Measuring retention on game servers (without lying to yourself)

Most server owners watch a single number: players online right now. It feels good when it is high and bad when it is low, and it tells you almost nothing about whether your community is growing.

Why concurrency lies

A weekend spike doesn’t mean you’re growing. It means it’s the weekend. To know whether last month’s event actually kept people around, you need to follow cohorts: groups of players who first joined in the same week, tracked over time.

Reading D1 / D7 / D30

• D1: did they come back the next day? This is mostly about first-session experience: queue times, spawn confusion, griefing.
• D7: did they make it through the first week? This is about whether they found a reason to log in again: a base, a clan, a goal.
• D30: did they become a regular? This is your real growth engine.

On wipe-driven games (Rust, ARK, 7 Days to Die) read retention per wipe, not across them. A reset deliberately churns players; comparing wipe-to-wipe is the only honest baseline.

What to do with it

If D1 is fine but D7 collapses, your onboarding works but your mid-game doesn’t. If D1 is weak, fix the first ten minutes before you spend a cent on promotion. Tickr computes these cohorts automatically. No spreadsheets required.