. Separate activity metrics from impact measures. Clicks and cost per lead help manage execution. Never report them as evidence that the business grew.

4. Write down the split between capture and creation. Decide how much goes to converting today’s demand and how much to building tomorrow’s, and defend it as a choice rather than inheriting it from last year’s dashboard.

5. Say when the problem is upstream. If the offer, the price or the segment is the real constraint, say so before promotion is asked to make up for it.

A brief that cannot answer these is asking a campaign to fix a problem nobody has named. A brief that can is a marketing plan, not a media plan.

AI makes the same mistake easier to repeat

The digital era taught marketers to equate progress with more speed, more data and tighter optimisation. AI now offers more of each. It can produce more content, analyse more data, automate more decisions and let a small team put out the work of a large one.

That raises the cost of acting without a marketing foundation. A weak offer can now be promoted at a scale nobody could afford before. An undifferentiated position can generate hundreds of content variations. A badly chosen metric can be optimised every hour. A broken customer experience can be automated end to end. Left undirected, AI will speed up every habit that narrowed marketing in the digital era, and the company will get more efficient at producing and measuring activity while staying unclear about whether any of it produces growth.

The lesson is not to slow the technology down. It is to put the decisions back before the tools. Digital was the first time the profession had more to measure than to decide. AI is the second, and this time we know how the trade ends.

We believe marketing is a commercial discipline, not a promotional service. Its job was never to promote whatever the business decided to build. It was to help the business decide what to build, for whom, at what price, and how that turns into growth that lasts.

We believe measurement should serve judgement, not replace it. A dashboard can tell you what moved. It cannot tell you what mattered.

Above all, we believe AI hands marketers the same choice digital did, faster. We can use it to produce more of the same at greater speed, or we can use it to take the execution load off the profession and put the time back into customers, markets and strategy.

Digital made marketing measurable. It also reduced it to tactics. Putting marketing back into the decisions about customers, offers and price does not mean giving up digital tools. It means using them to carry out a plan that marketing helped write.

Back2Marketing exists to help marketers get back into the decisions about customers, offers and price.

Evidence informs. Judgement decides.

New essays land here first: subscribe and the next one will find you.

And if you want to build the judgement, not just the output, Back2Marketing’s B2B Marketing Fundamentals course teaches marketing from first principles.

Sources

• American Marketing Association, Definitions of Marketing, on marketing as the creation, communication, delivery and exchange of value.

• Philip Kotler, Marketing Management (planning framework), on research, segmentation, targeting, positioning and the value proposition preceding implementation.

• Les Binet and Peter Field, The Long and the Short of It, IPA, 2013, on short-term activation versus long-term brand building, and the risk of using very short-term online metrics as primary performance measures.

• Les Binet, Unlocking Marketing Effectiveness in the Digital Age, Think with Google, March 2025, on early click-based misattribution (“not every click translated to a genuine, incremental conversion”) and combining marketing mix modelling, experiments and attribution.

• Brett Gordon, Florian Zettelmeyer, Neha Bhargava and Dan Chapsky, A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook, Marketing Science, vol. 38, no. 2, 2019. 15 US advertising experiments; observational methods “often fail to accurately measure the true effect of advertising” even after conditioning on extensive demographic and behavioural variables.

• Randall Lewis and Justin Rao, The Unfavorable Economics of Measuring the Returns to Advertising, Quarterly Journal of Economics, vol. 130, no. 4, 2015. 25 large field experiments with major US retailers and brokerages, most reaching millions of customers; median confidence interval on return on investment more than 100 percentage points wide.

• John Dawes, Ehrenberg-Bass Institute for Marketing Science with the LinkedIn B2B Institute, The 95:5 Rule, 2021, on out-of-market buyers and advertising’s role in building brand-relevant memories; the ratio presented as a heuristic.

• Charles Goodhart, Problems of Monetary Management: The UK Experience, 1975, origin of the principle that a measure loses its usefulness once it becomes a target.

The question I am asked most often about artificial intelligence is whether it will replace marketers.

Every earlier wave of marketing technology gave us new ways to do the work or new ways to measure it. AI does the work. That is a different kind of change, and it is why so many capable people are looking at their own role and asking how much of it a machine could now do.

This is not an article about whether marketers will be replaced. It is an article about what employers will keep paying marketers for once a machine can produce most of what marketers used to make.

The short version: the money, the headcount and the promotions are moving away from the people who produce the work and toward the people who decide what work is worth producing, where the budget goes and what is good enough to run.

Execution is becoming abundant. Judgement is becoming scarce.

Sixty-five per cent of CMOs expect AI to dramatically change the CMO role within two years, according to a Gartner survey of 402 senior marketing leaders published in February 2026. In the same survey, only 32 per cent said significant changes are needed to the CMO skill set. Two thirds can see the change coming. Fewer than a third think it will change what they need to be good at.

Very few marketers will be replaced outright. Many will find that the part of the job that used to define their value is becoming the part a machine does best. To see where the pay goes next, it helps to see how it got to where it is.

Three shifts pushed marketing toward execution. AI pushes the other way.

