Answer in brief
Re-engagement attribution credits an ad when a person who already has the app opens it, or acts in it, after interacting with that ad. It is measured apart from install attribution, with its own window, so that returning users are not counted as new ones. It answers whether retargeting spend brings back people who would not have come back alone.
Install attribution asks who brought the user. Re-engagement asks who brought them back.
An install is counted once. Everything a user does afterwards is normally credited to the channel that won that install. That rule breaks down when you pay for ads aimed at existing users: a discount shown to lapsed shoppers, a reminder shown to people who left a cart. If the purchase that follows is credited to the original install channel, the retargeting campaign looks like it did nothing.
Re-engagement attribution fixes this by opening a second attribution record. The retargeting ad becomes the owner of what the user does for a set period after they return.
The three settings that define it
| Setting | Question it answers | Effect of a larger value |
|---|---|---|
| Re-engagement lookback window | How long after the ad click can the app open be and still count? | More opens are credited to retargeting. |
| Inactivity period | How long must the user have been away before a return counts? | Fewer, more meaningful re-engagements. |
| Re-engagement attribution period | For how long are later events credited to the retargeting ad? | More revenue moves from the install channel to retargeting. |
Names differ between tools, and not every tool exposes all three. The ideas are the same everywhere: how fresh the click must be, how absent the user must have been, and how long the credit lasts.
How the match is made
There is no store visit, so the store referrer plays no part. The usual carrier is a deep link. The person taps the ad, the link opens the installed app directly, and the SDK reads the campaign details from the link at that moment. Where the ad platform shares a device identifier or a click ID, the MMP can also match a later open that did not come through the link.
Check whether views count at all. Apple's AdAttributionKit, for example, does not create re-engagement postbacks from view-through ads.[1]
Why re-engagement numbers are easy to inflate
Existing users open the app without any help. Show an ad to your most active users and many will open the app within a day, as they would have anyway. Each of those opens is a re-engagement on paper. Three guards keep the number honest:
- Exclude recently active users from retargeting audiences, or require an inactivity period.
- Judge the campaign on a paid event, such as a purchase, and not on app opens.
- Hold back a small group that sees no ads and compare how many return.
What to do next
- Decide what counts as a lapsed user for your app. A daily-use app and a travel app need very different inactivity periods.
- Give every retargeting ad a deep link, so the open can be matched deterministically.
- Report installs and re-engagements in separate rows.
Check that your plan measures it at all. Airbridge lists retargeting and re-engagement attribution, for campaigns that target existing active users, in its plan comparison.[2] For how the returning user's later purchases feed the ad platform, see the entry on the postback. For judging whether those users stay, see retention.
Questions and answers
What is the difference between re-engagement and reattribution?
Re-engagement usually means an ad brought back someone who still has the app installed. Reattribution, also called reactivation, usually means a lapsed user came back or reinstalled after a period of inactivity. Vendors draw the line differently, so check the definition in your MMP.
Does a re-engagement count as an install?
No. The person already installed the app once. An MMP records a re-engagement as its own conversion type so install counts and cost per install stay clean.
Continue reading
Guide
Last-touch attribution and why it is the common defaultLast-touch attribution gives all credit for an install to the final qualifying click or view. How the winner is picked, why it is the common default, and its limits.
Guide
Deterministic versus probabilistic matchingDeterministic matching ties an install to a click with an exact identifier. Probabilistic matching infers the link from device signals. When each is used.
Guide
View-through attributionView-through attribution credits an install to an ad that was shown but not clicked. How the view window works, the risk of over-crediting, and how to set it.
Guide
The Google Play Install ReferrerThe Google Play Install Referrer API gives an app the referrer URL, click time and install-begin time of its own install. How MMPs use it for attribution.
Guide
Self-attributing networksA self-attributing network decides for itself whether its ad led to an install and tells the MMP the answer. What that means for your reports and settings.
How it works
Best MMP attribution starts with knowing how the match is madeHow an MMP attributes an install on Android and iOS, what lookback windows do, and which settings to check before you trust a report.
Sources
- Apple: https://developer.apple.com/documentation/adattributionkit/receiving-ad-attributions-and-postbacks Checked 2026-11-10.
- Airbridge: https://www.airbridge.io/en/pricing Checked 2026-11-10.
Cite this page
Airbridge. “Re-engagement attribution.” bestmmpattribution.com. Reviewed 2026-11-10. https://bestmmpattribution.com/re-engagement-attribution