Be unmissable to someone

By Adrian Watkins. Roughly an eight minute read.

I spent a recent morning with a room of founders in Malaysia, hosted by Beyond4Tech, talking about something that sounds simple and almost never is: who is your customer, really?

It’s a warm, generous group, the kind that asks the awkward questions rather than the polite ones, and we kept circling back to the same quiet truth for an hour. Most founders believe their job is to be wanted by everyone. The more people who might want the product, the thinking goes, the bigger the prize. It feels like ambition. It’s actually the thing most likely to sink you.

I’ve run, advised, built, and closed businesses across the world, and I’ve been guilty of this myself more than once. So none of what follows is theory. It’s a set of things I’ve learned the hard way, backed up by people who have studied it far more rigorously than I have. The good news is that most of it is about noticing, not guessing.

The opposite of what most founders think

Your job isn’t to be wanted by everyone. Your job is to be unmissable to someone, and to offer a polite, unflinching no to everyone else.

That sounds harsh until you sit with it. A product that’s for everyone stands for nothing. It has no edge, no obvious reason to choose it, and no story that travels from one customer to the next. You go wide, and in going wide you lose the credibility and the authenticity that make people pick you in the first place.

The cost of getting this wrong isn’t abstract. When CB Insights worked through why failed startups actually died, reading the post-mortems of hundreds of venture-backed companies that shut down, the single biggest killer wasn’t weak technology or thin funding. It was poor product-market fit, which accounted for 43% of cases (CB Insights, Why Startups Fail: Top 9 Reasons). Sit with that number. Getting on for half of these companies didn’t fail because they couldn’t build. They failed because they built for a customer who, in the end, wasn’t quite anyone.

Poor product-market fit is what “for everyone” looks like a year later, once the money has gone. The deeper point is that fit isn’t something you decide from the founder’s chair. It’s something the market confirms, and keeps confirming. You don’t pick your customer. You watch which customers are picking you, and then you find the discipline to drop, or at least quietly deprioritise, the rest.

You don’t pick your customer, you notice who picks you

That single shift, from inventing your customer to noticing your customer, changes almost everything that comes after it.

When I started AIinASIA in 2020, the lazy instinct was to cover AI for everyone. There were already a hundred sites doing exactly that, louder and far better funded than me. So I asked a narrower question: who am I genuinely the best person in the world to write for? The answer was the operator in Asia who has to adopt this stuff and make it work, not the tourist who likes to admire it from a distance.

I didn’t guess that. The market confirmed it. The pieces written for that one reader travelled. The generic ones sat there and died. Today the site reaches around ten thousand readers a month with no paid distribution, because the right people instantly recognise it was built for them rather than at them.

The clarity came from subtraction, not addition. You start broad because you have to, then you remove the customers who take too long, need endless customisation, give you every good buying signal and never actually buy, and complain the loudest while knowing full well you’re a startup. What’s left, once you have the courage to subtract, is the shape of the person you’re truly for.

The behaviour test, not the opinion test

So how do you know you’ve chosen right? You watch what people do, not what they say.

This is the part founders find hardest, because opinions are flattering and behaviour is honest. “I’d definitely buy that” in a survey is worth nothing. People are kind, they want to encourage you, and they’ll happily tell you your idea is wonderful while never reaching for their wallet. A signature, a renewal, a referral is worth everything, because it costs the other person something.

It’s the same logic behind my day job at SQREEM, where I run global commercial operations and governance: the Large Behaviour Model (LBM) exists because the behavioural signal is what tells you who your real customer is, faster than any survey can.

Y Combinator, which has now funded thousands of startups and watched most of the failure modes up close, puts the same idea more bluntly in its essential startup advice: it’s better to make a few users love you than a lot of users merely like you. Paul Graham’s test for whether a market even exists is whether there are people who want what you’re building so urgently that they’ll use it, bugs and all, from a company they’ve never heard of. Not a lot of them. Some. Love is a behaviour. It shows up as urgency, and urgency shows up in your numbers.

A founder I mentored learned this in a single evening. She had a genuinely good product that could help many small businesses in many different ways, so she was selling everything to everyone, slowly, and building features endlessly just to keep each deal moving. We stopped guessing and went to her actual paying customers. We surveyed them, then looked hard at the ones who were happiest. The pattern was unmissable: clinic owners who had just hired their first admin member of staff. That wasn’t her opinion. It was sitting in her wins the whole time. She just hadn’t looked. She rebuilt around that one person, and reached profit fast.

Your earliest and best data isn’t a dashboard. It’s the pattern hiding in the handful of customers who were easiest to win.

Speed is the tell

If you want a single signal to watch in the early days, watch speed.

The right fit reaches a decision quickly. They feel the pain now, they already have a budget line for solving it, and they can see the value without you bending the product into a custom shape for them. You want momentum: multiple, fast conversions, deal after deal, rather than one prestigious account you’ve been nursing through a six-month internal war.

