Customer segmentation: What it is and how to do it

The way you do customer segmentation can make or break your product.

If you do it well, you can tailor your value proposition, packaging, and pricing to win the hearts of a well-defined group of customers. Do it poorly, and you’ll end up with a product that no one likes.

Segmentation brings focus. A product for everyone is a product for no one. The more you focus on a specific segment, the higher the chance of actually winning and dominating that segment.

At the end of the day, it’s better to win a smaller but well-defined part of the market than to strive to satisfy everyone.

The latter is impossible. Don’t even try.

What makes a good customer segment?

Let’s first define what constitutes a good customer segment. After all, there are hundreds of possible ways and criteria to segment people. You could even segment people by the color of their underwear. That would be concerning, though.

There are two core characteristics of a good segment. A good segment should be:

• Targetable — You can adjust your distribution channels to target that specific segment and filter out the rest
• Homogenous — People in the segment should react in a similar manner to the product, packaging, messaging, and so on

An ideal segment would also be self-referencing — meaning people in the segment tend to recommend products to one another.

In other words, proper market segmentation allows you to act differently on different segments.

If you can’t do that, why even bother?

Criteria for identifying customer segments

Why demographics rarely work

Segmentation by demographic data, such as gender, nationality, age, etc., is still among the most popular approaches. Mostly because it’s the most straightforward criterion.

It’s also the stupidest.

Let’s take a look at the characteristics of a good segment again. Ask yourself, if you segment by age and gender, is such a segment:

• Targetable? Kind of; you can set up ads to target specific demographics, but not much more
• Homogenous? No; people can be of the same age, gender, and nationality and still represent completely different categories of thinking
• Self-referencing? Somewhat; although people with similar demographics are more likely to refer new products to each other, since they are not homogenous, the recommendation frequency and effectiveness are limited

How to define user segments

Luckily, there are better ways to identify criteria for segmentation. My favorite approach to defining user segments involves the following three steps:

1. Find people experiencing similar needs and pains
2. Find a subgroup that values a similar solution
3. Focus on a group with a similar willingness to pay

1. Needs and pains

Nothing connects people better than experiencing common pain points. Pains don’t differentiate by age, gender, or anything else.

You can have a 9-year-old child, an 80-year-old grandma, and a 35-year-old corporate worker experiencing the same pain point and having the same need, albeit for different reasons.

For example, the child might want to improve their English grammar skills to ace a test in school. The corporate worker might want to sound professional. And the grandma might have problems remembering the proper structures.

Although they are of entirely different demographics, a common need makes them a relatively homogenous group.

2. Perceived value

One need can be satisfied by many solutions. Understanding how valuable people find a given type of solution is another great way to segment or further the subsegment of the user base.

For example, there might be a significant group of people experiencing a pain point associated with their daily commute. The segment is not homogeneous enough, though. There’ll probably be a subgroup of people that:

• Want to eliminate the commute altogether — You won’t satisfy them with faster trains or buses; all they want is a remote job
• Love their work but hate mass transit — They’ll see the most value in individual solutions such as electric bikes or scooters
• Prefer mass transit — Their biggest dream is a direct metro line between their home and office
• Are used to the status quo — They complain and experience the pain but are unwilling to do anything about it; they don’t care enough

Although these customers experience the same pain, they value different solutions differently. Therefore, they should be treated as separate segments. You have to act differently to win these different groups.

3. Willingness to pay

A truly homogenous and self-referencing segment is one that’s willing to pay similar money for similar features.

Back to our commute example: say you have defined a group of people who experience the pain of daily commuting. You decide to narrow down the focus to the ones who want to commute on their own while avoiding mass transit. There’s still one critical factor left on the table: different people are willing to pay different amounts to have the problem fixed.

You will probably find people willing to pay a lot just to make their commute more enjoyable. You could satisfy these people with some premium electric bikes, for example. But there’ll also be a more frugal group of people who can’t afford such a bike. A monthly electric scooter subscription might be a more suitable solution for them. And so on, and so on.

If you can’t win the segment in a similar manner, then it’s too heterogeneous. Narrow it down by treating groups with different willingness to pay as different (sub)segments.

How to prioritize customer segments

If you follow the criteria I proposed and segment by pains, perceived value, and willingness to pay, you will end up with more than one segment — in other words, you will end up with too many segments.

This is one of those good problems to have; it’s better to have 20 well-defined, actionable user segments than one vague, massive segment. But how do you decide which segment to focus on first?

There are two criteria worth considering:

• Segment attractiveness
• Business strength

Segment attractiveness

How attractive is the segment? How much value can you capture by catering to the needs of a specific segment?

Some criteria to evaluate segment attractiveness include:

• Compulsion — How strong is the need users experience?
• Size — How big is the segment?
• Growth — Does it grow or decline?
• Competitors — Do other companies focus on that segment?
• Entry barriers — How hard is it to enter the segment?

Business strength

How well is your company positioned to tackle the specific segment? Depending on your current business model, some segments might be easier to enter than others.

You can assess business strength by looking at your:

• Capability — Do you have the skills, knowledge, and expertise to cater to the needs of the specific segment?
• Credibility — Do you already have credibility in the segment you want to tackle? If not, how quickly can you build a strong reputation there?
• Distribution — Can you use your existing distribution channels to target people in the segment?
• Alignment — How well does entering the segment align with the company’s mission, vision, and strategy?

Segments that are both attractive and easily winnable are your best bets. As a second priority, choose either the most attractive (big bets) or most winnable (minor optimizations) ones. Ignore the rest.

A 4-step customer segmentation framework

Proper segmenting brings you focus. In the end, it’s better to win one or two small segments than do so-so in one big segment.

A good segment is homogenous, targetable, and self-referencing. Although demographic segmentation is the most straightforward approach, it doesn’t meet the characteristics of a good segment. In most cases, it’s just useless.

To select your next customer segment, you can use this four-step framework:

1. Find a unique need or pain point shared by a group of people
2. Understand what people in this group value the most
3. Discover how much these people are willing to pay for things they value the most
4. Prioritize customer segments by comparing segment attractiveness with your business strength

When in doubt, subsegment further. It’s very rare to define too narrow a segment. More often than not, people define broad, unactionable segments.