Before digital, marketing was a generalist craft. One person or one small team understood the customer, shaped the message, chose where it ran and judged whether it worked. The tools barely changed, so what made a marketer valuable was judgement about customers and messages, exercised with patchy data and a slow feedback loop.

Digital broke that. Search, then social, then mobile, then video, each with its own mechanics, metrics and learning curve. No one person could hold it all, so organisations specialised. The generalist gave way to the paid search specialist, the SEO lead, the automation manager and the social team. That was the right response. When running one channel well takes real expertise and running it badly wastes real money, you hire someone who does only that.

Attribution changed what marketers were believed for. Once activity could be tracked, the monthly argument about whether the campaign worked became a dashboard. That bought marketing a credibility it had never had, and I would not give it back. It also meant the channels that could prove their contribution got the budget, the headcount and the ambitious people, while the work that compounds slowly, brand and memory, was starved.

Then the technology stack became a job in itself. Scott Brinker’s marketing technology landscape held roughly 150 tools in 2011 and 14,106 in 2024, nearly a hundredfold increase in thirteen years, and his 2026 edition reported almost no growth, which suggests the stack has peaked. Every one of those tools needed someone to own it, integrate it and justify its licence fee.

Each of those three shifts moved pay and status closer to execution. Being good at marketing came to mean being good at running one discipline inside it. Marketing teams were built for a world where producing the work was expensive, slow and scarce.

AI removes exactly that constraint, and it is the first shift in thirty years that moves pay in the other direction.

A marketing team built around the cost of producing work will not survive that cost falling to nearly nothing.

The cost of producing marketing work has collapsed

Professionals with access to an AI assistant finished business writing tasks 40 per cent faster and produced work rated 18 per cent higher in quality, in a 2023 experiment with 453 people that Shakked Noy and Whitney Zhang published in Science. In a larger field experiment with Boston Consulting Group, 758 consultants using a frontier model completed 12.2 per cent more tasks, 25.1 per cent faster, at quality rated more than 40 per cent higher than the control group, when the task sat within the technology’s capabilities.

Anyone running a marketing team in the last two years has seen the same thing. The first draft, the variant set, the campaign summary and the first cut of a segment now come back in an afternoon rather than a fortnight.

The number that matters most for your career sits further down the consulting study. The bottom half of performers improved by 43 per cent. The top half improved by 17 per cent. AI is not mainly a tool that makes the best people better. It makes average execution hard to tell apart from good execution. The gap that justified a premium for competent production is closing.

When something becomes abundant, the market stops paying a premium for it. That is not a comment on whether it matters. It is a comment on what commands a price. Until now the scarce thing in marketing was the ability to produce and run the work. AI has made that abundant. What stays scarce is the ability to decide which work is worth making.

When producing the work is cheap, deciding what to produce is what gets paid for.

The labour market is already paying for judgement and not for execution

If that is right, the effect should already be visible in employment data. It is.

Employment of workers aged 22 to 25 in the occupations most exposed to AI is roughly 19 per cent below where it would be had it kept pace with less-exposed peers, according to the August 2026 version of a Stanford Digital Economy Lab paper by Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, based on ADP payroll records. Experienced workers in the same occupations show no comparable gap. The paper finds the same pattern among marketing and sales managers.

The damage is concentrated where the job consists of executing well-defined tasks. That is what entry-level roles were built around, and it is what the models now do cheaply. Where the job consists of decisions built on experience, employment is holding.

Indeed’s Hiring Lab reached a similar place from a different direction. Of roughly 2,900 skills found in US job postings, 46 per cent of the skills in a typical posting can be partly or fully transformed by generative AI, and fewer than 1 per cent can be fully performed by it. Indeed describes software developers as shifting from doing the work to directing it. The same shift is under way in marketing.

For a decade a marketer’s pay sat close to the work itself: producing assets, running campaigns, optimising channels. That is the layer AI is absorbing fastest. The pay is moving to the decisions above it: which customers to go after, what to be remembered for, how to split the budget, what to stop doing, and whether any of the activity was worth the money.

Employers are not paying less for marketers. They are paying less for marketers whose job is production.

Cheap execution without expertise makes teams confidently wrong

None of this is an argument against expertise, and it is not “specialists out, generalists in”. A world where execution is cheap is a world with more execution in it, and more need for people who know how a discipline behaves, where it breaks and what good looks like when a machine produces it at volume.

The consulting experiment shows why. The researchers included a task that looked like something AI would handle well but sat just outside its real capability. Consultants without AI got it right 84.5 per cent of the time. Consultants with AI got it right 60 to 70 per cent of the time. The tool made them faster, more confident and wrong more often, because they stopped checking the output. The study’s lead author has elsewhere called the pattern falling asleep at the wheel.

Every marketing team using AI now has a version of that task. The positioning draft that reads well and misstates the one thing the product does differently. The segment analysis that is fluent and built on the wrong customer definition. The campaign that is on brief and says nothing anyone will remember. Someone has to be able to look at plausible output and catch the line that is wrong. That person is not a luxury. Without them the team produces errors faster.

Cheap execution without an expert to check it does not add value. It subtracts value, at scale.

