The buyer was never choosing you

Most brand strategies still imagine a moment of decision. A consumer sits at a table, lays out three options, weighs them carefully, and picks the best. The strategy deck has a funnel. The funnel has a stage labelled "consideration." The deck assumes that consideration is where brands compete.

Decades of buyer behaviour research say this picture is almost entirely wrong. Buyers do not deliberate. They retrieve. When a need surfaces — a hunger, a deadline, a worry, a Tuesday night — the brain produces a small, fast, unimpressive list of brands that are already there. Whatever sits on that list wins most of the time. Whatever is missing from it does not even get a chance to lose.

Jenni Romaniuk and the Ehrenberg-Bass Institute have spent twenty years stress-testing this idea across categories, geographies, and product types. Their conclusion is uncomfortable for the industry that pays them: the buyer is not choosing your brand against competitors. The buyer is choosing one of the brands their memory served up, in the second after the need appeared. Marketing's job is not to win the comparison. It is to be on the list when the need shows up.

This is what Romaniuk calls mental availability, and the mechanism that builds it is called the Category Entry Point. A CEP is a cue , a context, a moment, a question, an emotion — that triggers a category need and pulls a small set of brands into mind. "Quick lunch on the way to a meeting" is a CEP. "Birthday gift for a colleague I barely know" is a CEP. "Insurance for a car I just bought used" is a CEP. The more CEPs a brand is linked to in memory, and the stronger those links, the more often the brand gets retrieved when buying happens.

The brand that wins is rarely the brand that was best considered. It is the brand that was already there when the question arrived.

This article is about what that shift means. Not as a slogan, and not as a vague nod to "salience." As an operating model for how brands actually grow — and as a diagnostic for why so many strategies that look smart on a slide produce nothing in the market.

This is the eighth piece in the Catchlight series, and it names the mechanism the earlier seven have been circling. "The Science of Being Seen" argued that visibility is the scarcest input in marketing. "The Answer Is Always Seven" showed how AI systems inherit the biases of the text they retrieve from. The short pieces on loss aversion, scarcity, and social proof catalogued the heuristics buyers actually use at the moment of choice. "The Half-Life of Marketing" established that brand effects decay in the absence of reinforcement. "The Feed You Didn't Choose" traced how algorithms now reshape what buyers encounter before they ever reach a category decision. Each of those articles describes a condition of buying. This one describes the mechanism underneath them all: buyers do not deliberate, they retrieve. Category Entry Points are the cues that fire that retrieval. The brand that wins is the brand already sitting in memory when the cue arrives — and everything the previous seven issues described is, in the end, an argument about how to put it there.

The research behind mental availability

The intellectual debt here is to Andrew Ehrenberg, whose work in the 1960s and 70s on repeat-buying laws established that markets are remarkably regular. Across categories, brand penetration — the share of buyers who buy at all — explains almost all of the variation in brand size. Loyalty differences are small and largely a function of penetration. Big brands are big because more people buy them, not because their buyers are more devoted.

Byron Sharp's How Brands Grow (2010) compressed this into laws marketers could repeat. The Double Jeopardy law: small brands suffer twice, with fewer buyers and slightly lower loyalty. The Duplication of Purchase law: a brand's buyers also buy competitors in line with those competitors' market shares. Both laws point to the same conclusion. Buyers are not loyal in the way brand teams imagine. They have repertoires, and brands compete to be in the repertoire.

Romaniuk extended this work into memory. In Building Distinctive Brand Assets (2018) and Better Brand Health (2023), she argued that the size of a brand's repertoire footprint is determined by mental availability, the probability that a brand is thought of in buying situations , and physical availability — the ease of finding and buying it. Mental availability is built by linking the brand to as many relevant CEPs as possible, with associations that are fresh, fluent, and consistent over time.

The empirical support is extensive. In a 2021 study published in the Journal of Advertising Research, Romaniuk and colleagues analysed buyer behaviour across 18 categories and 14 countries. Brands with broader CEP coverage grew faster, recovered faster from share dips, and were more resilient to competitive pricing. The relationship was not subtle. CEP breadth explained roughly 60% of the variance in market share growth across the sample.

The Ehrenberg-Bass position is not that targeting, segmentation, or differentiation are useless. It is that they have been wildly overweighted relative to the simpler problem most brands actually face: being mentally available when the buyer is in market. And buyers are in market constantly, in tiny moments, distributed across hundreds of contexts the brand team never sees on a research debrief.

Why CEPs Beat Personas

The persona is one of the most beloved and least useful artefacts in marketing. It produces a fictional buyer with a name, a job title, a set of frustrations, and a stock photo. It implies that the buyer is a stable type whose decisions can be predicted from demographics and attitudes.

Buyers do not work this way. The same person buys premium and budget products in the same category, in the same week, depending on context. The same parent buys organic baby food on Tuesday and a supermarket own-brand on Saturday. The relevant unit of analysis is not the buyer. It is the buying occasion — and the cue that triggered it.

