I Scored 671 Products to Find Out Why Things Go Viral. The Answer Was Humbling.

I spent a year convinced that great products win.

Build something beautiful. Make it fast. Nail the onboarding. If the product is good enough, people will come. That’s the story we all tell ourselves.

Then I scored 671 real products — companies from Crunchbase, Product Hunt, and my own research, tracked over three years — across five categories. (Some of that data was AI-generated, which caused its own problems.) Team quality. Product excellence. Distribution. Market timing. Competitive moat.

I compared every score against what actually happened. Did the product explode? Grow steadily? Spike and die? Go nowhere?

The results wrecked everything I believed.

The number that broke my brain

Moat — how hard it is for users to leave and for competitors to copy you — predicted 38% of whether a product would explode.

Market timing predicted 31%.

Together, those two things account for nearly 70% of the prediction. Before you even look at the product itself.

Product quality? Ten percent.

The founding team? Zero. Not low. Not “marginal.” The math found that knowing who built it tells you nothing about whether it’ll blow up.

I stared at that for a long time. Not the productive kind of staring — the kind where your brain needs a minute to let go of something you believed for years.

Someone at a demo asked me: “You’re telling me the team doesn’t matter?” “That’s what the math says,” I told him. He didn’t look convinced. I wasn’t either, at first.

Why this makes sense (even though it hurts)

Think about Slack.

Is Slack the best chat app ever made? Not really. There are dozens of alternatives. Some are arguably better. Cheaper. Faster.

But every team that uses Slack has years of message history in it. Hundreds of integrations wired into it. Custom workflows that took months to build.

New hires learn Slack on day one and it becomes muscle memory by day ten.

The product got people to try it. The moat is what made leaving feel like moving apartments. You could do it. You just really don’t want to.

Moat breaks down into five things: how strong the network effect is, how painful switching would be, how much of your data is locked in, how deeply it’s woven into your daily routine, and how hard it would be for a competitor to just copy it.

I see this pattern over and over in the data. A decent product with a deep moat beats a beautiful product with no moat. Every time. It’s not close.

The market doesn’t check resumes.

The timing question

Zoom didn’t win because it was the best video tool in 2020.

Zoom won because in March 2020, every human on the planet suddenly needed video conferencing and the existing options made people want to throw their laptops out the window. Zoom showed up at the exact moment the frustration peaked.

I see this across hundreds of cases. Products that launch when people are genuinely angry with the current options grow dramatically faster than products that launch into “what we have is fine” markets.

You can build the best thing in the world. If nobody’s frustrated with what they already have, your beautiful thing will grow slowly.

That’s not a product problem. It’s a timing problem. And you can’t manufacture timing. You can only be ready when it arrives.

The founder thing

This is the one that makes people angry.

The data says founding team quality has zero connection to whether a product blows up.

I don’t mean “it helps a little.” I mean the model assigned it a weight of zero after seven calibration rounds across 671 cases — rounds I later discovered were grading themselves with yesterday’s answer key. As in: once you know the product’s moat and the market’s timing, knowing who built it tells you nothing additional about whether it will explode.

A first-time founder with no funding can ride a perfect wave into a moated product and blow up. A dream team with $50 million can launch into a satisfied market and watch their beautiful product go nowhere.

We don’t like this because we want to believe the best people win. Sometimes they do. But the market doesn’t check resumes.

The fast starts that mean nothing

Some products in my data grew fast and then collapsed just as quickly.

Every one of them had the same profile: high virality, low moat. People heard about it, tried it, liked it fine — and left the moment something newer showed up. There was nothing keeping them.

Virality without moat is a sugar rush. The energy arrives fast and disappears faster. The data is brutal about this.

What I actually changed after seeing this

I stopped polishing and started locking in. Every hour I used to spend making things 5% prettier, I now spend building something that makes users 50% stickier.

I stopped lying to myself about market timing. If people aren’t frustrated, no amount of product quality creates urgency. I’d rather wait for the right moment than launch into indifference.

I stopped evaluating startups by their team bios. The data says it doesn’t predict what I thought it predicted. That was uncomfortable to accept, because I’d been doing it for years.

Is this prediction system perfect? No. It gets the exact tier right 65% of the time. Within one tier, 93%. Honest enough to trust. Imperfect enough to stay humble about.

But the biggest thing it taught me isn’t in the numbers.

It’s that my instincts about why products succeed were almost entirely wrong. And I’d been wrong for a long time without knowing it. Which makes me wonder what else I’m wrong about.

---

Thinking about product-market fit? I’ve stared at this data more than is probably healthy — mo@fadaly.net.