Customer segmentation built for the teams who use it.

Segments shaped by the decisions they serve, each one sized, a tool that finds them in your data, and working sessions that put them to use across your teams.

View segmentation case studies

Four steps, each with a test the segments must pass

  1. Build

    The segments differ in what people need, how they buy and what they’re worth.

  2. Size

    Each segment has a size, so the plan can show where growth could come from.

  3. Find

    Each segment can be found again, in a later survey or, where they hold the right variables, in your customer records.

  4. Share

    Each segment is described in words every team uses, so everyone works from one picture.

An illustration, not a client project.

A drawing: a market of people sorted into five groups on a sheet, with a bar sizing each group; a question card placing one person; and the group profiles fanned out, one marked as the team’s first.

Each step ends with something your teams can use.

A segmentation is useful when your teams can recognize each segment, find it in their data and act on it.

Build

We agree with your team what the segments are for, because the use sets what they’re built on: needs, value or media habits.

Where needs aren’t yet well understood, qualitative work comes first. Our data scientists then compare solutions with different numbers of segments, and you choose with us.

The segments, described in words your teams use, with the survey and the models behind them.

Size

A survey built to represent your market sizes each segment, in every market the plan covers. Your own sales or customer data can add what each segment is worth.

A size for each segment in each market, and its value where your data allows.

Find

A typing tool places people in segments: a short set of questions for future surveys, or a model that works from the records you already hold. Before anyone relies on one, it needs:

  • the variables it was built on, asked or recorded the same way;
  • a check, segment by segment, on people it wasn’t built from;
  • a route for people with little on record, such as new customers;
  • care with outside data, which places people less exactly than your own records.

The typing tool, with a record of how well it places people, segment by segment.

Share

Your team names the segments with us and chooses which to put first. Then we take them to the teams who’ll use them, in sessions built around their own decisions.

Segment profiles, a report for the wider business, and working sessions with the teams who’ll use them.

A segmentation in use

Every member in a segment, from data the venue already held.

A European sports venue sells membership to individuals and companies. Each team, from account managers to marketing, had its own picture of the members.

  1. From the members’ answers

    A member survey came first, and its answers alone grouped members by their needs and motivations.

  2. From the venue’s own records

    With the venue’s data team, we rebuilt the segments from fields held for every member, such as membership type and event attendance, judging each way by how closely it reproduced the needs-based segments.

  3. Into the business

    A needs-based segmentation, a tool that places every member from the venue’s own data and a report, then a closing workshop that took the segments into the teams that serve members.

What happened next

The client changed its email schedule by segment, and reports that the segments informed its brand positioning and its planning of events.

Segmenting a venue’s members from its own records

Another of our segmentation programs: EXPLOR Award winner, Insights Association, for a segmentation and market model, voted by the Corporate Researchers Conference.

Before you start a segmentation.

What sample size does a segmentation need?

There’s no single number. It depends on how many variables build the segments, how distinct they turn out, the smallest segment you need to read and how many markets you compare. We plan from the smallest group the work must read in each market, and boost the sample where a small segment matters.

Can a segmentation and a choice model run in one survey?

Yes, when the segments need to inform price or the range. The same people answer the needs questions and the choice tasks, so the trade-offs can be read segment by segment. The sample is sized for the smallest segment the choice model must read, which usually pushes it up.

Conjoint and choice modeling

When does a segmentation need refreshing?

When the segments stop predicting behavior. Common triggers are a shift in the category, the channels or the customer base, segment sizes drifting, or a typing tool that places fewer people correctly. It helps to set the triggers at launch and check them at each planning cycle.

Who builds the segmentation?

Our own data scientists and statisticians build the segments, the sizing and the typing tool. Our qualitative researchers run the interviews and ethnography that shape the survey or bring segments to life.

The partners

Our thinking on segmentation and customer understanding

Starting a segmentation, or rebuilding one?

Bring the question your segments need to answer. If you have an older segmentation, we’ll suggest what we would keep and what we would rebuild.

A partner or director will reply.