What STP is
The three-step sequence is most often associated with Philip Kotler's marketing-management textbooks, though the component ideas are older and have several authors. The framework is a sequence rather than a list: each step takes the previous one as its input. Segmentation produces a map of the market; targeting chooses which parts of the map to pursue; positioning decides what the chosen customers should think the offer is, relative to the alternatives they would otherwise consider.
What STP makes visible is choice. A business that has not run the sequence is usually serving whoever turns up, with a message written for nobody in particular and a price set by habit. Running it exposes the implicit decisions that have already been made and asks whether they are the right ones. It also makes trade-offs explicit: a target segment is a decision not to serve others, at least not with the same offer, and positioning for one group is a decision to be less attractive to another.
What the framework hides is that the segments are not found, they are drawn. Two analysts given the same data will divide a market differently, and the division that looks natural in a spreadsheet may not match how anyone actually buys. It also freezes a moving picture: segments shift, competitors reposition, and a map that was accurate at the last planning cycle may be out of date by the next. STP is a way of making decisions, not a description of the market that can be made once and filed.
The three steps
| Step | Question it answers | Output | Criteria for a good answer |
|---|---|---|---|
| Segmentation | How does this market divide into groups that want different things? | A small set of named segments with the basis for the split stated | Segments differ in needs or behaviour, can be identified in data, are large enough to matter and can be reached separately |
| Targeting | Which of these segments will we serve, and with how much of our effort? | A ranked choice of target segments and the targeting pattern (one, several, or all with one offer) | The segment is attractive on size, growth, competition and economics, and the business can serve it better than the alternatives can |
| Positioning | What should each target segment believe the offer is, relative to what they would otherwise choose? | A positioning for each target: frame of reference, point of difference, reasons to believe | The position is true, different from the alternatives the segment considers, and matters to that segment |
The table reads top to bottom for a reason. A positioning chosen before targeting describes the company rather than a customer's choice, and a target chosen before segmentation is usually the segment the company already serves by accident, renamed as a decision.
Segmentation criteria
A segment is a group of buyers who respond differently from other groups to some element of the offer: the product, the price, the channel or the message. The basis for the split can be who they are (demographics, firmographics, location), what they do (purchase behaviour, usage, channel habits), or why they buy (needs, jobs to be done, the situation that triggers the purchase). Needs-based and behavioural splits are usually more useful than descriptive ones, because they predict response; a segment defined by age tells you who the people are but not what they want.
The classic test of a segment asks whether it is measurable, substantial, accessible, differentiable and actionable. In practice the two that matter most are differentiable, meaning the segment would react differently to a changed offer, and accessible, meaning the business can reach it without also paying to reach everyone else. A segment that fails either test is a label, not a strategy. The segmentation guide covers the bases and the method for building segments from data, and the audience research guide covers where the evidence for a split comes from; this page stays at the level of the decision.
Targeting choices
Targeting has two parts: evaluating segments and choosing a pattern. Evaluation weighs each segment's attractiveness against the business's ability to serve it. Attractiveness covers size, growth, how many competitors already serve it and how well, and the economics of serving it, which for a digital business reduces to the relationship between acquisition cost (CACCost to acquire one new customer.) and customer lifetime value (LTVTotal value a customer brings over time.). A large segment that is expensive to acquire and quick to churn is less attractive than a smaller one that stays. Ability to serve covers product fit, channel access, credibility with the segment and the organisation's capacity to run more than one offer at once.
The pattern is the shape of the choice. An undifferentiated pattern serves the whole market with one offer, which suits commodities and very early businesses. A differentiated pattern serves several segments with a tailored offer each, which is most mature businesses. A concentrated pattern serves one segment deeply, which is the niche strategy and the usual right answer for a small business entering a market with established players. At the extreme, targeting narrows to individual accounts, which is what account-based marketing does: the segment becomes a named list, and the targeting decision is how much personalised effort each tier of the list receives. Whatever the pattern, the test is the same: can the business serve the chosen segments distinctly, or is it pretending to differentiate while sending everyone the same thing?
