Distributions: Why User Data Lies

By the end of this lesson, you should be able to: explain why user metrics are almost always lopsided, say what the average hides, pick the right summary for the question you're answering, and give a better answer than "just drop the outliers".

The average user doesn't exist

Playtika makes free-to-play mobile games, like casino and puzzle games. In the first quarter of 2026, it reported these numbers.1

So on an average day, Playtika earns about $0.94 for every player who shows up. But only 4.5% of players pay anything that day. That's about 1 in 22.

Now let's do one small calculation. If $0.94 per player comes from just 4.5% of players, then each paying player brings in about $0.94 ÷ 0.045, which is roughly $21 a day. Everyone else brings in nothing.

So who is the "average player" who spends $0.94? Nobody. About 95 out of 100 players spend $0, and the rest spend around $21 on average. The average sits in a gap where almost no actual player is.

This isn't special to games. It's what user data almost always looks like.

What lopsided data looks like

User data almost always has this shape: most users do a little. A few users do a lot. And those few account for a big share of the total.

You see it in spending, like the game above. You see it in activity too: most people open an app a few times a month, and a small group opens it many times a day. You see it in content: most people post nothing, and a small group posts most of what everyone else reads.

This shape has a name. It's called a heavy-tailed or skewed distribution. The "tail" is the long stretch of a few users with very large values.

Why does user data come out this way? A few reasons.

People need the product by very different amounts. A small business that runs its whole team on a work tool uses it far more than someone who tried it once.

Use builds on itself. People who use a product more get more value from it, so they use it even more. Small differences at the start grow over time.

There's a floor, but no ceiling. Nobody can spend less than $0 or open an app fewer than zero times. But there's no limit at the top. So the data piles up near zero and stretches out to the right.