3 pixelated ghosts on a blue background - market segmentation for small business

You know you’re good at what you do. Your clients stick around and your work holds up over time, yet the enquiries coming through are a mixed bag of price shoppers and tyre kickers, with the odd great client turning up more by accident than by design.

So you do what most business owners do and widen the net. The website copy gets broader so nobody feels left out, another service gets added to the list and the language gets softened until it could appeal to just about anyone, which somehow makes things worse rather than better.

In my experience the problem usually isn’t your marketing at all, it’s who you built the brand for. If you designed it around the average customer in your market, you designed it for someone who doesn’t exist. That’s really what market segmentation is about, it’s also the step most businesses skip on their way to a new logo or website.

In this guide I’ll walk you through what market segmentation is, the four common types and where each one falls short, why the average customer is a myth and how to find the group that’ll actually choose you. If you’re starting out without a research budget, there’s also a free AI prompt further down you can use to build your first customer profile.

Contents

What is market segmentation?

Market segmentation is the process of splitting a market into groups of buyers who want different things and choose differently, so you can decide which group to build your business around.

Most definitions stop at “groups of people with shared characteristics” like age, income, location or industry, which is where it tends to go sideways. Two people can share every one of those characteristics and still make completely different decisions when it comes time to buy.

The way I see it, a useful segment isn’t a group of people who look alike, it’s a group of people who choose alike. Pretty much everything else in this guide hangs off that idea, so it’s worth holding onto.

The pilot who didn’t exist

In the late 1940s the US Air Force had a problem it couldn’t explain. Pilots were losing control of their planes and crash investigations kept coming back without a mechanical cause.

Attention eventually turned to the cockpit, which had been designed around the average body measurements of pilots taken back in 1926. The thinking was that pilots had simply got bigger over the years, so in 1950 researchers at Wright Air Force Base measured 4,063 pilots across 140 physical dimensions to build a new, more accurate average.

A 23-year-old lieutenant named Gilbert Daniels asked a different question. Rather than working out the new average, he wanted to know how many pilots actually were average.

He took the ten dimensions that mattered most for cockpit design, including height, chest circumference and sleeve length. He defined average generously as anyone who fell within the middle 30% of the range on each one, then counted how many pilots sat inside that range on all ten.

The answer was zero, not one pilot out of more than four thousand, which meant the cockpit had been carefully engineered for someone who wasn’t there. As Todd Rose tells it in The End of Average, the fix wasn’t a better average at all. It was adjustability, with seats, pedals and straps designed to fit the range of real people instead of the arithmetic.

I think about that story a lot, because most of the brands I’m asked to fix were built exactly the same way.

Why designing for the average customer fails

Designing for the average customer fails because the average is a blend of many different buyers, so a brand built for the blend ends up fitting none of them particularly well.

The average customer in your market wants reasonable quality at a reasonable price with reasonable service. Build for them and you end up with a business that’s reasonable at everything, which is fine for plenty of people but the first pick for none of them.

You can see it on most business websites. Lines like “We work across all industries”, “Quality, reliability and value” and “Tailored solutions for businesses of every size” were each written to avoid excluding anyone, yet put together they give nobody a reason to choose you.

The consequences are probably ones you’re already living with. You keep attracting the wrong clients because nothing in your brand tells the right ones they’ve found their place. You get dragged into price conversations because when you look like everyone else, price is the only thing left to compare. You lose tenders to competitors who aren’t as good as you, simply because they looked more certain about who they were for.

It feels like a marketing problem, but it’s really a segmentation problem. The surface keeps getting fixed while the foundation stays built for a statistical ghost. If your messaging is where it shows up most for you, that’s usually where I’d start the conversation about brand messaging and creative.

The four types of market segmentation

There are four common types of market segmentation: demographic, geographic, psychographic and behavioural. Each one is useful in its own way, yet each one also has a blind spot you need to know about.

Demographic segmentation groups buyers by who they are, such as age, gender, income, occupation, business size or industry. The upside is that the data’s easy to find and easy to target in ads, which makes it handy for sizing a market. The downside is that it describes people without explaining why they buy, so two 45-year-old business owners on the same revenue can want opposite things from you. Think of a financial planner who targets business owners aged 50 to 65. That’s useful for choosing where to advertise but pretty useless for writing a single line of copy that makes one of them feel understood.

Geographic segmentation groups buyers by where they are, whether that’s country, state, city or service radius. It’s essential for service businesses with a physical footprint and it’ll sharpen your local search. What it can’t tell you is how someone decides, because a Gold Coast business owner and a Brisbane one can be far more alike than two Gold Coast owners with different priorities. A builder who only takes jobs within a set radius is making a sensible operational call, but a service radius isn’t a positioning.

