Attribution
Part of Gaming campaign attribution
Separating app installs from post-install value
Define an acquisition cohort, identify useful in-app actions and compare installs with later behaviour on matching terms.
An install counts as acquisition under a given rule. Post-install value covers what acquired users do later. Keep acquisition and later actions separate, define one cohort, and give every cohort the same time to reach the outcome before comparing.
Choose the acquisition event
Decide whether the cohort begins with downloads or first opens. Apply that definition consistently. Record the source, platform, attribution rule and reinstall treatment. Never combine downloads and first opens into one new-user total.
A network’s attributed acquisitions can be used to assess what it credited. App events can describe later behaviour. A rate for attributed acquisitions requires a supported way to identify those acquisitions in the later event data. If only separate aggregate totals exist, show them separately; their ratio is not a cohort rate.
Define a useful later action
Select actions that reflect the game’s objective, such as completed onboarding, a return to play or a completed purchase. Specify the trigger, whether repeat events count, and which app versions record it. A button tap must not count as a purchase.
For purchase outcomes, define which transactions count and how events are recorded. Check that the same purchase is not counted more than once. State any known coverage limits.
| Cohort measure | Calculation | Essential qualification |
|---|---|---|
| Activation share | Acquired users completing the defined action ÷ acquired users eligible to complete it | Use a linked cohort and the same observation period. |
| Purchasing share | Acquired users with a completed purchase ÷ eligible acquired users | Define completed payment and whether repeat purchases count. |
| Observed value per acquisition | Valid recorded value from the cohort ÷ acquired users | State currency, refunds, coverage and cohort maturity. |
These are analytical definitions, not universal platform metrics. State whether the denominator contains downloads, first opens or attributed users. Keep numerator and denominator within the same platform and eligible population.
Key Metrics for Post-Install Value Analysis
- Activation Share
- Percentage of users completing onboarding
- Purchasing Share
- Proportion of users making a completed purchase
- Observed Value per Acquisition
- Average revenue from attributed users
Compare cohorts at the same age
Users acquired yesterday have had less time to finish onboarding or purchase than users acquired weeks ago. Compare acquisition cohorts after the same observation period. Keep that period distinct from the attribution window that determines whether an ad receives credit.
Break out materially different app versions, devices and Australian eligibility rules where data supports it. If onboarding changed during the comparison, label the change. A lower later-action rate could reflect a different audience, changed experience or a missing event; the ratio does not identify the cause.
Use the result
Report acquisition volume and cost beside the number reaching the chosen useful action, with the cost boundary stated.
Where later value is sparse or delayed, an earlier verified action can help diagnose the path, but it should not be renamed as value. Attributed purchases or completions remain observed outcomes under stated credit rules. Estimating how many were added by advertising requires a suitable comparison.
