How to brief a segmentation your business will actually use

In our first piece, “before you commission a segmentation, ask whether you actually need one“, we argued that the business challenge should come before the methodology. Then, in You know what your customers do. But do you know why?, we looked at the kind of understanding a segmentation needs to create, and why behaviour alone does not always tell the whole story.

Once you know that segmentation is the right approach, and you have a clearer idea of what you need to understand, the next question is how to set the project up so the resulting segmentation is something the business can actually use.

That starts long before fieldwork, and in our experience, it starts with the brief.

A good segmentation brief needs to do more than describe the market you want to understand or list the questions you would like the research to answer. It should give everyone involved a clear sense of why the work matters, what the business hopes to do differently as a result, and who will eventually need to use it.

That matters because segmentation projects have a habit of accumulating ambition. They are significant pieces of research, often involving multiple teams and markets, so it is understandable that people want to make the most of them. The danger is that the brief gradually becomes a home for every unanswered customer question in the organisation.

The strongest briefs create enough focus to make good choices about what the segmentation really needs to explain, while leaving enough space for the research to uncover something new.

What should a segmentation brief actually do?

A segmentation brief should create clarity around the business problem before it starts prescribing the research solution.

We often receive briefs that contain a clear request for a segmentation, but less detail on the decisions sitting behind it. There may be a desire to understand customers better, identify priority audiences or refresh an existing model. All are reasonable objectives, but they still leave an important question unanswered: what will the organisation be able to decide once it has that understanding that it cannot decide confidently today?

Different decisions can require very different segmentations.

A business looking for growth audiences may need a broad view of the market and the needs that separate one opportunity from another. A team wanting to personalise communications may place more weight on whether segments can be identified within existing data. Innovation may need a stronger understanding of unmet needs and occasions, while a portfolio team may be interested in where audiences overlap and where different propositions have permission to play.

It is also important to be clear about who the segmentation is intended to represent. Are we looking at current customers, category users, prospective customers, the wider market, or a more specific audience? That choice can fundamentally shape the opportunities the research is able to see, so it is worth resolving before the work begins.

These are not simply different ways of using the same answer. They can affect what needs to be explored, what data matters and how the eventual segmentation should be evaluated.

The brief therefore needs to establish the job the segmentation is there to do before worrying too much about what the finished model will look like.

Start with the decisions you need to make

One useful question at the beginning of a project is: if this segmentation works brilliantly, what decisions will become easier?

The answer might involve where to focus investment, which audiences to prioritise, how to develop different propositions, where future growth could come from, or how to create a more consistent view of customers across the organisation.

Being explicit about those decisions gives the research something to work towards and helps distinguish what is genuinely important from what would simply be interesting to know.

That becomes particularly useful when the brief starts to expand. Marketing may want the segmentation to shape communications, innovation and product teams may want it to reveal unmet needs, customer teams may want to connect it to CRM, while senior leadership wants something simple enough to use when making investment choices.

None of those needs is unreasonable. The challenge is deciding which are fundamental to the segmentation, and which can be explored around it.

Trying to satisfy every possible use case in the core model can create a segmentation that is technically rich but strategically muddy. The clearer the organisation can be about the decisions that matter most, the easier it becomes to design something that genuinely supports them.

Who needs to be involved?

Segmentation rarely belongs to one team once it reaches the business, so it is risky to design it as though it does.

Insight may commission the work, but marketing, innovation, product, strategy, customer experience, data teams or senior leadership may all need to make decisions from it later. If those people only encounter the segmentation at the final presentation, it may be too late to discover that they needed something different from it.

Bringing the right stakeholders into the conversation early gives us a fuller understanding of the decisions the work needs to support. Different teams often see different parts of the same challenge, and those conversations can expose assumptions or tensions that need resolving.

It also creates ownership. People are more likely to trust and use a segmentation if they understand why particular choices have been made and can see how the work connects to the problems they are trying to solve.

That does not mean designing the research by committee. The aim is to understand the different needs, then make clear choices about what the segmentation should and should not be expected to do.

Be clear about what does not belong

There is usually a point in a segmentation project where somebody says, “While we have everyone in the survey, could we also ask about…?”

Sometimes the answer is yes. But enough additions can quickly turn a focused segmentation study into a large customer survey that happens to contain a segmentation somewhere inside it.

Every extra objective has consequences. Questionnaires get longer, the analytical possibilities multiply and it becomes harder to see which measures are genuinely important to the segmentation itself.