Psychographic segmentation groups buyers by what they value, believe and care about. It gets you closer to motivation and gives you tone, story and personality to work with. The catch is that it’s easy to fill with wishful thinking, since everyone says they value quality. Who Gives A Crap, the Australian toilet paper company that donates 50% of its profits to clean water and sanitation non-profits, is a good example of it done well. It isn’t built for every household. It’s built for people who want an everyday purchase to do some good and who’ll enjoy a joke on the packaging. Plenty of people don’t care about either of those things, but the ones who do are fiercely loyal.

Behavioural segmentation groups buyers by how they actually choose, meaning the outcome they’re after, what they’ll trade off and how they decide. It’s the only type that explains choice directly, which makes it the one you can build a brand around. It’s also harder to research because you have to talk to real clients rather than pull a report. Aldi is the best local example I know of. When its first Australian stores opened in 2001, each one stocked around 900 products, most of them brands nobody had heard of. It didn’t try to out-choice Coles and Woolworths. It was built for the shopper who’ll happily trade choice for a lower price and a quicker shop, so if you wanted six kinds of pasta sauce, Aldi was never for you and it didn’t need to be.

For most established businesses, behavioural segmentation is where the real answers are. I’d use demographics and location to find and size the group, then use behaviour to decide what to build.

Market segmentation types compared

Segmentation type What it tells you What it can’t tell you Best used for
Demographic Who the buyer is Why they choose one option over another Sizing a market, targeting ads
Geographic Where the buyer is What they value or how they decide Service areas, local SEO
Psychographic What the buyer cares about Whether those values change what they buy Tone, story, brand personality
Behavioural How and why the buyer chooses Where to find them (on its own) Positioning, offer design, messaging
The “average customer” A mathematical midpoint Anything about a real person Nothing you should build a brand on

When broad appeal does work

Broad appeal works for market leaders with the scale, budget and distribution to be everywhere at once.

It’d be unfair of me not to make the other side’s case properly. For example Bunnings isn’t trying to be the perfect fit for one narrow group, it wants to be the obvious place for anyone with a weekend project.

There’s serious research behind that approach too, which comes out of our own backyard. Byron Sharp and the Ehrenberg-Bass Institute at the University of South Australia argue that brands grow mainly by reaching more buyers across the whole category. They’ve also found that competing brands sell to surprisingly similar customer bases, which is why Sharp is openly sceptical of narrow targeting.

I take that work seriously, most of it, though, was built on big consumer brands in categories where people buy often and think little. Choosing who fits out your office, rebuilds your website or manages your project is a different kind of decision entirely. It’s rare, expensive and personal, so the buyer shortlists two or three businesses and asks which one gets them.

If you’re a $4M firm competing against someone bigger, trying to be for everyone means fighting on their terms. They’ll win every contest that comes down to who’s more visible to more people. Your reach can be as broad as you like, but the brand itself still has to be built for someone.

How to find a segment worth building for

You don’t need a research agency for this. What you need is clear answers to five questions.

  1. Start with your best clients, not your biggest market. List the ten clients you’d clone tomorrow: the most profitable, the easiest to work with and the most likely to refer. Ignore everyone else for now.
  2. Ask why they chose you, not who they are. Ring three of them and ask what they were trying to fix when they found you, who else they looked at and why they didn’t go with them. Write down their exact words rather than your interpretation of them.
  3. Test whether they choose differently. Put that group next to your other clients and see whether they decide on different things. If the answer’s “not really”, you’ve found a demographic rather than a segment, so keep digging.
  4. Check it’s worth winning. Interesting doesn’t always mean profitable. Is the group big enough, valuable enough and reachable enough? Could you realistically become their best option?
  5. Decide who you’re not for. This is the one that hurts. When I rebuilt my own brand, the hardest conversation wasn’t about the name or the logo, it was deciding who we’d stop chasing. Narrowing means some people will look at your brand and think “not for me”, which is exactly the point.

For most small and mid-sized businesses, that means building around one primary segment with room for a second where the fit’s strong. Each segment needs its own positioning and message and in my experience very few businesses under $10M have the budget to do more than two properly.

If you’ve made the call and you’re stuck on where to narrow, my guide on how to choose a niche picks up from here and if you’d like a sharper picture of that one person, my guide on how to define your ideal customer goes deeper.

Free AI prompt to build your first customer profile

If you’re just starting out, you might not have ten great clients to learn from yet, let alone a budget for research. That’s where a free AI tool like ChatGPT, Claude or Gemini can help, as long as you use it the right way.

