The 60/40 rule turns up in almost every presentation on brand building. The original source says something else. On the chart in question, in the free IPA presentation by Binet and Field, the label reads optimum: ~40%, tilde and all, marking an estimated peak on a curve that runs fairly flat across the top. One slide earlier sits the number the rule is named after, and it is not a recommendation: brand channels 61% against activation channels 39% describes what the UK market spent that year.
This article checks four of those figures against their own source. Not to knock them down, because they mostly hold up. Rather because the authors write caveats that disappeared somewhere on the way into the trade press, and those caveats change what you can do with the numbers. After that I put three Dutch measurements next to them that get quoted far less and that sit closer to a conversation with a mid-size board.
Our reason for doing this is practical. We back brand promises with figures, and a figure a client can check themselves is worth more than one that impresses. Every source below is free and public. You can verify this entire article in an afternoon.
What does the source of the 60/40 rule actually say?
Two things, compressed into one number. The free IPA presentation accompanying the book "The Long and the Short of It" contains a bar chart of UK media spend split 61% brand channels against 39% activation channels. That is a photograph of the market, not a recommendation.
The actual finding is on the next slide: a curve plotting the number of very large business effects against the activation share of budget, labelled optimum: ~40%. The tilde is the whole point. It marks an estimated peak, and the curve is flat enough at the top that being a few percentage points off barely matters.
The dataset behind it is solid and sits right there on slide two: a meta-analysis of the IPA effectiveness databank covering 996 campaigns, 700 brands and 83 categories. Anyone writing that it covers close to a thousand cases is on solid ground. Anyone inventing a different number is not.
There is one more limitation that rarely travels with the figure. The full report is a paid book. The free presentation is therefore the only publicly checkable carrier of this number, which is exactly why you should always link to it rather than to a summary of a summary.
We no longer write "the 60/40 rule". We write "an optimum around 40% activation". That is more precise than almost every agency website out there, it costs you nothing in argument, and a client can check it in two clicks. We apply the same discipline to everything we claim about brand identity and revenue.
Why doesn't 60/40 apply to B2B?
Because the same authors found a different number for B2B, and attached a warning to it. Binet and Field produced the B2B cut of that same databank for the B2B Institute. It says efficiency appears to peak at roughly 46% of budget to brand and roughly 54% to activation.
Immediately after comes the sentence that almost never gets quoted along with it: the ratio should not be followed too precisely, it is a guiding principle, and the small sample supports only a rough estimate. The underlying chart draws on B2B cases in the databank from 1998 to 2018.
So two things go wrong at once in day-to-day practice. Anyone applying 60/40 to a B2B brand is quoting the authors against their own cut of the data. And anyone presenting 46/54 as a hard number is doing precisely what that same paragraph rules out.
What survives, then? The direction. For B2B the centre of gravity sits slightly closer to activation than it does for consumer brands, and that makes sense: smaller buying populations, longer decision cycles, more familiar names in the consideration set. But the exact percentage is noise, and an agency selling it as precision is selling precision that isn't there.
This is the kind of nuance that wins a pitch. When a client says "but 60/40", the right answer is neither yes nor no. The right answer is to open the source and show what it says. That conversation ends differently from an argument about who has the better source.
Is the 95-5 rule research or a rule of thumb?
A rule of thumb, and the official publication says so in as many words. On page three the author writes that the 95% figure is not intended as a precise rule but is used as a heuristic to convey that the vast majority of businesses are not in market at any given moment.
The derivation is arithmetic from a single example. Companies change their bank or their law firm roughly once every five years, so about 20% per year, so roughly 5% in market per quarter. No meta-analysis, no sample, a calculation.
That does not make it less useful. It makes it useful in a different way. Because if it is a calculation, you can run it for your own category, and that is worth far more than the default number. The question is not "does 95-5 hold?" but "what is the interpurchase interval in my category?". Sell something customers replace every two years and roughly half your market is in market within a year, and roughly an eighth of it per quarter. That is a completely different media plan from the one 95-5 implies.
We use that calculation in brand projects as the first test of a content strategy. The longer the interpurchase interval, the stronger the case for recognition over activation. The shorter it is, the more you can steer on the moment itself. It is the same reasoning that sits underneath defining brand values: establish how the buyer behaves first, then decide what you say.
One more detail completes the irony. In the most recent IPA presentation, 95-5 is used as an indictment, not a recommendation. Agencies quoting the rule to sell targeting at the in-market group are quoting it exactly the wrong way round.
What was actually measured in 2025?
Something that moves the entire budget conversation. At the 2025 IPA effectiveness conference, Les Binet and Will Davis presented an analysis of the databank breaking down what explains the variance in incremental profit. The result: budget explains 89% and return explains 11%. Binet's own summary on the next slide is that budget is eight times more important than return.
That is an uncomfortable finding for a profession that mostly talks about efficiency. The rest of the presentation makes it more uncomfortable still. Their 2025 survey of senior marketers produces a run of figures about how small the thinking has become:
| What marketers do | Share |
|---|---|
| Target subsegments rather than the whole category | 56% |
| Ignore the over-45s | 62% |
| Target the small group currently in market | 68% |
| Use a narrow media mix | 53% |
See the original slides for the full run. In 48% of cases activation dominates the brand budget, which means the 60/40 debate often isn't even had. And media return itself barely moved: from 3.07 before the pandemic to 3.15 after it.
