Segmenting by need, not just demographics

Most customer segmentations start with demographics: age bands, income brackets, household types, regions. For business customers, the equivalent is firmographics: company size, sector, turnover. These are easy to collect and easy to target in media plans. They are also, on their own, poor guides to what customers actually want.

Two 45-year-old homeowners on similar incomes can approach the same purchase in completely different ways. One wants the cheapest reliable option with no fuss; the other wants expert advice and is happy to pay for it. Two manufacturing firms of the same size may value a supplier for entirely different reasons. Demographics describe who customers are. Needs describe why they buy, and that is what shapes products, messages and service.

What a needs-based segment looks like

A needs-based segment groups customers by what they are trying to achieve and what matters most to them when they choose. It describes the job they want done, the outcome they value and the trade-offs they are willing to make.

Consider a garden centre. A demographic view might divide customers into retirees, young families and new homeowners. A needs-based view might find quite different groups: keen gardeners who want rare plants and expert advice; tidy-uppers who want a presentable garden with minimal effort; social visitors for whom the café and a pleasant browse are the main attraction; and project buyers tackling a one-off job such as a new patio. Retirees appear in every one of those groups. So do young families. The needs, not the age, determine what each group values and how the business should serve them.

Start with what customers are trying to achieve

Needs come from understanding customers, not from a workshop of assumptions. Customer interviews are the best starting point: conversations about real recent purchases, what prompted them, how people chose and what they would do differently. Listen for differences in goals, priorities and frustrations rather than differences in who people are.

From the interviews, draft a set of candidate needs and the statements that capture them, in customers’ own words where possible. “I want it sorted quickly without having to think about it.” “I want to understand the options before I commit.” “I want the best, and I will pay for it.” These become the raw material for the segments.

To test how common each need is, and how needs cluster together, follow the interviews with a survey of a larger sample. Ask customers to rate the importance of each need and describe their recent behaviour. Statistical clustering can then suggest natural groupings, though judgement is always required to turn clusters into segments the business can use.

Combine needs with behaviour

Needs on their own can be hard to act on. You cannot easily target someone in a marketing system because they “want to understand the options”. The answer is to link each needs-based segment to observable behaviour: what people buy, how often, through which channels, at what price points and with what response to promotions.

That link allows existing customers to be assigned to segments using transaction data, and new customers to be recognised from early behaviour. A garden centre’s keen gardeners might be identified by purchases of specialist plants and attendance at workshops; tidy-uppers by bulk purchases of compost and ready-grown bedding at weekends in spring. The assignment will not be perfect, but it is usually good enough to tailor communications and offers.

• Purchase patterns: categories, frequency, basket size and price points.
• Channel preferences: online, in store, phone, through an account manager.
• Service use: advice requested, returns, support contacts.
• Response to offers: discounts, bundles, loyalty rewards, new products.
• Demographic and firmographic data, used as supporting description rather than the basis of the segment.

Name segments so people remember them

A segmentation that lives in a slide deck changes nothing. To shape decisions, it has to become part of how people talk. Short, descriptive names help enormously: “keen gardeners”, “tidy-uppers”, “social visitors” and “project buyers” are easier to remember and discuss than “Segment C” or a cluster number.

Support each name with a one-page profile: the core need, what they value and what frustrates them, how they behave, how to recognise them in data and what share of customers and revenue they represent. Add a short illustrative description of a typical customer, clearly labelled as illustrative, so the segment feels human. Avoid names that stereotype or patronise; they tend to leak to customers eventually.

Put segments to work

The test of a segmentation is whether it changes decisions. Once segments are defined, work through the main areas of the business and ask what each segment needs that it is not getting.

• Products and ranges: which segments are well served, which are neglected, and where new offers might fit.
• Marketing messages: what each segment wants to hear, through which channels, at which moments.
• Service design: how much advice, speed or personal attention each segment values.
• Pricing: where segments are price-sensitive and where they value quality or convenience more.
• Measurement: tracking growth, retention and satisfaction by segment, not just in total.

Not every segment deserves equal investment. Some are larger, more profitable or growing faster. Others may be expensive to serve for little return. A clear segmentation makes these choices explicit, so the business can decide deliberately where to focus rather than trying to please everyone equally.

Once we stopped talking about age groups and started talking about what customers were trying to do, our product and marketing teams finally agreed on priorities.

Marketing director, home improvement client

Keeping segments current

Needs change more slowly than behaviour, but they do change. New competitors, economic conditions and technology shift what customers value. Review the segmentation every year or two, refresh the profiles with new interviews and check that customers are still being assigned accurately. A segmentation that no longer reflects reality is worse than none, because it gives false confidence.

To begin, take your current segmentation, or your implicit one, and ask whether it explains why customers choose you. If not, run ten to fifteen interviews focused on recent purchases, draft candidate needs, and test them with a survey. Settle on four to six segments, give them memorable names, link them to behaviour you can observe and pick one decision, such as next season’s marketing plan, to shape with them first.

Source: Intelliversal analysis.

Frequently asked questions

Should we abandon demographic data?

No. Demographics remain useful for describing segments and for media targeting. They work best as supporting information alongside needs and behaviour, not as the basis of the segments themselves.

How many customer segments should we have?

Usually four to six. Enough to capture meaningful differences, few enough for teams to remember and act on.

How do we assign customers to needs-based segments?

Link each segment to observable behaviour such as purchase patterns, channel use and response to offers, then use transaction data to assign customers. Refine the rules as you learn.

Explore Consumer behaviour analysis