Slow, painful deals are usually a fit problem wearing a process costume. It’s tempting to blame procurement, or timing, or a difficult stakeholder, but very often the deal is slow because the customer isn’t really your customer. That big logo you’re chasing may be costing you the ten right-sized customers you could have won and delighted in the same period. When you find yourself fighting for every inch, it’s worth asking honestly whether you’re pushing the wrong fit uphill.

The referral is the strongest signal of all

Of all the signals, the one I trust most is the referral, because it’s the market doing your positioning for you.

Think about how this actually happens. People in the same line of work talk to each other. They sit in a coffee shop, they compare notes, they scratch their heads over the same problems. And one of them says, “I’ve got just the thing that solved this for me, it cost this much, and it worked quickly.” That’s a referral, and it’s gold, because it takes a customer who understood the fit clearly enough to spot it in someone else, and the conviction to bring them in.

So the referral is also a test you can fail. If a happy customer can’t describe you crisply enough to pass you on, your positioning is still blurry to them, however much they like you.

This is where the discipline pays off in hard commercial terms, not just in tidiness. McKinsey’s research on precision in how companies target and speak to customers found that getting it right typically lifts revenue by 10 to 15%, and that the fastest-growing companies derive about 40% more of their revenue from that precision than slower-growing peers do (McKinsey, The value of getting personalization right). A well-aimed message, built for a customer you actually understand, travels. A generic one gets ignored, and you pay to be ignored.

None of which requires waiting until you’re big. First Round Review, distilling go-to-market advice for early founders, makes the same case for starting now: even when you’re early, define your ideal customer so you can narrow your focus, learn and iterate faster, and spend your scarce time on the leads most likely to be transformed by what you do (First Round Review, The Most Common Go-to-Market Questions From Founders). The point of choosing isn’t neatness. It’s that everything you do next gets cheaper and faster.

A niche is a beachhead, not a prison

The fear sitting under every one of these decisions is the same: if I narrow down, the market will be too small. So let me meet it head on, because it’s the question that keeps founders broad and therefore invisible.

Being targeted isn’t the same as being finite. A niche isn’t a ceiling. It’s a controlled experiment for a company with limited funds, limited customers, and limited runway. You’re not shrinking your ambition. You’re choosing where to win first, then moving sideways from strength.

This is one of the most studied moves in business. Peter Thiel argues in Zero to One that every startup should begin with a very small market, because it’s far easier to dominate a niche than to win a slice of something huge, and that the discipline is to expand only once you own it. Geoffrey Moore built an entire methodology around the same instinct in Crossing the Chasm: pick one beachhead segment, concentrate everything on dominating it, then roll into adjacent segments like pins in a bowling alley. Harvard Business Review has made the case directly to founders, arguing that the go-to-market approach startups actually need is to win a narrow beachhead before broadening, rather than spreading thin from day one (Harvard Business Review, The Go-to-Market Approach Startups Need to Adopt).

The textbook example is Amazon, which sold nothing but books, globally, for years. Only once it owned that did it move into other categories, having already built a trusted brand and a buying experience people loved. The niche wasn’t the limit. It was the launch pad.

My own version is PromptAndGo. The global incumbents had more money, more content, and more reach. I couldn’t out-broad them, and I didn’t try. I out-specified them, with prompts built for the Asian operator’s market, register, and language. That’s a position a global player can’t copy without becoming a worse fit for everyone they already serve. You don’t beat a giant by being a smaller giant. You beat them by being unbeatable for the customer they can’t be bothered to love.

Saying no without flinching

None of this works if you can’t say no out loud.

This is where founders flinch, and I understand why. Saying no to a potential customer feels like saying no to revenue, and when you’re hungry that’s genuinely hard. But if you never define your anti-customer, the whole team chases anything that moves. Your effort scatters, your data turns to noise that can’t tell you what’s working, and you burn out your people and your runway at the same time. It’s the quiet road back to that 43% who built for no one in particular.

Naming the segment you won’t serve isn’t a loss. It’s what frees the time, the budget, and the focus for the people who actually fit. Give yourself permission to say no, the same way you give yourself permission to test, to play, and to be wrong in small, controlled ways.

Near the end of the session in Malaysia, one of the founders summed the whole hour up better than I had. She had been working through her own customers, and she put it like this: rather than starting from broad demographics, she had begun looking at which segments were giving her a small return for a large slice of her budget, and cutting them. “Our product is not for everyone,” she said, “so we eliminate, one by one.”

That’s the entire discipline in a sentence. Not a grand act of positioning genius, just the honest, repeated work of looking in the mirror, asking why each customer is really there, and removing the ones who don’t fit until the picture comes clear.

---

If you want the short version, I distilled this into this week’s Friday Frame. This essay is the long-form writeup behind it.

Thanks for reading,
Adrian

This is written as part of the democratising.ai engine.