The scarce skill is not generalism. It is knowing enough about each discipline to judge its output, and then deciding how the disciplines fit together: how paid, owned and earned combine, how brand and performance trade off, how this quarter’s number is weighed against next year’s growth.

The job that execution crowded out

Picture a marketing team you will probably recognise. The paid media is well run. The content calendar ships on time. The automation flows fire. The SEO trend line points the right way. Every specialist is hitting their number, every dashboard is green, and the company still misses its revenue target.

In the post-mortem nobody can explain why, because the explanation does not live in any single channel. Demand was captured efficiently but never created. Sales told buyers one story and the campaigns told them another. The budget followed last quarter’s attribution instead of next year’s buyers. No individual was wrong. The plan as a whole was.

That gap, between well-run parts and a plan that produces revenue, is the job marketing was always supposed to do. In practice it means four things. Understanding who buys, why, and what they weigh when they decide. Choosing where to compete and what to be remembered for. Splitting the budget between building memory among buyers not yet in the market and converting the few who are. Making sure sales, product and finance are working to the same view of where next year’s revenue comes from.

For as long as most of us have worked, nobody could staff that job properly, because producing the work consumed everyone. The channels had to be fed. The stack had to be run. The reporting had to be built. The deciding happened in the gaps, in the weekly meeting, in the planning cycle, if there was time. There was never time.

AI opens that time up. It can now support most of the production: the research summary, the positioning draft, the campaign assets, the optimisation, the sales enablement. What it cannot do is decide where to compete, how to position, what to prioritise and how to weigh this quarter against the long term. Someone still has to make those calls and be answerable for the revenue number rather than a channel number.

The buyer’s side makes that harder to avoid. The average B2B purchase now involves around 13 internal stakeholders and nine external influencers, according to Forrester’s 2026 State of Business Buying research, and buyers increasingly begin their research inside AI tools and then check what they are told against peers and communities they already trust. The vendors that survive that check are the ones that built memory before the purchase began. A marketing function organised as separate channel teams cannot serve a buying process that behaves as one decision.

The money makes it harder to avoid too. AI already takes 15.3 per cent of the average marketing budget, and 56 per cent of CMOs say they lack the budget to deliver their 2026 strategy, according to Gartner’s 2026 CMO Spend Survey of 401 marketing leaders. Someone has to decide what the AI spend displaces and how its contribution will be judged. In the same survey, 70 per cent of CMOs named becoming an AI leader as a critical goal and only 30 per cent said their organisation’s AI readiness was mature. That gap is a decision nobody owns.

The job AI hands back to marketers is the one execution never left time for: deciding where the money goes and being answerable for what it returns.

The decisions that do not automate

Take the most familiar decision in demand generation: how much to invest in creating future demand against capturing the demand that exists today.

The evidence here is as good as marketing evidence gets. Up to 95 per cent of business buyers are not in the market for a category at any given time, according to John Dawes’s research with the Ehrenberg-Bass Institute and the LinkedIn B2B Institute, which is why most marketing works, if it works at all, by building memory for a purchase that has not started. Les Binet and Peter Field, analysing hundreds of campaigns in the IPA databank, showed the long-term cost of over-investing in short-term activation and proposed a rough 60:40 balance of brand to activation, with their B2B work pointing nearer an even split. Treat both as rules of thumb rather than constants.

That evidence tells you what tends to work, on average, across categories. It cannot tell you how to weight brand against activation in your market, for your category, against your targets, this year. The moment that question stops being academic arrives near the end of a hard quarter, when someone senior points at the brand line of the budget and asks what it did for pipeline this month. Every marketer who has held a budget knows that moment. The research is on your side, and the research is not enough, because the right answer depends on your cash position, your board’s patience, your competitor’s noise and your own conviction. An AI assistant can recite Binet and Field. It cannot tell you whether your business can survive the eighteen months brand investment takes to pay back.

Evidence informs. Judgement decides.

The brand-versus-activation call comes up once a year, at budget time. The next decision comes up every day. A piece of work is on brief, error-free and finished. Is it good enough to run? When anyone can generate a competent asset in seconds, competent stops being an advantage. Your competitors have the same tools and get the same output. Plenty of machine-drafted work this year has been fluent, on brief and forgettable. The willingness to say “not yet” to work that is merely fine, and to know what would make it better, is becoming one of the few durable advantages a marketer has. Call it taste. It is not a soft skill. It is the quality control on everything the machine produces.

Learn the tools last

The obvious response is to go and learn the tools. But tools change every few months, and fluency with this quarter’s features is the fastest-depreciating asset you can hold. The durable investment is in the things that make your judgement worth paying for. Check where you stand on each of them before you book another tool course.

1. When did you last hear a customer explain a buying decision in their own words? AI can summarise the research you give it. It cannot sit in a sales call and hear the objection underneath the stated one. Marketers who keep doing primary work hold an advantage over everyone whose understanding comes from prompting the same models as their competitors.

2. Could you explain your company’s margins, customer lifetime value and cost to acquire a customer without looking them up? The marketers who allocate budgets and defend them speak in commercial terms. That conversation is getting harder as boards ask what the AI spend is returning.