AI can’t tell you what your customers think because it’s never met them. What it’s good at is asking you better questions than you’d ask yourself, noticing when your answers sound like wishful thinking and turning a messy brain dump into a profile you can test with real people.

Copy the prompt below into whichever tool you use, fill in the bits in square brackets and answer its questions as truthfully as you can. The more specific you are, the more useful the output gets.

I'm building a small business and I want a customer profile based on how people choose, not just who they are. Act as an experienced brand strategist and interview me. Ask one question at a time and wait for my answer before moving on.

Here's what I know so far:
- What I sell: [your product or service]
- Where I operate: [city, region or online]
- Who I think my customer is: [your best guess]
- Real customers or conversations so far: [describe them, or write "none yet"]

Work through these areas with me:
1. The problem the customer is trying to fix when they go looking for someone like me, plus what usually triggers the search.
2. Who else they'd consider, including doing it themselves or doing nothing.
3. What they'd trade off to get what they want, such as price, speed, choice, quality or convenience.
4. How they make the decision and who else has a say.
5. The words they'd use to describe the problem, not the words I'd use.

When we're done, give me:
- Two or three possible segments that choose differently from each other, with a short description of each.
- The segment you think I'm best placed to win, why you think that and the biggest risk in that call.
- Five questions I can ask real people this week to test whether you're right.
- A one-paragraph profile of the customer I should build my brand for.

Push back if my answers sound like wishful thinking or describe everyone. Flag anything you've assumed rather than heard from me.

Treat whatever comes back as a hypothesis rather than an answer. The five questions at the end are the most valuable part, because the only way to know whether a segment is real is to put those questions to real people and listen to how they respond. If what they say doesn’t match the profile, trust them over the AI every time.

Signs you’ve built your brand for the average

  • You describe your ideal client as “anyone who needs what we do”.
  • You keep attracting the wrong clients and can’t work out why.
  • Your website could swap logos with three competitors and nobody would notice.
  • Most sales conversations end up being about price.
  • You lose tenders to businesses that aren’t as good as you but looked more professional.
  • You’ve tried a new website, ads or a logo refresh and nothing changed.
  • You hesitate before sending someone the link.

If three or more of those land, the fix isn’t another tactic but a brand strategy decision about who you’re for.

FAQs

What is market segmentation in simple terms?

Market segmentation means dividing your potential buyers into groups that want different things and choose differently, then deciding which group to build your business for. The point isn’t to label people. It’s to decide who you’re designing for so your offer, pricing and message fit someone properly instead of everyone loosely.

What are the four types of market segmentation?

The four main types of market segmentation are demographic (who buyers are), geographic (where they are), psychographic (what they value) and behavioural (how and why they choose). Most businesses lean on the first two because the data is easy to get. Behavioural is the most useful for positioning because it’s the only one that explains choice directly.

Why doesn’t the average customer exist?

The average customer doesn’t exist because an average blends many different people into a midpoint that no individual actually matches. The clearest proof is the 1950 US Air Force study, where Gilbert Daniels found that none of 4,063 pilots fell within the average range on all ten key body measurements. Customers work the same way, so if you average their needs you get a person nobody is.

Is market segmentation the same as choosing a niche?

No. Market segmentation maps the different groups in your market, while choosing a niche is committing to one of them. You segment first, then choose. Skipping segmentation is how businesses end up picking a niche by gut feel or by industry label alone.

Is it risky for a small business to target a narrow segment?

Targeting a narrow segment is usually less risky than targeting everyone, because a business that’s the best fit for someone tends to win more work at better margins than one that’s an okay fit for all. You won’t stop serving people outside the segment. You’ll just stop building your brand around them.

Can I use AI to build a customer profile?

Yes, free AI tools like ChatGPT, Claude or Gemini can help you build a first customer profile by interviewing you, testing your assumptions and suggesting possible segments. What they can’t do is tell you what real customers think, so treat the profile as a starting hypothesis and test it by talking to actual buyers before you build your brand around it.

My takeaways

The average customer is arithmetic rather than a person, so nobody who matches it is ever going to ring you.

Demographics tell you who someone is while behaviour tells you why they choose, which is why I’d always build on the second.

Broad appeal suits the market leader, but for everyone else it means fighting on someone else’s terms.

Narrowing feels like giving something up, yet it’s really what it costs to become the obvious choice for someone.

If you’re starting from scratch, a free AI prompt and a handful of real conversations will get you further than guessing ever will.

Whatever stage you’re at, stop building a cockpit for a pilot who doesn’t exist.