Why Did 70% of Players Vanish on Day One?
A day-one retention number is a fever reading, not a verdict. When 70% of players don't come back, the game isn't "bad" — something specific in the first session is ushering them out. The good news: the first session is the most fixable part of your game, because you control every pixel of it. Here's how to find the leak instead of guessing.
Q: What's a "normal" day-one retention?
For a healthy mini-game, day-one retention usually lands between 35% and 50%. Below 30% and you're keeping almost no one; above 55% and something about your first session is working absurdly well. The number isn't the diagnosis, though. Two games at 40% D1 can be in completely different health, because what matters is where those players go after the first session.
Q: How do I find which step is leaking?
Don't guess. Instrument the first session like a checkout funnel: time-to-first-action, level-1 completion, first reward claimed, first failure, first ad seen. Most teams discover the leak isn't the game — it's a 12-second load, an account wall before play, or a tutorial that lectures instead of letting anyone touch anything. The first 90 seconds decide the week.
Q: Is it a difficulty spike or just a bad first impression?
Those are different cures. A difficulty problem shows up as players who reach level 3–5 and quit; a first-impression problem shows up as players who never finish level 1. If your quit curve is front-loaded into the first session, it's onboarding, not balance. Fix the welcome before you touch the difficulty dial — balance tweaks won't save a game people abandon in minute two.
Q: What should the first 90 seconds actually contain?
One satisfying action, fast. Let the player win something trivial in the first ten seconds so the loop is legible, then withhold a little so they want more. The mistake is a slow reveal: logo, menu, account, settings, then maybe you play. By the time they touch the fun, half are gone. The fun should be the first thing they meet, not the prize for surviving the menu.
Q: What A/B test moves day-one retention the most?
The test that removes friction, not the one that adds polish. Kill the account wall, shorten the load, or skip the tutorial — those beat "prettier first screen" every single time. And measure against players who return on day 7, not installs, or you'll optimize a vanity curve that looks great and earns nothing.
| Step | Where teams lose players | Quick fix |
|---|---|---|
| Launch → first tap | >12s load, account wall | defer account, lazy load |
| First level | tutorial lecture, no agency | playable first 10s |
| First failure | unfair spike | forgiving early levels |
| First reward | reward hidden | surface it early |
| First return prompt | none at all | notification + streak |