3. Could you explain, from evidence, why brand investment matters when most buyers are not in market? Ehrenberg-Bass on how brands grow, Binet and Field on effectiveness, behavioural science on how buyers decide. This knowledge is the raw material judgement is made from, and it is what lets you evaluate machine output rather than accept it.

4. Do sales, product, finance and marketing agree on where next year’s revenue comes from? If the answer is no, that is the job. Getting four functions to one view is work no model can do, and the thirteen-person buying committee on the outside is a fair picture of the coalition you need on the inside.

5. Do you know what the current models do well and where they fail? Working knowledge of the tools is now expected. The point is to sequence it. AI fluency multiplies the value of market knowledge, commercial literacy and evidence, and multiplies nothing when they are absent.

If you can answer the first four, the fifth will make you more valuable than you have ever been. If you can only answer the fifth, you are learning to operate the thing that is replacing you.

Let’s be clear: this shift is not painless and it is not automatic. Marketers whose identity is built on executional craft are the most exposed, and the employment data says the squeeze is already real at entry level. Moving from execution to decisions is a change in skill, not a change in title. Not everyone will make it. But the direction is clear enough to act on now, while acting is still a choice.

The tasks are going. The job is not.

Will AI replace marketers? It will keep removing tasks, and some of those tasks have defined whole jobs. That is real, and anyone whose job is made of those tasks is right to worry.

But a job is not a list of tasks. Marketing has always been about understanding a market, choosing customers, building memory and preference among them, and being answerable for the growth that results. Digital, attribution and the stack each buried that purpose a little deeper under the mechanics of producing the work. AI is the first shift that clears the mechanics away. What remains is what was always underneath: understanding markets, making sound decisions under uncertainty, and getting an organisation to act on them.

For an experienced marketer, the practical response is to move your time from producing the work to the decisions above it, and to make that move visible: own a revenue number, not a channel number, and be the person in the room who can say why the budget is split the way it is.

For a junior marketer the answer is harder, because the production tasks that used to teach the job are the tasks going first. The judgement this essay describes was learned by doing a thousand small things and seeing which ones worked. Juniors will have to learn it another way: sitting in on customer calls, reading the sales pipeline, seeing the budget decisions being made, and being asked for a view earlier than any of us were. Leaders have to arrange that on purpose. It will not happen by itself, and if it does not happen the next generation of judgement never gets built.

We believe marketing is a commercial discipline, not a promotional service. Its value lies in the decisions it makes about markets, customers and money, not in any single channel or task.

We believe AI should raise the contribution marketers make to a business, not shrink it. It should absorb the production that crowded out the real work and hand the time back to understanding customers, making decisions and getting the organisation to act on them.

Above all, we believe judgement and taste become more valuable as execution becomes cheaper. Tools will change. The frontier will move. The need to decide what is worth doing, and to know good from merely fine, will not.

AI is not replacing marketers. It is changing what marketers are paid to do.

Back2Marketing exists to help marketers do that job well.

Evidence informs. Judgement decides.

New essays land here first: subscribe and the next one will find you.

And if you want to build the judgement, not just the output, Back2Marketing's B2B Marketing Fundamentals course teaches marketing from first principles.

Subscribe now

Sources

• Gartner, Survey Reveals CMO “AI Blind Spot”, press release, February 2026. 402 senior marketing leaders in North America and Europe surveyed August to October 2025; 65 per cent say advances in AI will dramatically change the role of the CMO in the next two years; 32 per cent say significant changes are needed to the CMO profile and skill set.

• Gartner, 2026 CMO Spend Survey, published May 2026. 401 CMOs and marketing leaders in North America, the UK and Europe surveyed January to March 2026; AI at 15.3 per cent of marketing budgets; 70 per cent say becoming an AI leader is a critical goal for 2026; 30 per cent report mature AI readiness; 56 per cent say their organisation lacks the budget to deliver its 2026 strategy.

• Scott Brinker, chiefmartec, Marketing Technology Landscape Supergraphics, 2011 to 2026. Roughly 150 tools in 2011; 14,106 in 2024; 15,384 in 2025; the 2026 edition reported near-zero growth.

• Shakked Noy and Whitney Zhang, Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence, Science, 2023. 453 professionals; 40 per cent faster completion and 18 per cent higher rated quality on business writing tasks.

• Fabrizio Dell’Acqua et al., Navigating the Jagged Technological Frontier, Harvard Business School working paper with Boston Consulting Group, 2023; published in Organization Science, 2026. 758 consultants; 12.2 per cent more tasks, 25.1 per cent faster, quality more than 40 per cent higher inside the frontier; bottom-half performers improved 43 per cent against 17 per cent for the top half; on the outside-frontier task the control group was correct 84.5 per cent of the time against 60 to 70 per cent with AI. “Falling asleep at the wheel” is the title of Dell’Acqua’s separate 2022 study of HR recruiters.

• Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence, Stanford Digital Economy Lab, August 2026 version. ADP payroll data; employment of workers aged 22 to 25 in the most AI-exposed occupations roughly 19 per cent below where it would be had it kept pace with less-exposed peers (earlier versions reported 13 and 16 per cent); no comparable gap for experienced workers; marketing and sales managers, in the fourth exposure quintile, show the same pattern.

