Distribution is the moat

What a decade of backing 250 founders taught me - and why AI just made it the only thing that matters.

For ten years I've watched brilliant builders lose to mediocre sellers.

AI just poured petrol on it.

When anyone can ship a product in a weekend, the product stops being the moat. So let me tell you what a decade of Startmate demo days actually taught me about the thing that wins.

Someone once introduced me on stage like this:

"We've got someone who has put 250 companies through a pressure cooker and watched what happens to each of them."

Honestly, that's the most accurate description of my last decade I've ever heard.

I've backed 250+ startups - the Startmate portfolio is now worth over $4.5 billion across 350+ companies employing 3,500+ people.

A cybersecurity unicorn in Bugcrowd, hundreds-of-millions outcomes in Upguard and Morse Micro, and the ones I backed in my own tenure: Amber, Mindset Health, Heaps Normal, Sitemate, Lyrebird Health and Andromeda Robotics.

I've personally interviewed close to a thousand founders and reviewed thousands of applications.

And here's a little-known fact about me: I track every single founder I meet in a scorecard. Product. Market. Founder fit. Distribution. Then every 12 months I go back and mark my own homework - was I right, would I invest in them again, yes or no.

So this isn't vibes. It's a decade of receipts. Here's what they say.

1. The builders broke my heart

When I started, I over-indexed hard on the builder.

I'm non-technical, so I was mesmerised by the people who could just... make the thing. 

Beautiful products. Scalable. Never broke. Big logos on the CV. At Startmate we'd tell every founder they were the top 3%, that the expectation was to become a cohort-of-one Canva. And the builders looked exactly like what a top 3% founder is supposed to look like.

I remember a wave of them in 2018 and 2019. Heads of engineering, CTOs from great scale-ups. Genuinely some of the most talented builders I've ever met.

And they'd freeze the second you put a customer in front of them.

They didn't want to sell. Didn't want to market. Stayed inside their own network and never branched out. They kept doing the thing they were already great at - building - and quietly avoided the thing that actually grows a company.

"Build it and they will come" is the most expensive lie in startups.

2. A terrible product with customers beats a beautiful product with none

Here's the one that flips people's brains.

A shit product with lots of customers is more impressive than a great product with lots of customers.

Think about it. If your experience is bad and people still pull out a credit card and put up with it, the pain must be so real they'll tolerate anything to get it solved. That's product-market fit screaming at you.

The best example I know is Casey Ellis. Before Bugcrowd had a website - before it had a product - Casey sold $500,000 in preorders off a Wufoo form. No product. No site. Just a real problem, a form, and the guts to get in the room and sell the outcome.

That crowdsourced-security idea he was pre-selling? Bugcrowd is now a US$1 billion company, having raised a US$102 million round led by General Catalyst in 2024.

They got in front of the most senior person they could find. Sold the outcome. Started the waitlist. Took pre-orders. Got credit cards down before a single line of code existed.

That's not a growth hack. That's the whole job.

3. So what actually is "distribution"?

Everyone wants the one channel. There isn't one.

My rule is simple:

Lean into what feels easy to you and hard for everyone else, and go to where your customers already live. Where those two overlap - that's your unfair advantage.

It's not "do sales." It's not "do content." It's not "build a community." It's: where do your customers actually hang out, what do they actually read and watch, and which of those can you show up in with energy instead of dread?

Small example from my own life. I ran a little productivity community called Puddle Pod. It's grown to cohort after cohort on $0 of marketing, purely word of mouth. Not because word of mouth is some genius growth strategy - but because writing and community are the things that give me energy. That's the point. The right channel for you is the one you'll still be doing on day 400.

And distribution is a day-400 game.

4. How I actually spot it in a room

After close to a thousand of these, the tells are boringly consistent.

Speed of reply. The single best predictor. When an interviewer once asked me for my most underrated founder metric, I didn't say vision or IQ. I said communication speed. The best founders reply to a customer email instantly. Not "later today." Now.

If they're slow to reply to me, they're slow to their customers.

Do they track the funnel? The ones hiding from sales don't track pipeline, don't know their conversion, can't tell you why the last deal died. They're busy, not effective.

Are they opinionated? Which brings me to the word I keep coming back to.

5. Taste is just being opinionated

Taste sounds fluffy. It isn't.

Taste is being fiercely opinionated. About what belongs on the page and what doesn't. About what great looks like. The more rogue and controversial the opinion, the better. In a pitch, the tell is a single sharp insight into the customer's problem that everything else in the company keeps bending back towards. Every decision traces to it.

Grace Brown at Andromeda refused to build a realistic humanoid. She made Abi deliberately cartoonish and animated, because lifelike robots trip the uncanny valley and make people recoil. Everyone else was chasing realism. She chased warmth - and it's why Abi goes viral every time it appears.

Andy Miller at Heaps Normal had a different heresy: non-alcoholic beer was never about health. It's about still belonging - holding a beer at the pub without having to explain yourself. He refused to market it like a wellness product, and built a movement instead.

The Naoumidis brothers bet Mindset Health on hypnotherapy - a thing most people write off as pseudoscience - because the clinical evidence was actually there and everyone else was too embarrassed to touch it.

And here's the uncomfortable bit:

There's no such thing as faking taste. You either have it or you don't. When you don't, everything comes out looking AI-generated and bland.

Which is exactly the trap of this new era. Performative hustle is the fake version too. Setting yourself "100 emails a day" and grinding it out feels like work. But if you're not tracking what converts, if you're not living for the demo itself, you're just generating inputs. Motion isn't distribution.

6. Now here's why all of this just got 10x more important

Building is free now.

Claude and Codex got so good that anyone can build almost anything. Which means the beautiful product is no longer rare. It's the default.

I see stunning AI products every single day sitting in a complete void. No oxygen. The AI does a brilliant job and nobody ever finds out.

I now get DMs constantly that go something like this:

"I built this AI product, I've got a couple of interested people, but I don't know how to get customers - do you want to buy it? Or come lead my sales and marketing?"

Last month alone I got three of these. One had a genuinely brilliant product - beautiful, fast, solved a real problem - and four users.

Founders trying to hand off the actual company, because the only part they enjoyed was the build.

You see it clearest in my own backyard. There are hundreds of AI consulting and implementation shops now, Hourglass included. The ones that survive aren't the best builders. They're the ones who cracked the other muscle - getting those first customers. Every week another one messages me asking how to do distribution and marketing well. That's the tell for the whole industry.

When everyone can build, building is worth nothing. Distribution is worth everything.

7. What to do on Monday

If you take one thing from a decade of my notes, it's this:

Start caring about your customers more than your product.

Figure out where they live. Then do everything you can to spend time with them. Set yourself a goal every single day and every single week to get one notch closer - a call, a coffee, a demo, a message. Whatever it is.

The build will keep getting easier. The customers won't come to you.

Go to them.

Michael Batko spent nearly a decade as CEO of Startmate, where he backed 250+ startups now worth over $4.5 billion. He now runs Hourglass AI (thehourglass.ai), an AI consulting services and implementation partner that helps companies find and automate their manual work.