For a mid-size brand that translates into one question that rarely gets asked. Not: are we splitting our budget well? But: is there enough budget to produce an effect we can measure at all? That is a less pleasant conversation and a more honest one. We would rather have it at the start of a project than afterwards at the evaluation, and it is why in what branding does to your advertising costs we talk about size and not only about ratio.
Which Dutch figures say anything about this?
Three, and all three get quoted far less than the British canon. That is a shame, because they cover our own market.
The first comes from research into brand orientation among mid-size Dutch B2B companies, carried out with Erasmus University. From it: 63% of companies have no internal programme helping staff understand and carry the brand values, and only 43% measure the brand's impact on commercial performance. The sample is small, 30 companies of fifty to two hundred and fifty staff, so present it as indicative and never as "Dutch mid-market".
Those two numbers explain each other. If you don't measure brand impact you can't defend brand budget, and if you can't defend it you don't get it. That is the chain a brand project actually breaks on, and it is an organisational problem before it is a creative one.
The second comes from Exact's annual mid-market barometer, with a sample of more than seventeen hundred Dutch owners and managers. From it: 93% of mid-market firms are investing this year. So the money is being spent. It's just that marketing sits below technical and technological tooling in the list of destinations. Brand is not a cost-cutting target. It is a priority that loses.
The third is the best Dutch effectiveness material there is, and it appears almost nowhere in trade content. Together with bvA, VIA and Effie Awards Netherlands, SWOCC coded 439 Dutch Effie entries from eleven years against the factors that explain campaign success. The public result: campaigns running longer than two years win an Effie in more than half of cases, while campaigns shorter than six months have a much lower hit rate. See SWOCC's write-up.
That last one is the argument you use against short-term thinking, and it comes from our own market rather than from British data fifteen years old. We use it as standard in the conversation about why a brand project is not a campaign, as we do in building an e-commerce brand.
Which figures are better left alone?
Three kinds, and all three circulate in brand content.
The first kind is attribution without a report. There is a statistics page doing the rounds with a run of Dutch brand figures, each attributed to a well-known institute and a year, with no report title, no date and no findable publication. When we looked one of those studies up in the named institute's own index, it wasn't there. A brand name plus a year is not a source. A source has a title and an address.
The second kind is citation drift. About the same survey of B2B marketers, one official party writes that 96 out of a hundred expected an effect within two weeks, and the other party that co-ran the study writes that it was 95. One percentage point apart, between the two organisations that did it together, and neither with a link to the sample. That isn't fraud. It is precisely the mechanism by which figures in this field slowly drift.
The third kind is the number without a method. A widely quoted figure about the value of AI traffic comes from a single blog post by a tooling vendor. The only mention of scope on the entire page is a count of topics. Missing: number of websites, number of sessions, the definition of a conversion, the measurement period and the attribution model. We don't quote that, not even with a citation, because citing a number that has no method lends it authority it hasn't earned.
The rule we distilled from this is short. A figure can go in if you have opened the primary carrier, the number appears verbatim, and you can name the sample and the year. If it doesn't clear that bar, you write the sentence qualitatively. A qualitative sentence always beats invented precision.
How do you check a brand statistic yourself?
In about ten minutes, in four steps. This is literally the routine we run before a number is allowed into a proposal or an article.
Step one: find the primary carrier. Not the article quoting the figure, but the report, the slides or the dataset behind it. If you get no further than one agency blog pointing at another agency blog, that is your answer.
Step two: find the number verbatim in that document. If there is a tilde, an "approximately" or a range, it belongs in your sentence too. The tilde in "optimum: ~40%" is not a detail. It is the meaning.
Step three: note the sample and the year. Thirty companies is not a thousand, and data from 2011 is not data from 2025. Both are fine, as long as you say so.
Step four: find the caveat. Authors of serious research almost always write down where their figure stops applying. That sentence rarely survives into the quotes, and carrying it with you is exactly what makes you credible.
What you gain is more than accuracy. It is a different kind of conversation. A client who notices that you have read the source starts deliberating with you instead of negotiating. We see that difference in every discovery session, and it is cheaper to arrange than any brand promise.
What we do with this
We turned it into a house rule. Every figure we publish must carry a clickable link to the primary source where that number appears verbatim, and on sensitive topics that source has to be a regulator, a government, a standard or original research. An agency blog does not count as a source, even when it happens to quote the same number. The rule runs automatically: an article with an unlinked figure doesn't go live here.
That is stricter than necessary and it costs time. It buys two things. The texts are checkable, which in a market full of borrowed figures is distinctive in itself. And it forces us to write about what we actually know, which in practice means we make a sentence qualitative more often than we go hunting for a number that happens to fit.
For brand projects the same principle applies one level deeper. A position you can't support is an assertion. A position you can tie to a measurement is an argument. That difference decides whether a brand story survives the moment someone looks at it critically, and that is exactly what we work on in our brand and strategy projects.
Want to know which figures circulate in your category and which of them have a source, get in touch. We'll walk through them with you and you get the list, even if you take nothing else from us. In six years we have run that exercise often enough to know there is always at least one figure in there that nobody ever looked up.