The same applies to the model. Just because a variable is available does not mean it should help define the segments. As we discussed in “You know what your customers do but don’t know why”, the inputs need to earn their place. The measures at the heart of the segmentation should represent differences that matter for the decisions the business needs to make, rather than simply reflecting whatever data is easiest to collect.

A good brief creates a clear enough definition of success that we can challenge complexity when it is not adding much value.

Think about how people will use it before you build it

One of the most important parts of the briefing conversation is understanding what “use” will actually look like.

If the segmentation is intended to shape strategy, senior teams may need a relatively simple framework that allows them to compare opportunities and make choices. Marketing teams may have different requirements around recognisability, targeting and media activation. If the business wants to connect the segmentation to its customer database, allocation becomes important and the research needs to collect the right information to make that possible.

These considerations should influence the design from the beginning. They are much harder to bolt on once the segmentation already exists.

That does not mean activation needs to be fully mapped out before the research starts. We still want room for the work to surprise us and for new uses to emerge. But there is a big difference between leaving room for discovery and leaving the question of use until the end.

The aim is to understand enough about how the segmentation will live inside the organisation to make sensible decisions about its design now.

It is also worth thinking beyond the immediate decisions the segmentation needs to support. Over time, new questions will emerge, new data will become available, and different teams may need to use the framework in ways that are difficult to predict at the outset. Having an early view of what might happen after launch, whether that involves allocation, activation, connecting other data sources or further research, can help create a segmentation that is useful now without boxing the organisation in later.

What should you agree at kick-off?

By the time the programme begins, there should be a shared understanding of the business challenge, the priority decisions, who needs to use the segmentation and what success will look like.

Success does not mean deciding in advance that there must be five segments or specifying what each should look like. Being too attached to a particular answer before seeing the evidence rather defeats the point.

It is more useful to define success in terms of what the segmentation needs to enable. Perhaps it needs to reveal genuinely different growth opportunities. Perhaps, teams need to recognise and use the segments consistently across markets. And maybe the framework needs to connect with existing customer data or support decisions across brand, innovation and experience.

Agreeing those criteria early is particularly helpful when the analysis produces more than one statistically viable answer, which it often will. The strongest solution is not necessarily the cleverest model. It is the one that best balances meaningful differences, analytical confidence and the practical job the organisation needs it to do.

How much prescription should be in the brief?

There is a balance to strike. A vague brief makes it difficult to design the right programme, but an overly prescriptive one can lock the research into assumptions it should really be testing.

The best briefs tend to be clear about the problem and flexible about the answer.

They tell us what is happening in the business, what decisions are coming up, where the uncertainty lies and what needs to become clearer. They should also set out constraints that genuinely matter, such as markets, existing data, business systems or requirements for future use.

What they do not need to do is dictate every methodological choice in advance.

If we already know which variables should define the segmentation, how many groups we expect to find and what each one ought to look like, we are not leaving much room for the research to tell us anything new.

A good brief should give the work direction without predetermining the answer.

So, how do you make a segmentation actionable?

Being actionable starts much earlier than the final debrief.

It comes from being clear about the decisions the segmentation needs to support, involving the right people, being disciplined about what goes into the study and thinking realistically about how the eventual framework will be used.

That is why actionability should not simply mean “easy to understand”. A segmentation can be beautifully brought to life and still be difficult to act on if it was designed around the wrong questions. Equally, a statistically sophisticated model can struggle to gain traction if nobody can see how it connects to the decisions they make.

Getting the brief right helps make the eventual segmentation more actionable because the project has been designed around real decisions from the outset. Teams have a clearer understanding of what the framework is for, where they expect to use it and what it needs to enable.

That does not guarantee the segmentation will land successfully, but it gives the project a much better chance of producing something that connects naturally to the decisions people need to make.

A good brief is the beginning, not the end…

A strong segmentation brief gives the research direction. It helps clarify the business challenge; decide which questions matter and design the work around the people who will eventually use it.

But even a well-designed segmentation reflects what we know at a particular point in time. Markets change, new behaviours emerge, categories evolve and organisations ask new questions. If the segmentation is going to remain useful, the understanding around it needs to keep developing too.

That means thinking beyond the initial project. Rather than treating the final segmentation as a finished piece of research, the opportunity is to use it as a framework that new learning, data and experience can continue to build upon.

That is the question we’ll explore next in Why segmentation should be a living framework, not a one-off project: how do you keep a segmentation relevant and useful long after the initial work has been delivered?