This is the first practical break with persona thinking. CEPs replace "who" with "when." When does the buyer want this category? What were they doing? Who were they with? What were they trying to get done? What were they worried about? Romaniuk's framework — Why, When, Where, While, With/For Whom, How Feeling — turns the question of segmentation into a structured map of buying contexts.

A CEP map for a coffee brand will surface dozens of distinct entry points: the morning commute, the post-lunch slump, the meeting where you need to look prepared, the weekend with a newspaper, the gift for a host. Each one triggers a slightly different short-list. A persona collapses all of this into "the busy professional, 28-45." That is not a buyer. It is an apology for not having done the research.

The second break is more significant. Persona-based strategy assumes the brand is competing with other brands aimed at the same persona. CEP-based strategy assumes the brand is competing with whatever else gets retrieved at the same moment. For a midweek dinner CEP, a frozen pizza brand competes with a meal-kit subscription, a takeaway app, a packet of pasta, and the leftovers in the fridge. None of those competitors share the persona. All of them share the cue.

The work of CEP research is not glamorous. It is interviews, diary studies, and structured prompts that surface the moments buyers actually experience. The output is not a poster of a fictional person. It is a list of cues, ranked by frequency and brand strength, that tells the team where the brand is already strong and where it is invisible. Strategy then becomes a matter of choosing which cues to defend and which to attack.

How memory actually works in buying

The reason CEPs matter is that human memory is associative, not deliberative. Cognitive psychology has known this since the 1970s. Endel Tulving's distinction between semantic and episodic memory, John Anderson's spreading activation models, and Daniel Schacter's work on memory reconstruction all point to the same picture: memory is a network of nodes connected by associations, and retrieval is a function of how strongly a cue activates the right nodes.

This has direct consequences for marketing. A brand is not a thing the buyer "knows about." It is a node in memory connected to other nodes — colours, sounds, situations, feelings, faces, slogans. When a cue fires, activation spreads from that cue to whatever nodes it is connected to. The brands that come to mind first are the ones with the strongest, freshest, most numerous connections to the cue.

Daniel Kahneman's System 1 framework, popularised in Thinking, Fast and Slow (2011), describes the experiential consequence of this. Most decisions are made by the fast, automatic system that does not deliberate. It produces a feeling, a recognition, a sense of fluency — and the conscious mind ratifies the choice afterwards. Buyers experience this as preference. Researchers know it as availability bias and processing fluency. Brands that are easier to retrieve are felt to be better, regardless of objective quality.

The implication is that marketing investment is, at base, an exercise in shaping the structure of memory. Every ad, every package, every shelf placement, every search result is either reinforcing existing associations or attempting to create new ones. The brands that grow are the ones whose investments are coherent over time — same assets, same cues, same emotional textures — because consistency is what allows associations to harden into reliable retrieval.

Inconsistency is the most expensive mistake in this model. A brand that changes its colour, its slogan, its spokesperson, its tone, and its target every two years is not building memory. It is wiping the chalkboard before the chalk dries. The buyer's brain has no way to consolidate what the brand stands for, and the brand quietly disappears from the retrieval set.

This is also where the half-life problem from last week's article connects. Marketing investment decays not because attention is short — though it is — but because memory traces fade unless they are refreshed. CEP-linked associations are no exception. A brand that owned "morning coffee on the commute" five years ago does not own it today unless it has spent five years reminding the market. Mental availability is a stock that depreciates.

The Diagnostic: Where your brand is invisible

The most useful thing a CEP framework offers a strategy team is not a new theory. It is a diagnostic. Once you map the entry points in a category and measure how often your brand is retrieved at each one, the gaps are obvious — and they are usually nothing like what the brand team expected.

The standard diagnostic runs in three steps. First, generate the list of CEPs in the category through qualitative research. Romaniuk recommends 20–50 CEPs for most categories, structured along the W framework. Second, run a quantitative survey in which buyers are asked, for each CEP, which brands come to mind. The output is a matrix: rows are CEPs, columns are brands, cells are mental market share at that entry point. Third, compare your brand's share at each CEP against your overall market share. The pattern reveals strength, weakness, and opportunity.

The findings are usually humbling. A brand that holds 18% market share might dominate three CEPs at 40%+ and be effectively invisible at fifteen others. The temptation is to celebrate the strongholds. The opportunity is in the gaps. Each missing CEP is a buying occasion the brand was not retrieved for — and therefore not chosen at. Multiplied across the category, those absences are where competitors quietly grow.

The diagnostic also exposes the cost of category-generic advertising. A campaign that says "we make great coffee" links the brand to the broad concept of coffee but not to any specific entry point. It buys attention and forgets to attach the attention to a moment. The buyer remembers seeing an ad. The buyer does not remember the brand when the morning slump arrives. Mental availability did not move.

The contrast is campaigns that target specific CEPs explicitly. Snickers spent a decade linking itself to the cue of "hungry and irritable" through "You're not you when you're hungry." The campaign did not invent the cue. The cue existed in every buyer's life. Snickers attached itself to the cue with such force that the cue now retrieves the brand automatically. Sales growth followed for years, well beyond the period most brand campaigns sustain.

