Market segmentation research: find, size and target the segments that matter

Questions this report answers

• Which distinct groups of buyers exist in our market, and what really separates them?
• How large is each segment, and what share of category spend does it control?
• Which segments are growing, and which are already well served by competitors?
• What does each segment expect from the product, the price and the service?
• Which segments should we target first, given our strengths?
• How can sales and marketing recognise a segment member in the CRM or in media buying?
• Which messages and channels work best for each priority segment?

When you need it

Segmentation earns its keep when a single proposition no longer fits everyone. Typical triggers:

• Growth has stalled and every campaign speaks to "the average customer", who does not exist.
• A new proposition or brand architecture needs a clear target, and teams disagree about who it is.
• Value tiers in the CRM tell you who spends most but not why, so retention offers miss the mark.
• Several countries are in play. Eurostat's 2024 price level index for restaurants and hotels was 124.6 in the Netherlands and 53.2 in Bulgaria (EU27 = 100), so a segment labelled "price-sensitive" can mean very different things across borders.

In B2B, firmographics such as company size or sector describe accounts but rarely explain buying behaviour; needs-based segments of decision-makers are usually more useful, and our B2B research page covers how to reach them.

Segmentation is not the right first step if you cannot act differently for different groups, for instance with one product sold through one channel. If you have no data on the category at all, start with consumer research to understand habits, then segment.

Sample table of contents

1. Executive summary
   the segment map on one page and the recommended targets
2. Approach and sample
   basis variables, fieldwork, statistical method
3. Market overview
   the category size used as the base for sizing
4. Choosing the solution
   how many segments and why, with stability tests
5. Segment profiles
   needs, attitudes, behaviour, demographics, media use
6. Size and value
   share of buyers, share of spend, expected growth
7. Brand position by segment
   awareness, usage and share of wallet
8. Prioritisation
   attractiveness against your ability to win
9. Personas
   a narrative portrait of each segment, drawn from interviews
10. Typing tool
    a short question set that classifies new customers
11. Activation plan
    proposition, channels and messages per target segment

Methodology and sources

What to segment on

Demographics describe segments; needs and attitudes explain them. We therefore build segments on needs, attitudes and behaviour, and use demographics and media habits to profile and find them.

Qualitative groundwork

Ten to twenty in-depth interviews generate the needs statements and trade-offs in buyers' own words, which keeps the survey from reflecting only internal assumptions.

Survey design

A CAWI survey (CATI for many B2B audiences) includes 30–50 attitudinal statements and behaviour, spend and media questions. MaxDiff often replaces rating grids, because forcing a choice between statements separates people better than agreement scales. A practical rule is at least 100 respondents per expected segment, so a five- or six-segment solution needs roughly 800–1,200 interviews. Our sample size calculator helps you check the precision for each segment.

Analysis

Factor analysis reduces the statements to underlying dimensions. Cluster analysis, using k-means or latent class models, then produces solutions of three to eight segments, compared on statistical fit, size, clarity and whether the business can act on them. A discriminant model or decision tree becomes the typing tool, with its classification accuracy reported.

Quality

Straightliners are removed with particular care, because identical answer patterns create artificial clusters.

What you receive

• Segmentation report as PDF and an editable deck
• One-page persona sheet per segment
• Excel sizing model linking segments to category value
• Typing tool: short questionnaire plus scoring algorithm with accuracy figures
• Respondent-level data with segment labels
• Activation workshop with marketing, sales and product teams

How the project runs

1. Kick-off and mapping the decisions the segments must support — 1 week
2. Qualitative interviews — 1–2 weeks
3. Questionnaire design and pilot — 1 week
4. Fieldwork — 1–2 weeks
5. Cluster analysis and choice of solution — 1–2 weeks
6. Personas, typing tool and report — 1 week
7. Activation workshop — 1 day

Common mistakes and how to avoid them

• Segmenting on demographics alone. People of the same age and income often want very different things; use needs and behaviour, then describe with demographics.
• Choosing the best statistical fit over the most usable solution. A segmentation nobody can explain in one sentence per group will not be used.
• Too many segments. Teams can work with four to seven; beyond that, merge or nest them.
• No typing tool. Without a way to classify customers, the segments stay in a slide deck.
• Treating segments as permanent. Refresh the solution after a major market shift or every few years.

Who it suits

Updated: 5 Nov 2026