Market segmentation: what it is, types and how to apply it with customer data
By Paco Herranz · November 12, 2026 · 12 min read
Market segmentation means dividing a market into groups of people with similar needs, characteristics or behaviour, so you can give each group the offer and message that fits it best. It stops you talking to everyone the same way and puts your budget where it performs best.
In this guide you will see what it is for, the four types of segmentation with examples, how it differs from segmenting your own customers and how to do it step by step with the data you already have.
What is market segmentation?
A market is never uniform. Among "electronics buyers" there are people looking for the lowest price, people who want the latest release and people who only buy during sales season. Segmenting means recognising those differences and grouping the people who are alike in something that matters to your business.
Segmentation comes before two decisions: which segments you target (your target audience) and how you position yourself with each one. Without segmenting, neither the product nor the message has a clear recipient.
Characteristics of good segmentation
Not every group is a useful segment. For market segmentation to help you decide, each segment needs these characteristics:
Measurable
You can know how many people it contains and how much they buy.
Substantial
It is large enough, or valuable enough, to deserve its own action.
Accessible
You can reach it through your channels: if you cannot identify those people or send them anything, the segment stays theoretical.
Differentiable
It responds differently from the others. If two segments react the same way to everything, they are really one.
Actionable
You have something different to offer it: a product, a message, a price or a channel.
Stable for long enough
It does not change from one day to the next, or, if it does, your segments update themselves.
What is market segmentation for?
- Investing better. You concentrate budget on the groups with the most potential instead of spreading it across everyone.
- Communicating better. Each segment gets a message about what matters to it, and that shows in conversion.
- Deciding better. Which products to launch, what prices to set or where to open depend on which segments exist and what they are worth.
The four types of market segmentation
| Type | What it looks at | Common variables | Retail example |
|---|---|---|---|
| Geographic | Where the customer is | Country, region, city, climate, urban or rural area | A different assortment for stores in Monterrey and Mérida depending on the climate |
| Demographic | Who they are | Age, gender, income, household size, occupation | A product line for families with young children |
| Psychographic | How they think and live | Values, interests, lifestyle, personality | A campaign for customers who prioritise sustainability |
| Behavioural | How they behave with your brand | Purchase frequency, average order value, channel, loyalty, product usage | An offer for customers who only buy during El Buen Fin |
The first three describe the person; behavioural segmentation describes what they do. For a business that already has customers, behaviour is usually the best predictor of what they will buy, because it is based on facts rather than assumptions.
Geographic segmentation
It groups people by where they live or shop. It is the simplest and the first one applied in retail with physical stores: assortment, prices and campaigns change by region, city or even each store's catchment area. Its limit is that two neighbours can buy in completely different ways.
Demographic segmentation
It uses objective characteristics of the person: age, gender, income, education, household size or life stage. It is useful for defining products and messages, and it is the most widely used in advertising. Its limit is that it describes who someone is, but does not predict well what they will buy.
Psychographic segmentation
It groups people by values, interests, lifestyle and attitudes. It explains the why behind many purchases (people who prioritise health, seek status or want to save time) and helps with brand tone and communication. It is the hardest to measure: it usually comes from surveys or from inferring interests from behaviour.
Behavioural segmentation
It groups people by what they do with your brand: what they buy, how often, how much they spend, through which channel, whether they respond to promotions, whether they are loyal or only buy in season. It is the best predictor of future behaviour and the one you can act on immediately, because you already have the data.
Types of consumers by behaviour
A practical way to apply behavioural segmentation is to classify your customers by how they relate to your brand. These are the most common consumer types in retail, and what to do with each one:
Loyal
High frequency · high spendThey buy often and spend well. They do not need discounts; they need to feel recognised: early access, priority service, loyalty programme perks.
Occasional
Low frequency · no patternThey buy now and then, with no clear pattern. The goal is to build a habit with reminders at the moment they usually need the product.
Deal hunters
Only on promotionThey only buy on promotion, for example during El Buen Fin or the sales. It is worth knowing who they are so you do not give away margin all year, and send them offers at the right moment.
New
A single purchaseThey have just made their first purchase. The second purchase decides whether they stay: this is the segment where good onboarding pays off most.
At risk
Falling frequencyThey used to buy regularly and have started buying less often. You need to act before they leave; we explain how in our churn guide.
Dormant
No purchases for monthsThey have not bought for a long time. A well-targeted reactivation campaign can win back some of them at a much lower cost than acquiring new customers.
How to choose which segments to target
Segmenting is pointless if you then try to reach every segment with the same intensity. To decide where to put your effort, assess each segment with five questions:
How much is it worth?
Segment size multiplied by average spend or, better, by customer lifetime value.
Is it growing?
A small but growing segment can be worth more than a large, stagnant one.
Is it profitable?
Some segments buy a lot but only on discount, and leave little margin.