• Indeed Hiring Lab, AI at Work Report 2025: How GenAI is Rewiring the DNA of Jobs, September 2025. Analysis of roughly 2,900 skills found in US job postings; 46 per cent of skills in a typical posting subject to hybrid or full transformation; fewer than 1 per cent fully performable by GenAI today; software developers described as shifting from “doing the work” to “directing the work”.

• Forrester, The State of Business Buying, 2026, January 2026. Typical buying decision includes 13 internal stakeholders and nine external influencers; buyers increasingly start research in AI search tools and validate with trusted human sources.

• John Dawes, Ehrenberg-Bass Institute for Marketing Science with the LinkedIn B2B Institute, The 95:5 Rule, 2021. At any given time, up to 95 per cent of business buyers are not in the market for a given category; treated as a heuristic rather than a constant.

• Les Binet and Peter Field, The Long and the Short of It, IPA, 2013. IPA Effectiveness Databank; roughly 60:40 brand-to-activation balance; reliance on short-term measures undermines long-term effectiveness.

• Les Binet and Peter Field with the LinkedIn B2B Institute, The 5 Principles of Growth in B2B Marketing, 2019. B2B balance indicated nearer an even split, labelled tentative by the authors.

Every generation of marketers is convinced it is living through the transformation of the profession. Ask a marketer in the 1950s and the answer would have been television. Ask in the sixties and it was the database, which turned mass broadcasting into targeting and direct response. Ask in the mid-nineties and it was the internet, which rewrote distribution and shifted power from brands to consumers. Since then it has been search, then social, then the smartphone that put all of it in our pockets.

Now it is artificial intelligence.

This is not another article about AI. It is an article about what becomes more important when the ability to create is no longer the constraint.

To be fair, the current wave deserves the attention it is getting. Generative AI is not simply another channel or software upgrade. It changes who can create, how quickly ideas can be developed and how much a small team, or one person, can accomplish.

A founder can describe an application in plain language and have a working prototype without first assembling a development team. A subject-matter expert can build a website, create a course, publish a newsletter and produce credible marketing materials with tools that were out of reach a few years ago. Marketers can lean on AI for research, analysis, writing, coding, design and the administrative grind.

The productivity evidence is real. In an experiment involving professional writing tasks, Shakked Noy and Whitney Zhang found that people using generative AI finished 40 per cent faster, and the quality of their output improved by 18 per cent. A separate field experiment with 758 Boston Consulting Group consultants found that AI users completed more tasks, worked faster and produced higher-quality work when the task sat within the technology’s capabilities.

These are not marginal gains. Capabilities once confined to specialists or large organisations are becoming available to almost anyone.

The barrier to building is falling

For most of modern business history, turning an idea into a product took serious resources. Software needed engineers. A credible brand needed designers and agencies. Manufacturing needed capital and suppliers. Reaching customers at scale needed distribution relationships or a large media budget.

Those barriers have not vanished. There is still a wide gap between a prototype and a secure, differentiated, commercially sustainable product. Expertise still matters. Quality still matters. Capital still matters.

What has changed is the starting point.

More people can now get from an idea to something tangible. The time and cost of testing a concept are falling, and work that once needed a team can increasingly be attempted by an individual. In one controlled GitHub experiment, developers using an AI coding assistant completed a programming task 55 per cent faster than those working without it. The technology did not remove the need for engineering skill, testing or judgement. It reduced the effort of turning an instruction into working code.

As the cost of creating something falls, more things get created. Markets will fill with applications, services, courses, newsletters, consultancies, products and content. Plenty will be competent. Some will be excellent. A good deal will be noise.

The supply of attention, however, does not increase simply because the supply of things competing for it has multiplied.

Lower barriers to production do not remove the commercial challenge. If anything they sharpen it. When building is easy, the hard part is knowing what to build, who to build it for and why it will matter to them. It is knowing how to make an unfamiliar product relevant, how to earn attention in a crowded market, and how to turn a first visitor into an audience, an audience into customers, and a promising product into a business that lasts.

None of those are technology questions. They are marketing questions.

The bottleneck hasn’t gone away. It has moved.

Most of the AI conversation ends at the moment of creation. Someone builds an app over a weekend, stands up a website in an afternoon or generates a month of content before lunch. The speed is impressive. The harder commercial questions are still sitting there.

A product can be technically possible without solving a problem that matters. A website can look credible without giving anyone a reason to visit. A business can publish constantly without building an audience. An automated email sequence can run flawlessly while pushing an offer nobody wants.

AI reduces the effort of bringing an idea into the world. What it cannot do is establish that enough people have the problem, are unhappy with the alternatives, are willing to change their behaviour and can be reached at a cost that makes sense. It cannot tell you whether the offer is distinctive enough, whether buyers will trust an unfamiliar provider or whether the economics will support growth.

Theodore Levitt warned about this decades ago in Marketing Myopia. Businesses become vulnerable, he argued, when they define themselves by the products they make rather than the customer needs they serve. The danger is greater still when a product can be built before the market has been understood.