Positioning
Positioning is the decision about perception: what the target segment should think the offer is, compared with the alternatives they would otherwise choose. It has three parts. The frame of reference is the category the segment places the offer in, which decides what it is compared with. The point of difference is the claim that makes the offer preferable within that frame. The reasons to believe are the evidence that the claim is true: product facts, proof, reputation.
Positioning is done per target segment, which is where STP earns its sequence. The same product can be positioned as the premium choice for one segment and the convenient choice for another, provided the two segments do not compare notes. The positioning guide covers how to find a defensible position in a crowded category, and the positioning statement framework covers the sentence that records the decision. This page's concern is narrower: that the position is chosen for a segment that was deliberately targeted, from a map that was deliberately drawn.
Worked example
A regional accountancy firm runs STP before a website rebuild
A twelve-partner accountancy firm in the Midlands serves anyone who asks and is about to rebuild its website. Before writing a word, the partners run STP. Segmentation by need rather than by company size produces four groups: owner-managed trading businesses wanting compliance done invisibly, growing firms wanting management accounts and a fractional finance director, contractors wanting the cheapest compliant option, and high-net-worth individuals wanting tax planning. Targeting weighs each: contractors are the largest group but bring an average fee of £900 a year and churn after 18 months on average, so the lifetime value of £1,350 barely covers the £1,100 it costs to win one. Growing firms number 40 clients, pay an average of £14,000 a year and stay for over 6 years. The firm chooses growing firms as the primary target and owner-managed businesses as the secondary one, declines to compete for contractors, and positions itself to growing firms as "the finance function you would hire at twice the size", with the fractional finance director service as the point of difference and three named client stories as the reasons to believe. The website is then written for two readers rather than for everyone.
Limits
STP assumes the market can be divided into stable, reachable groups, and that is sometimes untrue. Markets where needs change with context, where the same person is a different buyer on Monday and on Saturday, resist segmentation by person; the jobs-to-be-done view in this library is the usual correction. The framework is also only as good as the data behind the split, and the data available tends to be descriptive (age, location, company size) rather than motivational, so segments drift towards what can be counted rather than what predicts behaviour.
The sequence is static. It is run at a planning moment and recorded, but segments shift as the market matures and competitors reposition in response to each other, so a position that was distinctive at the last review may now be shared. Nothing in STP prompts the review; the discipline has to come from the planning cycle. STP also says nothing about execution: a correct target and a sharp position are wasted if the channels, the product and the price do not follow, which is what the marketing mix checks. And digital advertising has created an illusion of segmentation: an ad platform's audience is a targeting mechanism, not a segment, and choosing one in a campaign is not the same as choosing whom to serve.
Finally, the framework rewards narrowing, and narrowing can be overdone. A segment so specific that it cannot support the business, or a position so distinctive that nobody is looking for it, is the framework working exactly as designed and still producing a bad strategy. The KPIA metric tied directly to an objective. that catches this is simple: whether the chosen segments, served as positioned, can produce enough revenue.
For a product with no customers yet, the guide to running STP for a new product launch works from hypotheses and a single beachhead segment. Turned inward, STP for customer retention segments the existing base by value and adds a decision about whom to let go.
Common mistakes
- Segmenting by what is easy to count. Age bands and company sizes are available in every dataset and rarely explain why anyone buys. Segments built on needs or behaviour are harder to draw and worth more.
- Targeting the segment already served. The largest current customer group becomes the "target" without any evaluation, so the exercise confirms the accident of history rather than testing it.
- Choosing every segment. A plan that targets all four segments with four messages from a team that can run one is a concentrated strategy pretending to be a differentiated one.
- Positioning for the company, not the segment. "Trusted, professional, innovative" describes the firm's view of itself and gives the segment nothing to compare.
- Mistaking a platform audience for a segment. An interest category in an ad account is a way of reaching people, not a decision about which people to serve.
- Running STP once. Segments move and competitors reposition; a map from several years ago is a historical document, not a strategy.
- Ignoring the economics. A segment that is large and reachable but costs more to acquire than it ever pays back is attractive on every criterion except the one that matters.
