That is the whole test, and applying it honestly eliminates most of what gets called segmentation. If you would treat two groups of customers identically, they are not two segments. They are one segment you have described twice.
This matters because the standard approach produces grids that look like analysis and change nothing. Age brackets, income bands, job titles, a persona with a stock photograph and an invented name. A first year business can spend a weekend on that and end the weekend with exactly the same business it started with.
Demographics are usually the wrong axis
Two customers can match on every demographic you could record and be completely different to serve.
Same age, same income, same part of town, same size of business. One knows what they want, decides in a day, and pays on time. The other needs three conversations, changes the scope twice, and pays in sixty days. Those are different customers in every way that affects your business, and no demographic field would have told you apart.
The axes that actually predict behavior are about circumstance and conduct rather than identity.
What triggered them to look. Somebody whose system failed this morning behaves nothing like somebody planning a renovation for spring, even when they are demographically identical. Urgency changes price sensitivity, decision speed, and tolerance for inconvenience, which is three of the four things you care about.
How they found you. A referral arrives already convinced and asks different questions than somebody who found you in a search result and is comparing four options. The work of selling to them is not the same work.
How much attention they consume. Some customers cost you an hour of overhead on a job and some cost you a day. This is invisible in revenue and decisive in profit, and almost nobody tracks it.
Whether they return. A customer who buys once and a customer who buys quarterly are the same transaction and completely different assets.
Start from the customers you have
Segments are discovered rather than invented. You cannot reason your way to them from an empty spreadsheet.
Take your last twenty customers. Not the ones you wish you had. Write each one on a line with four things: what made them contact you, how they found you, how much trouble the job was relative to what you charged, and whether they have come back or referred anybody.
Then look for clusters. They will be messier than the frameworks suggest and more useful. Most first year businesses find two or three groups, and frequently one of them is a group they had not noticed they were serving at all.
If you have fewer than twenty customers, you do not have enough to segment and the exercise is premature. What you have instead is a list of every customer, which is better, because you can hold all of them in your head and should.
The test again, applied
For each cluster you find, ask what you would do differently for them. If the honest answer is nothing, merge them back together.
Different how you reach them. Different what you say first. Different what you charge or how you structure the price. Different what you include. Different how much you follow up. Different whether you take the work at all.
If a distinction does not change one of those, it is an observation rather than a segment. Keeping it costs you nothing in money and something in attention, which at this stage is scarcer.
The segment most businesses miss
The one you are serving badly and profitably by accident.
Nearly every first year business has a group of customers it did not set out to serve, acquired through some channel it was not deliberately working, who buy something adjacent to what the business thinks it sells. They are frequently the most profitable group and they are invisible because nobody looked.
The reverse exists too and is more painful. A group the owner considers core, talks about constantly, and loses money on every time. That one is hard to see because identity is tangled up in it, and the clustering exercise is useful mainly because it surfaces the arithmetic without an argument.
Naming them properly
Name a segment by what the customer is trying to accomplish, not by what they are.
Small commercial property owners is a description of people. Landlords who need a problem fixed before a tenant moves in next week is a description of a situation, and the second one tells you what to say, what to charge, and how fast to respond.
The test of a good segment name is whether it implies the action. If reading the name does not suggest what you should do, the name is doing decoration rather than work.
How many you should have
Two or three. Possibly one.
A first year business running five segments is not segmenting. It is describing its entire market in slightly more detail, and the practical effect is that nothing is prioritized, because everything is a priority.
The point of defining segments is to decide where attention goes, which requires that some groups receive less of it. A segmentation that does not cause you to deprioritize anybody has not made a decision.
If the thought of naming a group you will serve less feels uncomfortable, that discomfort is the exercise working. Every business that is good at something specific became that way by being less available to everybody else.
What to do with them once defined
Three things, in order of how much they return.
Change what you say to each one. The same service described for a different situation reads as a different service. This costs a paragraph and it is the highest return change available.
Change where you spend your time. If one segment produces work that is twice as profitable and half as difficult, that is an argument for where the next ten hours of effort go, and it is a stronger argument than any marketing advice.
Change who you decline. The segment you have identified as costing more than it returns is not a problem to be solved with better processes. It is a group you can stop pursuing, and doing so is usually the fastest improvement in margin available to a small business.
The maintenance nobody does
Segments drift. The groups that describe your business at month ten will not describe it at month thirty, because the work you have taken has changed who finds you, and who finds you changes who you serve.
Redo the exercise once a year with the last twenty customers. It takes an hour and it is the only way to notice that the business has quietly become something else, which it will, and which is frequently fine as long as it was chosen rather than drifted into.
The short version
Look at who you have actually served rather than who you imagined. Cluster them by what prompted the purchase, how they arrived, what they cost you, and whether they returned. Keep a group only if it changes what you do. Name it by the situation rather than the demographic. Hold yourself to two or three, and accept that defining them properly means deciding to be worse at serving somebody.
That is a morning of work and it outperforms any framework, because it is built from your actual customers rather than from a category you were told you belong to.