Convert a new-player cohort and D1, D7 and D30 retention percentages into retained-player counts. The same percentage can mean very different operational outcomes at different acquisition volumes. Read retention together with cohort size, acquisition source, platform, territory and build version. Relative reduction is (1 − D30 / D1) × 100 and is undefined when D1 is zero. A negative result means the D30 rate exceeds D1. This does not measure churn among D1 players: players returning on D30 need not be the same people who returned on D1. When evaluating a patch or onboarding change, compare similar acquisition cohorts over comparable dates. Collapsing mixed user groups into a single percentage can produce the wrong product decision.Game Retention Calculator
Retention percentage is not enough
A practical measurement rule
QUBU / PLAYER TESTS
Turn the retention result into a test plan
Record the player cohort, acquisition source and game build behind the D1, D7 and D30 counts. Then choose one change, such as the first-quest flow, and use the GDD’s success-measurement field to write the hypothesis, comparison groups and outcome to track. The next measurement will have a specific question to answer.
How to calculate D1, D7 and D30 retention
Game retention measures the percentage of players who return on a specific day after first playing. Use the same install cohort for both numbers: divide the players from that cohort who return on Day N by the cohort’s total new players, then multiply by 100.
Formula: Day N retention = returning cohort players on Day N ÷ total new players in the cohort × 100.
D1, D7 and D30 retention definitions
- D1 retention shows whether the first session, onboarding and early value give players a reason to return the next day.
- D7 retention shows whether the core loop, progression and first-week goals continue to create value.
- D30 retention shows whether long-term progression, content cadence and social or competitive systems sustain a habit.
Retention calculation example
If 1,000 players install on Monday and 320 of those same players return on Tuesday, D1 retention is 32%. If 120 return seven days later, D7 retention is 12%. Do not mix calendar-day active users with the original install cohort.
How to use the result
Compare cohorts after meaningful product changes rather than treating one benchmark as a universal target. Segment by platform, country, acquisition source and build version. Pair retention with session quality, progression, crashes and the events in our core gameplay loop guide. During a limited release, use the Steam wishlist measurement guide to decide whether the result is stable enough to act on.
Retention calculator FAQ
What is a good D1, D7 or D30 retention rate?
A useful target depends on genre, platform, audience and acquisition source. Compare like-for-like cohorts and use your own improving baseline before relying on a broad industry benchmark.
Should I use classic or rolling retention?
Classic retention counts players who return on the exact day. Rolling retention counts players who return on that day or any later day. State the method in every report so teams do not compare different definitions.
A complete D1, D7 and D30 cohort example
This is an illustrative dataset, not a QUBU project result or an industry benchmark. Start with 1,000 distinct new players and count distinct returning players from that same cohort.
| Metric | Returning players | Calculation | Retention |
|---|---|---|---|
| D1 | 320 | 320 ÷ 1,000 × 100 | 32% |
| D7 | 120 | 120 ÷ 1,000 × 100 | 12% |
| D30 | 50 | 50 ÷ 1,000 × 100 | 5% |
A player can count on several different days, but only once within each day’s numerator. Do not add D1, D7 and D30 percentages together: they are separate observations of the same cohort.
Measure retention without changing the denominator
- Define entry and return events. For example, first playable session and a later playable session. Record the player identifier, event time, build and acquisition source; exclude test accounts.
- Choose one day definition. Calendar days use the project’s timezone; elapsed 24-hour windows follow each player’s entry time. Keep this consistent between releases.
- Wait for the cohort to mature. A cohort that joined five days ago cannot provide a completed D7 or D30 result. Mark unavailable observations as pending, not zero.
- Count players, not sessions. Ten sessions from one player still represent one returning player on that day. Combine fully observed cohorts using total returning players divided by total cohort players, rather than averaging unequal cohort percentages.
What should you test after a retention drop?
Use the drop to choose a test, not to declare a cause. For D1, inspect first-session crashes, tutorial exits and time to the first meaningful action. For D7, inspect progression stalls and whether the core loop still offers useful choices. For D30, inspect longer-term goals and the content players actually reach. Compare matching acquisition segments and build versions before attributing a change to your design.
How does this calculator relate to analytics?
The calculator estimates returning-player counts from the cohort size and retention percentages you enter. It does not import event logs or validate your tracking. Calculate the observed percentages in your analytics system first, then use the tool to compare planning scenarios.
Measurement references: Amplitude’s official documentation explains exact-day and on-or-after retention and calendar days versus 24-hour windows. The worked dataset and test checklist above are editorial examples.
