Beyond demographics: why identical people buy differently.

Demographics are the oldest tool in marketing, and they were never very good. They survive because they are easy to collect, easy to count, and easy to put in a slide. You can buy a panel of women aged 25 to 44 in any country in the world before lunch. What you cannot buy is a panel of women who feel about your category the way your best customers feel about it, because that requires you to know what your best customers feel, and most brands do not.

The fundamental problem with demographic segmentation is that it confuses correlation with cause. Yes, on average, women buy more shampoo than men. Yes, on average, higher-income households spend more on cars. These are true statements at a population level. They have almost nothing to do with how an individual makes a decision. Inside any demographic group, there is enormous variance in what people want, how they buy, and what they will pay for. A 34-year-old in Kingston could be a loyal premium customer, an indifferent occasional buyer, a hostile former customer, or someone who has never heard of you. The demographic tag tells you which shelf to put her on. It does not tell you which person she is.

What actually predicts buying

If demographics are weak predictors, what is strong? Behavioural and psychographic variables, mostly, and within those, two stand out: the job a person is trying to do, and the way that person feels about doing it.

Clayton Christensen popularised the first one as "jobs to be done." The idea is straightforward and unfashionable. People do not buy products. They hire products to do jobs. A bottle of beer is hired sometimes to celebrate, sometimes to relax after work, sometimes to make a long bus ride more interesting. The same person hires the same brand for different jobs at different times. The same job sometimes hires different brands at different times. If you segment by demographics, you miss this entirely. If you segment by job, the world makes more sense.

The feeling layer is what neuromarketing adds. Two people doing the same job can feel completely differently about it. Both might be hiring a coffee shop for "fifteen minutes of escape on the way to work." One of them feels guilty about the spend and rushes through it. The other feels rewarded and lingers. Same demographic. Same job. Different emotional context. Different brand fit.

The implicit map

One of the more interesting tools that has come out of cognitive psychology in the last twenty years is the implicit association test, developed at Harvard by Greenwald, Banaji, and Nosek. The test measures how quickly people pair concepts with attributes under time pressure. The faster the pairing, the stronger the underlying mental association. You cannot game it. You barely have time to think.

When we run brand-version implicit tests, we are not asking people what they think about a brand. We are asking their brain what it knows about that brand, and reading the answer off the reaction times. The picture that comes back is often very different from the survey picture. A brand that scores 8 out of 10 on "premium" in a stated questionnaire might be slow to "premium" and fast to "ordinary" in the implicit test. The slow conscious answer was the polite answer. The fast unconscious answer was the operational one.

Segmenting on implicit associations is harder than segmenting on age, but the segments behave more like people. You end up with groups defined by how the category lives in their head, rather than how their birth certificate is filled in. In our work for Caribbean clients we keep finding the same broad pattern. Within any single demographic cell, there are usually three or four distinct mental models of the category, each with its own emotional fingerprint, and each requiring a meaningfully different conversation.

Two women, same shelf

Let us go back to the two women in Kingston. Both 34, both mothers, both shopping for the same household. Why does one buy the brand and the other not?

The answer almost never has anything to do with their age or their income, because their age and income are essentially identical. The answer has to do with three things: the job they are trying to do, the brands they associate with that job, and the feeling those associations carry.

Customer A might be hiring the category to "show care for my family" and your brand may have, through decades of advertising and family-table moments, become tightly associated with care. Customer B might be hiring the same category to "save effort on a Tuesday night" and may associate your brand with weekend occasions, which is a job she is not trying to do mid-week. Same product. Same shelf. Different job. Different fit. The demographic tag was useless. The job and the feeling explain everything.

People do not buy products. They hire products to do jobs, and the way they feel about the job is half of the choice.

The reach trap

Most large brands fall into the same trap. They sell to many segments at once, communicate to none of them in particular, and end up with a kind of grey-cloud brand presence in the consumer's head. The advertising checks the boxes. The packaging is acceptable. Nothing offends anyone. Nothing thrills anyone either.

This is a rational outcome of segmenting on demographics. If your target is "women 25 to 44," your communication has to be vague enough to work for the 32-year-old single professional and the 39-year-old mother of three and the 27-year-old who has just moved out of her parents' house. The vague middle is what you get. The vague middle does not move share.

Brands that win in tight categories tend to do one of two things. They either become extremely sharp about one segment and own it, or they build a brand strong enough at the symbolic level that it can plug into multiple jobs without losing coherence. Both moves require an understanding of what is actually going on in consumers' heads, not just their household composition.

What a useful segmentation looks like

A segmentation that is going to drive real decisions has to do three things. It has to be predictive of behaviour. It has to be addressable, meaning you can reach the people in each segment. And it has to be actionable, meaning you can change what you do as a result.

Demographic segmentation is addressable but not particularly predictive and rarely actionable beyond media buying. Psychographic segmentation is predictive but often hard to address. Behavioural segmentation, built from data on what people actually do, is the most reliable, especially when it is enriched with implicit emotional data showing why each segment behaves the way it does.

In practical terms, a good Caribbean segmentation for a category we have worked in looked something like this. There were four groups. One was small and intensely loyal, hired the brand for ritual and family identity, and would pay a meaningful premium. One was larger and casual, hired the category for convenience, switched on price. One was a "lapsed devotee" group, used to be loyal, drifted away after a packaging change. One was a near-rejector group, associated the brand with an older generation and viewed it as not for them. The four groups overlapped heavily on age, gender, and income. They were almost completely separable on emotional and behavioural variables. Different communication, different distribution, different pricing approaches, every time.

The Caribbean nuance

Regional context complicates and clarifies in equal measure. Caribbean consumers are not a single market. Jamaica and Trinidad and Barbados differ on many dimensions, and within each island the urban and rural pictures differ again. But there are also patterns that run across the region, including a strong orientation toward family and community contexts for purchase, a high salience of music and rhythm in brand association, and a particular sensitivity to authenticity in how brands present themselves locally.

None of these patterns show up in age and income data. All of them show up in implicit and behavioural data. A brand that segments well in the Caribbean is one that has done the work to understand which of these threads runs through which of its customers, and how loud the volume is on each thread for each segment.

The hardest move

The hardest move for an established brand is to stop describing its customers in the language it has used for twenty years and start describing them in the language they actually live in. The internal pressure to keep using the old segmentation is enormous. The slides exist. The targets are agreed. The media plans are built. Everything reinforces everything.

The brands that pull this off tend to do it in small steps. They keep the old segmentation in the room for media purposes. They build a parallel emotional segmentation for strategy and creative. Over a few quarters, the parallel one starts driving decisions. The old one fades into a media-buying tool, which is the one job it was actually good at.

The point of the exercise

None of this is about complexity for its own sake. The point of going beyond demographics is to be useful to the customer in a way that the demographic tag could never describe. Customer A wants to be helped to celebrate her family. Customer B wants to be helped to get through a Tuesday. Both of them are 34 years old in Kingston. Both of them are willing to spend money on the right brand. Both of them will be lost to a competitor that figures out their actual job before you do.

Marketing has been describing people by their birth dates and pay slips for too long. The brain does not work in those categories. The shelf does not either. The shoppers are not waiting for you to catch up. The work is to ask better questions about who they actually are, and to listen for the answers in places other than the answers they give.