Who else is competing for it?
If all your competitors target the same group, winning there costs more.
Does it fit what you offer?
Your product, price and channels must be what that segment is looking for.
Most businesses pick two or three priority segments, with a clear proposition for each, and handle the rest with low-cost actions.
Market segmentation vs customer segmentation
Most businesses use both: market segmentation to acquire and customer segmentation to sell more to the customers they already have.
Behavioural segmentation with data: RFM and CLV
Two methods let you segment your customers by behaviour with data you already have:
- RFM analysis. It ranks each customer by the recency of their last purchase, their frequency and their spend. It is quick and produces very actionable segments: champions, at risk, dormant. We explain it in RFM analysis: how it works and what to do with each segment.
- Customer lifetime value (CLV). It estimates how much each customer will be worth in the future, so you can decide how much to invest in them. See customer lifetime value: how to calculate it and segment beyond RFM.
How to segment step by step with the data you already have
Set the goal
Acquire, raise average order value, retain or reactivate. Each goal calls for different segments.
Bring each customer's data into a single profile
Ecommerce, point of sale, loyalty programme, email. Without unification, the same customer appears several times and segments come out inflated.
Choose the variables
That really change behaviour: frequency, spend, category, channel, tenure.
Create few segments
Between four and eight is manageable; with more, nobody will treat them differently.
Give each segment a specific action
And message.
Measure and update
Customers move between segments; segments should recalculate themselves with every purchase.
Before accepting a segment, test it: does it behave differently from the others in purchases, order value or campaign response? Is it a reasonable size? Can you activate it today in some channel? If the answer to any of these is no, merge it with another or drop it. And when you launch an action, keep a control group within the segment to measure how much it really sells.
Market segmentation with artificial intelligence
Classic segments are defined by hand: someone chooses the variables and the cut-offs. With enough customer data, artificial intelligence adds two things:
Discovering segments you had not looked for
Clustering algorithms find groups of customers who are alike across many variables at once, for example customers who buy little but always from high-margin categories, or who only buy in store after browsing online.
Segmenting by what will happen
Predictive models calculate each customer's probability of buying, churning or responding to an offer. With that you can build segments such as "high propensity to buy this category next month", which no manual cut-off achieves.
The condition, once again, is unified data: a model trained on duplicate customers or on half of their purchases produces segments as incomplete as the data.
Five segmentation examples
- Fashion retail. Customers who only buy in the sales versus customers who buy new arrivals at full price. The first get a sales alert; the second get early access to the collection, with no discount.
- Pharmacy. Customers who buy a product regularly versus occasional buyers. Regulars get a restock reminder; occasional buyers get seasonal care content.
- Restaurants. Weekday lunch customers versus weekend family customers. Different offers at different times.
- El Buen Fin season. Customers who only buy during sales events, regular customers and new customers acquired during the campaign. The third group gets a plan for their second purchase before they are lost.
- Loyalty programme. Active members, members with unused points and inactive members. Those with unused points get a reminder before they expire.
Common segmentation mistakes
- Segments that are too small or too many segments: nobody can prepare twenty different campaigns.
- Segments that are never activated: they look good in a presentation, but have no action attached in any channel.
- Segments that are never updated: calculated once in a spreadsheet and used for months.
- Segmenting only by demographics when you have behavioural data, which predicts much better.
Frequently asked questions about market segmentation
What is market segmentation? It is dividing a market into groups of people with similar needs, characteristics or behaviour, so you can give each group the offer and message that fits it best.
What is market segmentation for? To invest budget where it performs best, adapt the message to each group and make better product, pricing and channel decisions.
Why is market segmentation important? Because no market is uniform: treating everyone the same means the message does not quite fit anyone and the budget goes where it does not perform. Segmenting lets you choose where to compete and with what proposition.
What characteristics should good segmentation have? Each segment must be measurable, large or valuable enough, accessible through your channels, different from the others in its behaviour and actionable, meaning you have something different to offer it.
What are the types of market segmentation? The four main ones are geographic (where they are), demographic (who they are), psychographic (how they think and live) and behavioural (how they behave with your brand).
What is a customer segment? It is a group of your current customers who behave alike, for example by purchase frequency or spend, and to whom you can direct a specific action.
What is behavioural segmentation? It groups people by what they do: what they buy, how often, how much they spend and through which channel. For a business with customers it is usually the most useful, because it is based on facts.
How FLYDE does it
FLYDE brings each customer's data from ecommerce, point of sale, loyalty programme and email into a single profile. With Market you build segments by combining any variable, such as RFM, value, category or channel, and activate them in your channels in one click. Segments update themselves with every purchase, and customer profiling helps you discover groups you had not looked for.
If you want the segmentation models with the most impact in retail, download the whitepaper "The best customer segments for retail": eight models, from RFM to predictive, with the signal, the KPI and the most common mistake for each.