In categories where buying is private — insurance, finance, health, B2B procurement — the diagnostic is even more valuable. You cannot watch the buyer in the store. You cannot see the moment of decision. The CEP map is the only structured view of when the category is actually entering the buyer's mind, and where your brand is or is not present at that instant.

Building CEPs in an AI-Mediated World

The CEP framework was developed in an era when buyers retrieved brands from their own memory. That era is ending. A growing share of buying journeys now begin with a query to an AI assistant or a generative search engine, and the model retrieves the candidate set on the buyer's behalf. The cue still fires. The retrieval still happens. But the retrieval is no longer happening inside the buyer's head.

This is the structural shift the next decade of marketing has to absorb. Mental availability is being supplemented — and in some categories replaced — by model availability: the probability that a brand is named in the response when an LLM is asked the implicit question behind a CEP. "I need quick lunch near my office" becomes a query. The model returns a short list. The brands on that list win. The brands missing from it are not on the buyer's mental list either, because the buyer has outsourced retrieval.

The good news is that the underlying mechanism is the same. Models build their candidate sets from associations encoded in training data — text on the open web, structured data, reviews, third-party mentions. A brand that has spent ten years linking itself to a specific cue across its owned and earned channels has, by accident, built the exact corpus the model uses to answer the query. CEP-driven brand building is also the most effective form of LLM SEO, because it produces the kind of context-rich, situation-anchored content that models latch onto.

The bad news is that brands which optimised for a different game — keyword density, persona-based messaging, generic awareness — have built nothing the model can use. They are mentally available in some buyers' heads and model-invisible everywhere a query is mediated. The gap between those two states will widen as AI-mediated journeys take share from organic search.

The practical implication is that CEP work is now infrastructure for two parallel retrieval systems, not one. The brand team's job is to identify the cues that matter, build distinctive assets that attach the brand to each cue, and then ensure those associations exist in formats both humans and models can ingest — long-form content with explicit context, structured FAQs, third-party citations, reviews that describe situations and not just product features. The output is not "content marketing." It is a deliberate, situation-anchored corpus designed to make the brand retrievable across both kinds of memory.

This is also why the brands most exposed to AI disintermediation are the ones that never bothered with mental availability in the first place. They were riding paid search, retargeting, and personas. None of that survives a world in which the buyer asks a model and accepts the answer. Mental availability — and its newer twin, model availability — is the only durable defence.

What to do on Monday

If this argument is right, the work of brand strategy in the next twelve months looks different from what most teams are doing now. It is less about positioning workshops and more about cue inventories. Less about persona refreshes and more about situation research. Less about awareness scores and more about retrieval audits.

A pragmatic sequence for a brand team that wants to move on this:

First, run a CEP discovery exercise. Talk to thirty buyers about the last five times they bought the category. Capture the cues — what triggered it, what they were doing, who they were with, what they needed. Cluster the cues into 25–40 distinct entry points. This is unglamorous work and it should not be outsourced to a survey panel.

Second, run a quantitative CEP audit. For each entry point, ask buyers which brands come to mind. Compute mental market share at the cue level. Compare it to your actual market share. The gaps and the strengths will tell you, with unusual specificity, where to invest.

Third, audit your distinctive brand assets against the cues you want to own. Do your assets fire reliably when the cue is present? Are they consistent across the channels the cue lives in? If a buyer encounters your brand in the context of one of your priority cues, do they recognise the brand instantly? If not, the asset is doing nothing.

Fourth, audit your model availability. Ask the leading LLMs the implicit question behind each priority cue and see whether your brand is named. Where it is not, work backwards: what would need to exist on the open web for the model to retrieve you? Build it.

Fifth, hold the line on consistency. The biggest single cause of underperforming brands in CEP audits is not bad creative. It is changing the creative every eighteen months. Memory needs repetition to consolidate, and brand teams need the discipline to leave things alone long enough for the consolidation to happen.

None of this is novel. Romaniuk has been writing it down for fifteen years. Ehrenberg wrote the foundations sixty years ago. The reason it remains contrarian is that it asks brand teams to give up the things they enjoy — the workshops, the personas, the new launches, the rebrands — in favour of the thing that actually works, which is patient, situation-anchored, repetitive investment in being on the list when the buyer's need shows up.

The buyer was never going to deliberate. The door they walk through is the cue. The brand they choose is whichever one was already standing there.

References

Anderson, J. R. (1983). A Spreading Activation Theory of Memory. Journal of Verbal Learning and Verbal Behavior, 22(3), 261–295.

Ehrenberg, A. S. C. (1972). Repeat-Buying: Theory and Applications. North-Holland.

Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.

Romaniuk, J. (2018). Building Distinctive Brand Assets. Oxford University Press.

Romaniuk, J. (2023). Better Brand Health. Oxford University Press.

Romaniuk, J., Wight, S., & Faulkner, M. (2017). Brand Awareness, Journal of Product & Brand Management

Schacter, D. L. (1996). Searching for Memory. Basic Books.

Sharp, B. (2010). How Brands Grow. Oxford University Press.

Tulving, E. (1972). Episodic and Semantic Memory. In Organization of Memory. Academic Press.

Unsubscribe