Distribution Beats Product, and AI Made It Harsher

The short version

Adobe taught me that distribution beats product, and I spent a decade wrongly unlearning it at startups, where I fell in love with building the better thing. Now AI has made product cheap to build and distribution the only durable moat, so the lesson is back and harsher than ever. The best product rarely wins. The best distributed product wins, because distribution decides how many people ever try the thing and how cheaply you reach the next customer. AI collapsed the cost of building, which means product quality gets copied almost immediately and stops being defensible. What does not get copied is the channel, the brand, the install base, and the customer relationships. As building gets cheap, the moat moves entirely to distribution. Here is the lesson I had to learn twice, and why this time there is no excuse to forget it.

I learned distribution beats product at Adobe, watching technically superior competitors lose to products that were merely good but reached everyone. Then I spent ten years at startups carefully unlearning it, because at a startup you are the underdog with the better mousetrap, and the better mousetrap is the most seductive story in business. AI just proved Adobe right in the cruelest possible way.

What Adobe taught me

At Adobe I watched a pattern repeat until I had to accept it. The best product, on the merits, regularly lost.

A competitor would ship something technically superior, a cleaner tool, a smarter feature set, and it would lose to an Adobe product that was good enough and everywhere. Bundled. Defaulted. Embedded in the workflows people already lived in. Taught in every school. Distribution was not the thing that helped the better product win. Distribution was the thing that beat the better product.

The mechanism is simple and it compounds. Distribution decides how many people ever try your product, and how cheaply you can reach the next one. A great product that reaches a thousand people loses to a good product that reaches a million, every time, and the million-reach product uses that reach to fund reaching the next million. Product advantages erode as competitors copy them. Distribution advantages compound as scale lowers your cost to acquire. One curve goes down, the other goes up.

“

The best product rarely wins. The best distributed product wins. Product advantages erode as they get copied. Distribution advantages compound as they scale.

”

I knew this. I had seen it work on companies with genuinely better products than Adobe's. And then I left to do startups and spent a decade forgetting it on purpose.

The decade I unlearned it

At a startup, the distribution lesson is inconvenient, so you talk yourself out of it.

You are the small team with the better idea. The incumbent is bloated and behind. The whole founding story is that the superior product will win on merit, and you need to believe that story to get out of bed. So you pour everything into product, you obsess over the thing being better, and you tell yourself distribution is a later problem, a go-to-market detail you will sort out once the product is undeniable.

I did this. At more than one of my startups I built something genuinely better than what the incumbent had and watched it struggle to reach anyone, while a worse-funded but better-distributed competitor ate the market we were technically winning. The product was right. The reach was wrong. And the cheap-capital decade let me hide the mistake, because you could buy distribution with paid growth and call it product-market fit when it was really just an open funding spigot.

The hard truth, which I now believe, is that founders systematically underrate distribution because building is the part they love and control. We are builders. We assume a great product distributes itself. It almost never does. I had to learn the Adobe lesson a second time, the expensive way, on my own startups.

AI made the lesson harsher

Here is why this is not a nostalgia piece. AI did not just confirm the distribution lesson. It made it brutal.

For all of my career, product was at least somewhat defensible. Building the better thing took real time and money, so a head start on product bought you a window before competitors caught up. Distribution mattered more, but product still mattered. The better mousetrap had a shelf life of advantage.

AI collapsed that. When anyone can build a comparable product in weeks, the product head start has a shelf life measured in days. Your clever feature gets cloned almost immediately, because the thing that used to take a team a year now takes a competitor an afternoon with the same tools you used. Product quality stopped being a moat because it stopped being scarce. What is left, the only thing that does not get instantly copied, is distribution. The channel. The brand. The install base. The customer relationships and the trust. None of that clones in an afternoon.

“

AI made product cheap to build, so product stopped being a moat. The only thing that does not clone in an afternoon is distribution. The moat moved, entirely.

”

So the Adobe lesson came back wearing a harsher face. It used to be distribution beats product. Now it is distribution is the only moat, because product is no longer scarce enough to be one. The founders still telling themselves the better product will win are about to learn the lesson I learned twice, on a faster and less forgiving clock. This is the same reason I argue gross margin and distribution efficiency are now the PM's job, not a finance afterthought.

Where the lesson has limits

I will keep myself honest, because the cheap version of this take says product does not matter and that is wrong.

Product is the price of entry. A product that is bad does not get to play, distribution or not. And there is one product property that still functions as a moat: taste, the judgment about what to build and what to leave out, which does not clone because it is not a feature, it is a sensibility. That is the argument I make in why taste is the last moat, and it is the one part of product that survives the copying.

But notice that taste is upstream of distribution, not a substitute for it. Taste tells you what to build. Distribution decides whether anyone ever sees it. You need taste to deserve the market and distribution to actually take it, and in a world where build cost went to nearly zero, the binding constraint is almost always the second one. A tasteful product nobody can reach loses to a tasteful product that is everywhere.

What this means for how you build

The practical shift is to treat distribution as a first-class product surface, not a go-to-market afterthought you bolt on later.

Build the product so it reaches the next customer cheaply. Embedding, defaults, virality, partnerships, the brand people already trust. Design the distribution into the product the way you design the features, because in the AI era the distribution is the more durable half of what you are building. In my work at Falkster.AI, I think about how a customer outcome turns into a same-day prototype and how that prototype reaches the next customer with the same effort, because the build is the cheap part now and the reach is the moat.

Measure distribution efficiency the way you measure product quality. How cheaply do you acquire the next customer, and is it getting cheaper as you scale. That curve, not your feature set, is what competitors cannot copy in an afternoon.

Pick one thing to try this week

Take your product's single best feature, the one you are proudest of, and ask one honest question. How long would it take a competent competitor with today's AI tools to copy it. If the answer is weeks or less, that feature is not a moat, and you have been treating it like one. Now ask the harder question. What about how we reach customers does not clone that fast. Spend this week making that thing one notch stronger, embedding, brand, a channel, a default. That is the moat. The feature was just the price of admission.

Sources: a16z, on distribution and go-to-market · First Round Review, on building a distribution advantage · Lenny Rachitsky, on growth and distribution