The ability to build is not evidence of demand.

A founder may be able to launch globally on day one, but global availability is not global awareness. A product may be capable of serving a million customers, but scalability means nothing if the first hundred cannot be found. The fact that something exists does not mean anyone will notice it, understand it, remember it, trust it or choose it.

The bottleneck has moved from producing the product to creating the market conditions in which the product can succeed.

One long-form edition at a time, on making better marketing decisions. Free to subscribe.

Subscribe now

Marketing was never just promotion

One reason this shift is misread is that marketing has slowly been reduced to a narrow set of visible activities.

In most organisations, marketing has become shorthand for advertising, social media, content, email, campaigns and lead generation. These are legitimate parts of marketing. They are not the whole discipline.

The American Marketing Association defines marketing as the activities and processes involved in “creating, communicating, delivering, and exchanging” offerings that have value. Communication is in that definition, but it sits alongside the creation and delivery of value. It is one of four verbs, not the headline.

The Four Ps make the same point. Product, Price, Place and Promotion describe a commercial remit far broader than communications. If you are newer to the profession, this is worth committing to memory: promotion is one component of the marketing mix. It is not the definition of marketing.

Philip Kotler spent a career establishing marketing as the process of understanding markets, identifying customer needs, creating value and choosing which customers an organisation is best equipped to serve. In that conception, marketing begins long before a campaign is briefed. At its fullest, it shapes the product, the market, the proposition, the route to market and the commercial model.

That places marketing upstream in the business. It should help decide which problems are worth solving, how a market should be understood, which customers represent the best opportunity and why the organisation has a credible right to win. It should have a voice in the offer, the price customers will accept, the way the product is distributed and the position the business can occupy in the customer’s mind.

Promotion comes after those decisions. It cannot compensate for their absence.

Yet that is rarely how it plays out. Marketers are often brought in only after the product has been built, the commercial target set and the launch date agreed. Product builds, finance prices, sales picks the audience, and marketing is asked to generate demand. In many B2B companies it is worse than that: the marketing department is treated as a support function for Sales.

When marketing is confined to the final stage, its job is no longer to help the organisation understand and shape the market. Its job is to communicate decisions that have already been made.

AI did not cause this narrowing. It started much earlier.

I’d argue that digital marketing did a great deal to accelerate it.

How digital marketing changed what we valued

Let’s be clear: digital marketing brought enormous benefits. It expanded access to customers, made experimentation faster and allowed small businesses to reach markets that were previously out of range. It created new forms of customer experience and gave marketers a level of visibility into behaviour and performance that earlier generations could only guess at.

It also made marketing accountable. For organisations that had struggled to connect marketing spend to commercial outcomes, that was a real advance.

But every new capability changes what people pay attention to. Digital platforms delivered an immediate, constant stream of measurable activity: impressions, clicks, open rates, conversions, leads, cost per acquisition, return on ad spend. Because those numbers were visible, fast and easy to put on a slide, they became the easiest things to report, optimise and defend.

The problem was never measurement. The problem was letting what could be measured immediately define the value of marketing.

Les Binet and Peter Field identified this tension in their analysis of the IPA Effectiveness Databank. Their research distinguished short-term activation, which converts demand that already exists, from long-term brand building, which creates future demand and produces broader commercial effects over time. They also warned that growing reliance on very short-term online measures would undermine long-term effectiveness. Most of the evidence since suggests they were right.

Digital did not make strategy unnecessary. It made execution unusually absorbing. There was always another campaign to launch, another platform to learn, another audience to optimise, another dashboard to update, another conversion point to improve. Marketing teams organised themselves around channels and systems, and professional expertise became closely tied to the ability to operate them.

Over time, channel choices crowded out market insight and customer choices. Optimisation displaced diagnosis. A media plan was presented as a marketing strategy, and a collection of platform tactics was mistaken for a coherent approach to growth.

Not because digital marketing failed. Because it offered so many useful, measurable things to do.

The urgent gradually displaced the important.

AI could repeat the same mistake

If that sounds familiar, it should. The early AI conversation is following the same script. Marketers are being urged to find the best tools, write better prompts, automate more workflows, produce more content and deploy agents on ever more complex tasks. Do more. Produce more. Automate more. Drive efficiency at scale.

AI has unlocked a great deal that is useful. It reduces repetitive work, shortens production cycles and makes expertise more accessible. But these are execution capabilities.

AI is very good at helping answer the question, “How can we do this?”

It is far less inclined to answer, “Should we do this at all?”

AI can generate hundreds of content ideas without knowing whether the market needs more content. It can produce message variations without deciding which idea the brand should consistently own. It can summarise customer research without determining which customers the business should prioritise, or which it is willing to walk away from. It can build a campaign without taking any responsibility for the commercial reasoning behind it.

The research on the “jagged technological frontier” shows why this distinction matters. In the same Boston Consulting Group experiment, AI produced substantial gains on tasks inside its capabilities. On a task deliberately designed to sit outside that frontier, consultants using AI were 19 percentage points less likely to reach the correct answer than those working without it.

The lesson is not that AI should be avoided. It is that the technology’s value depends on the quality of the task, the context in which it is used and the judgement of the person directing and evaluating the work.

