D1, D7 and D30 retention calculator
Day N retention is the share of a cohort of new players who play again exactly N days after their first session. Enter the players who started on the same day and how many of them played on day 1, day 7 and day 30. The calculator gives each rate and fits a retention curve.
Fitted curve: retention on day t = 33.31% × t^(-0.573), a power curve fitted to your points. Our player LTV calculator uses the same curve.
How to use it
- In your game analytics, pick one start day and count the new players whose first session fell on it. Enter that number in "New players in the cohort (day 0)".
- Count how many of those same players played on day 1, day 7 and day 30, and enter them in "Played on day 1", "Played on day 7" and "Played on day 30".
- Read D1, D7 and D30 retention, and the players who did not play on day 30.
- Check the fitted curve against your three points, then copy the page link to share the result.
How it works
Day N retention = players from the cohort who played on day N ÷ players in the cohort × 100, rounded to two decimals. Day 0 is the day of the first session, so day 0 is always 100%. This is exact-day retention, also called classic or recurring retention: a player counts on day N only if they played on that day.
The curve is a power curve, retention(t) = a × t^(-b), fitted by least squares to the logarithms of your rates. The fit is our method, and our player LTV calculator uses the same one. A GameAnalytics paper on the retention curve uses the same power-function shape for game retention.
A cohort has a D30 value only once day 30 is over. Before that, leave the cohort out instead of entering a zero, or the rate looks far too low.
Dn retention = played on day n ÷ cohort × 100
did not play on day 30 = cohort - played on day 30
fitted curve: retention(t) = a × t^(-b), from a straight line through ln(rate) against ln(day)Worked example
Example with made-up inputs: a cohort of 2,000 new players
D1: 560 ÷ 2,000 × 100 = 28%. D7: 220 ÷ 2,000 × 100 = 11%. D30: 70 ÷ 2,000 × 100 = 3.5%.
On day 30, 2,000 - 70 = 1,930 players from the cohort did not play.
How to read the result
Compare cohorts with each other: one build against the next, one traffic source against another, or demo players against full-game players. In our practice, a drop in D1 points at the first session, and a drop between D7 and D30 points at the mid game or at content that runs out.
In the GameAnalytics paper, East Side Games aims for about 40% D1 and 20% D7 for its idle mobile games. That is a mobile target, not an indie PC one. Premium PC and web games often follow other patterns, so compare your own cohorts over time.
With small cohorts, a few players move the rate a lot: in a cohort of 50, one player is 2 points. Pool several start days, or wait for a larger cohort, before you act on a change.
Pick one definition and keep it
Some tools use rolling retention, which counts a player who returns on day N or any later day. Rolling retention gives higher numbers than exact-day retention. Compare games and builds only when they use the same definition.
Calendar days or 24-hour days
Analytics tools split days in two ways. Calendar retention counts a player on day 1 if they play on the next calendar date, so a player who starts at 11 PM can come back an hour later and count. Retention by hours counts day 1 as the second 24-hour window after the first launch. devtodev reports both. Use one rule, and one time zone, for every cohort you compare.
Count real players
Use the same player ID for every session, and remove test and bot accounts first. A player who reinstalls or plays on a second device can look like a new player if your analytics do not link the IDs.
The fitted curve and the share link
The tool fits a power curve to your D1, D7 and D30 and draws it from day 0 to day 30 next to your numbers, with the fitted formula under the chart. Copy the page link to share the result. The address keeps your inputs.
Limits
The calculator does not read your data or count players for you. It trusts that the day 1, day 7 and day 30 counts come from the same cohort and the same player IDs.
Estimate (fitted power curve). The curve between and after your three points is a fit, not a measurement, and it can be off for games with a long tail or a weekly rhythm. The tool does not give rolling retention, retention by hours or a benchmark for your genre.
Frequently asked questions
What is D1 retention?
D1 retention is the share of new players who play again on the day after their first session. If 1,000 players start on Monday and 300 of them play on Tuesday, D1 retention is 30%. It is the fastest signal of whether the first session works, because you can read it as soon as day 1 is over.
How do you calculate D7 and D30 retention?
Take the players whose first session was on one day. Count how many of them played exactly 7 days later, and divide by the cohort size. Do the same for day 30. With 2,000 players, 220 on day 7 and 70 on day 30, D7 is 11% and D30 is 3.5%. Players who skipped those exact days do not count.
Why can D30 be higher than D7?
With exact-day retention, a player can skip day 7 and return on day 30, so D30 can be higher than D7 in a small cohort, or after an update or a sale that brings players back. Rolling retention, which counts a player on day N if they return on day N or any later day, can never go up from one day to the next.
What is a good D1 retention for a game?
It depends on genre, platform, price and how players found the game. Some mobile studios aim for about 40% D1, but premium PC and web games behave differently, and we do not have a public benchmark that fits indie PC games well. Compare your own builds and cohorts over time, with the same definition each time.
What is the difference between classic and rolling retention?
Classic, or exact-day, retention counts a player on day N only if they play on that day. Rolling retention counts a player on day N if they play on day N or on any later day, so it is always the same or higher. This calculator uses exact-day retention. Manalens shows both, named recurring and rolling.
Sources
- Russell Ovans, "Game analytics 100: The retention curve" (GameAnalytics, April 2024), read
- devtodev documentation: SDK metrics (retention by hours and by calendar day), read
Last reviewed . All math runs in your browser. We do not send your inputs anywhere.
Next step
Manalens draws D1, D7 and D30 retention from your game events by daily or weekly cohort, with recurring (exact-day) or rolling retention, so you do not count players by hand. See retention →