A market segmentation analysis breaks your total market into smaller groups of people who share similar needs, then helps you decide which groups to focus on. The process moves through five core phases: grouping potential buyers into segments, categorizing your products, estimating the size of each segment, selecting your targets, and building marketing actions around them. Here’s how to work through each phase with real precision.
Define Your Objective Before You Segment
Before sorting anyone into groups, get clear on why you’re segmenting. Are you trying to grow by reaching new customers, or are you trying to increase revenue from existing ones? The answer shapes everything downstream, from which data you collect to which variables matter most. A company focused on growth needs to study people who aren’t yet buying. A company focused on retention should dig into its current customer base. Trying to do both at once with a single segmentation model usually produces a result too vague to act on.
If your buyers and end users are different people (a parent buying a product for a child, or a procurement manager choosing software that employees will use), your segmentation needs to account for both. Segmenting only end users when someone else controls the purchase decision is a common reason projects fail to produce useful results.
Choose Your Segmentation Variables
Segmentation variables are the characteristics you use to sort people into groups. They fall into four main categories, and most useful segmentation analyses combine variables from more than one.
- Demographic: age, gender, income, education level, family size, marital status, race. These are the easiest to measure and the most widely available in public data.
- Geographic: urban or rural location, climate, proximity to your business, regional preferences. This matters most for businesses with physical locations or products that vary by region.
- Psychographic: lifestyle choices, personality traits, personal values, opinions, interests. These reveal why people buy, not just who they are.
- Behavioral: purchasing history, product usage frequency, brand loyalty, how far along someone is in deciding to buy, whether they respond to promotions or pay full price.
The variables you pick should directly relate to how people make purchasing decisions in your category. This is where many analyses go wrong. One energy provider built an entire segmentation around household lifestyle values, only to discover those values didn’t predict energy usage or service preferences at all. The variables looked interesting but had no connection to actual buying behavior. Always ask whether a variable helps explain how people choose or use what you sell.
Gather Your Data
You need two types of data: information you already have and information you need to go find.
Internal data comes from your CRM, sales records, website analytics, customer service logs, and transaction history. This is your richest source for behavioral variables like purchase frequency, average order size, and product preferences. If your segmentation needs to map back to your customer database for personalized marketing, building it on variables that actually exist in your CRM is essential.
External data fills in the gaps. Government sources like the Census Bureau provide demographic and geographic data at no cost. The Bureau of Labor Statistics publishes detailed income and employment data. Specialized sources exist for specific populations: the Library of Congress maintains a consumer research guide listing data portals for various demographic groups, generational research organizations, and regional equity data. Market research firms also sell proprietary segmentation data, though these typically require a subscription.
Primary research, meaning surveys, interviews, and focus groups you conduct yourself, is the main way to collect psychographic and attitudinal data. A survey of 500 to 1,000 people in your target market can reveal values, preferences, and decision-making patterns that no existing database captures. The cost and effort are real, but psychographic insights often produce the sharpest segments.
Build Your Segments
With data in hand, you’re grouping potential buyers into clusters. Each segment should pass two tests simultaneously: people within the same segment should share similar needs and behaviors, and people in different segments should be meaningfully different from each other. If two segments look nearly identical in what they want and how they buy, combine them. Every additional segment adds cost because each one eventually needs its own marketing approach.
Five criteria help you decide whether a segment is worth creating:
- Profit potential: Segmentation costs money. A segment only justifies its existence if targeting it can lead to higher revenue or margins than a one-size-fits-all approach.
- Internal similarity: Members of the segment should respond to the same marketing message or product offering in roughly the same way.
- External difference: The segment must be distinct enough from other segments that it genuinely requires a different approach.
- Reachability: You need a practical way to get your message or product in front of this group. A perfectly defined segment you can’t reach through any available channel isn’t useful.