AI can make valuable work faster. It can also make poorly chosen work faster.

If marketing responds to AI by producing more activity without improving the decisions behind it, the profession will repeat the mistake of the digital era, only at far greater speed and scale.

A product does not create its own demand

The belief that a good product will find its own market is remarkably persistent, especially among founders and product-led organisations. Occasionally it happens. It is not a growth strategy.

Customers do not evaluate every available product and select the objectively best one. They operate with limited time, incomplete information and varying levels of interest. They rely on memory, familiarity, trust, recommendations and whatever is easiest to find or justify.

This matters enormously in B2B, where purchases are infrequent, involve several stakeholders and carry professional and organisational risk. At any given moment, most potential buyers are simply not looking.

Professor John Dawes’ 95:5 rule is a useful way to think about this. The exact ratio varies by category and should be treated as a rule of thumb rather than a constant. The principle holds: for many B2B purchases, the large majority of potential buyers are out of market at any given time. Advertising and other forms of brand building therefore work largely by creating and refreshing memories that can be retrieved when a buyer eventually enters the category.

For a new business, the implications are significant. Most of its future customers are not searching for it. They may already have a supplier, lack budget, be locked into a contract or have no immediate need. They are not being difficult. They are not in the market yet.

Demand-capture activity can help a brand compete for the minority who are in market now. Search advertising, comparison content, review platforms and sales outreach all matter at that late stage. But none of them explains how the brand entered the buyer’s mind before the search began.

Someone has to create the memory. The customer must encounter the brand, understand what it stands for and connect it to a relevant need or buying situation. Familiarity and trust have to be in place before the organisation appears on a shortlist.

The Ehrenberg-Bass Institute describes this as mental and physical availability. Brands grow by becoming easier for more category buyers to think of and easier to buy. That requires reach well beyond the small group shopping today, and consistent links between the brand and the situations that bring people into the market.

This is the work that happens long before a sale shows up in a dashboard. It is harder to measure through an attribution model, which is arguably the wrong instrument at this stage of the buying journey anyway. And it is the work that many emerging businesses underestimate, because its effects are less immediate than launching a product or watching a click.

From the first visitor to a scalable business

AI can help a business build its first website. It cannot deliver the first visitor. And once that visitor arrives, the real work starts: understanding what brought them there, whether the offer is relevant and what would persuade them to come back.

An audience is not accumulated traffic. It is a group of people who recognise value in what a business consistently provides. Building one requires a clear understanding of who the organisation wants to serve, the problems it can credibly help solve and the ideas it can contribute better than the alternatives.

Turning an audience into customers takes more than visibility. The offer must be clear, sensibly priced and easy to evaluate. The customer must believe the value outweighs the cost and the risk of changing. In B2B, the proposition may also need to survive procurement, finance, IT, legal and executive scrutiny.

Then comes scale, and a fresh set of questions. Which acquisition channels are repeatable? Do the customer economics hold? Can the brand grow beyond the founder’s personal reputation? Will the proposition stay distinctive once competitors respond?

None of this is separate from product development. It is building the business.

A technically scalable product with no repeatable route to market is not yet a scalable company. A large potential audience that cannot be reached economically is not yet a viable market. High engagement that never becomes customer value or revenue is not yet growth.

Marketing, done properly, is what connects the idea to the economic reality around it.

When execution becomes abundant, judgement becomes scarce

The falling cost of execution does not make marketing judgement less important. It makes it the differentiator.

Execution turns an instruction into an output. Judgement determines whether it was the right instruction in the first place.

Judgement is what separates an interesting idea from a commercially valuable one. It is what lets a marketer interpret incomplete or contradictory evidence, weigh short-term revenue against long-term growth, and choose between several plausible courses of action.

It is also required because marketing decisions are never made in a laboratory. Customers change, competitors respond, budgets tighten, and every organisation brings its own history, capabilities and politics to the table. The same evidence can lead to different choices depending on the market, the business model and the appetite for risk.

AI can contribute to that process. It can retrieve information, challenge assumptions, spot patterns, develop scenarios and speed up how a marketer explores a problem. It may even recommend a course of action.

But the marketer remains responsible for deciding whether the evidence is credible, whether the recommendation fits the context and which consequences the organisation is prepared to accept.

This is why more powerful technology demands stronger foundations, not weaker ones. The sharper the tool, the more it matters to know what good marketing actually looks like.

Evidence informs. Judgement decides.

AI should give marketers time to become marketers again

For years, marketers have said they do not have enough time for strategic work. Their days are consumed by campaign administration, reporting, decks, data cleaning, asset production, platform management, stakeholder requests and coordination.

AI can lift some of that. The evidence from writing, coding and professional knowledge work suggests that well-designed AI assistance can hand real time back to people while improving performance on suitable tasks.

The more interesting question is what organisations and marketers do with that time.

The path of least resistance is to demand more output. If a marketer can produce five pieces of content in the time it used to take to produce one, the organisation will ask for five. If reporting can be automated, it will ask for more dashboards. If campaigns can launch faster, it will launch more campaigns.

That captures the productivity gain. It does nothing to improve the marketing.

The bigger opportunity is to reinvest the time in the work execution has been crowding out: talking to customers, spending time with sales, studying the category, understanding competitors, improving the offer, sharpening positioning and finding the real barriers to growth. Marketers could sit with product teams on what should be built, with finance on how value should be priced and with sales on how the market actually buys.

They could write better briefs. Ask harder questions. Stop activity that does not serve the strategy. Think more carefully about how brands are built in memory, how future demand is created and how the organisation becomes easier to choose.

AI should not simply make marketers faster at the version of the job they inherited. It should give them the capacity to reclaim the version of marketing the business actually needs.

AI, used well, can help marketers get back to marketing.

Returning to first principles

Returning to first principles is not a case for resisting technology or retreating to some earlier era. Digital platforms, analytics, automation and AI have all expanded what the profession can do. The task is to reconnect those capabilities to the purpose they were meant to serve.

First principles provide stable footing when the tools are changing by the month. They ask marketers to begin with the market rather than the channel and with the customer rather than the campaign. They ask whether a real need exists, whether the organisation can create meaningful value, whether the offer is distinctive, and whether customers can easily think of it, find it and buy it.

They also give marketers a better yardstick for evaluating new technology. The question is not simply whether a tool can generate content, automate a workflow or cut cost. It is whether it helps the organisation understand customers more clearly, make a better decision, create more value, build stronger memories or make it easier to buy.

Technology should strengthen marketing capability rather than substitute for marketing thought.

Without first principles, a business can execute efficiently in the wrong direction. With them, AI becomes more useful, because it is being applied to better-defined problems, with stronger evidence and clearer standards for judging the outcome.

Let’s get Back To marketing.

Five questions to ask before you build

If the ability to create is no longer the constraint, the discipline lies in what you ask before you create. Before the next product, course, newsletter or campaign, answer these five honestly:

1. Do enough people experience this problem to sustain a business, and is it important enough that they would change their behaviour to solve it?

2. Is your audience dissatisfied with the current alternatives, or merely aware that alternatives exist?

3. Can you reach your target audience economically, through channels that are repeatable?

4. Why would your target audience trust an unfamiliar provider, and what will earn that trust before the first sale?

5. What will make your target audience think of you when they finally enter the market, given that most of them are not buying today?

If you cannot answer these, the product is not ready, however quickly it can now be built. If you can, your marketing strategy is starting from a good place.

The return of marketing

The AI era will create more builders than any before it. Individuals and small teams will develop products that once required substantial resources and organisations. The distance between an idea and its execution will shrink, and experimentation will get cheaper.

AI is a meaningful democratisation of capability. It will not democratise attention in the same way.

Customers will still have limited time and more choices than they can evaluate. Businesses will still need to decide which markets to enter and which problems are worth solving. New products will still need to become known, remembered, trusted and easy to buy. Organisations will still need to balance the demand they can capture today with the demand they must create for tomorrow.

None of that can be solved by production alone.

It requires marketers who understand the commercial role of the discipline, who can tell activity from progress, and who know how to use evidence without surrendering judgement. It requires marketing teams that contribute before the product is finished, not only after the launch date is set.

Back2Marketing exists to help restore that first-principles version of marketing.

We believe marketing is a commercial discipline, not a promotional service. It begins with understanding markets and customers, creating value and making choices about where and how a business can grow. Communication matters, but it cannot repair a weak proposition or create lasting demand for something the market does not value.

We believe digital and AI should increase the contribution marketers make to a business. They should reduce lower-value work, improve access to evidence and give marketers more time to think. They should not trap the profession in an ever-accelerating cycle of more content, more campaigns and more optimisation without a strategic foundation underneath.

Above all, we believe first principles become more valuable as the world becomes more complex. Tools will change. Channels will change. The economics of production will change. The need to understand people, create value, build demand and make sound commercial decisions will not.

Everyone can now build. Few know how to market.

That is the gap Back2Marketing exists to close.

Evidence informs. Judgement decides.

New essays land here first: subscribe and the next one will find you.

And if you want to build the judgement, not just the output, Back2Marketing B2B Marketing Fundamentals course teaches marketing from first principles.

Subscribe now

Sources

• American Marketing Association, Definitions of Marketing, on marketing as the creation, communication, delivery and exchange of value.

• Philip Kotler, The Past, Present and Future of Marketing, on the development of marketing and the centrality of customer value.

• Shakked Noy and Whitney Zhang, Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence, Science, 2023.

• Fabrizio Dell’Acqua et al., Navigating the Jagged Technological Frontier, Harvard Business School and Boston Consulting Group field experiment.

• GitHub, Quantifying GitHub Copilot’s Impact on Developer Productivity, controlled experiment examining task-completion speed.

• Les Binet and Peter Field, The Long and the Short of It, IPA analysis of short-term activation and long-term brand building.

• Ehrenberg-Bass Institute for Marketing Science, Advertising Effectiveness and the 95:5 Rule, on out-of-market buyers and the role of advertising in building memory.

• Ehrenberg-Bass Institute for Marketing Science, research on how brands grow through mental and physical availability.