- Simplicity of assignment: You should be able to look at a potential customer and determine which segment they belong to without excessive cost or guesswork. If the criteria are so complex that classification becomes its own research project, the segmentation is too granular.
Resist the urge to let internal opinions drive the groupings. Every sales team and executive has a mental model of “who our customers are,” often based on anecdote and pattern recognition. Those instincts can be useful starting points, but basing your segments on stereotypes rather than data leads to strategies built on assumptions that may not hold up.
Create a Market-Product Grid
A market-product grid is a simple matrix. Your customer segments go down the rows. Your product or service categories go across the columns. Each cell represents a potential combination: this type of customer buying this type of product.
For each cell, estimate the market size. How many people are in that segment, and how much might they spend on that product category? Some cells will be large and obvious. Others will be small or empty. The grid gives you a visual map of where the opportunity concentrates, and it often reveals combinations you hadn’t considered. Maybe a segment you thought of as secondary actually represents significant spending potential in a product category you’ve underinvested in.
You don’t need perfect numbers here. Reasonable estimates based on your data, industry benchmarks, and public demographic figures are enough to compare segments against each other and prioritize.
Validate Your Segments
Before you commit resources, test whether each segment holds up against three practical standards.
First, is it measurable? You should be able to estimate how much a segment will spend. If you can’t connect the segment’s defining characteristics to actual purchasing behavior, the segmentation may be intellectually interesting but commercially useless.
Second, is it accessible? Understanding a group and being able to reach them are different things. A segment of older retirees may not respond to mobile app notifications but might engage with print advertising. A segment of young professionals might be reachable through social media but ignore direct mail. Your segments need to align with channels you can actually use.
Third, is it substantial? The segment must be large enough and have enough purchasing power to justify the cost of targeting it. A group of people who love your product concept but can’t afford it isn’t a viable segment. Similarly, a segment of 200 people in a market of millions probably isn’t worth building a separate strategy around.
Select Target Segments
You don’t have to pursue every segment you’ve identified. In fact, trying to serve all of them equally is usually a recipe for diluted effort. Rank your segments by their estimated size, growth potential, competitive intensity (how many other businesses are already fighting for them), and how well your strengths align with their needs.
Some businesses choose a single segment and go deep. Others pick two or three and develop distinct approaches for each. The right number depends on your resources. A startup with a small marketing budget and one product might focus on a single high-potential segment. A larger company with multiple product lines might address four or five.
Build Segment Profiles
For each target segment, create a detailed profile that makes the group feel concrete to everyone in your organization. Include the demographic, geographic, psychographic, and behavioral characteristics that define the segment. Describe what drives their purchasing decisions, what they value, where they spend time, and what kind of messaging resonates with them.
These profiles become the working documents your marketing, sales, and product teams use daily. The more specific they are, the more useful they become. “Budget-conscious parents aged 30 to 45 in suburban areas who compare prices online before buying in store and respond strongly to promotional discounts” is far more actionable than “price-sensitive families.”
If your organization relies on a CRM, make sure the segment definitions can be mapped back to fields in your customer database. A segmentation that lives only in a PowerPoint deck and can’t be applied to real customer records rarely gets used after the initial presentation. The most effective segmentations are ones your team can filter, sort, and act on inside the tools they already use every day.
Put Segments Into Action
The final step is translating segments into differentiated marketing actions. This is where the four Ps come in: product, price, place, and promotion. Each target segment may need a different combination.
One segment might respond to a premium version of your product sold through specialty retailers and promoted via content marketing. Another might want a stripped-down version at a lower price point, distributed online, and promoted through paid search ads. The whole point of segmentation is to stop treating your market as one homogeneous group and start making deliberate choices about how to serve each group differently.
Track results by segment over time. Measure whether each segment’s actual purchasing behavior matches your projections. Segments shift as markets evolve, new competitors enter, and consumer preferences change. Plan to revisit and refresh your segmentation at least annually, comparing your original profiles against current data to see what still holds and what needs updating.