Adjust's attribution methods
Here, you'll learn how and when Adjust uses different attribution methods.
For instructions on managing attribution settings, see these articles:
Adjust supports two attribution methods—device matching and probabilistic matching for both clicks and impressions. The type of conversion and quality of user engagement determine which method we use.
- We use device matching and probabilistic matching to attribute installs (first app opens)
- We use device matching to reattribute (assign new attribution sources to) inactive users
Adjust follows this order when making attributions:
- Click-based device matching
- Click-based probabilistic matching
- Impression-based device matching
- Impression-based probabilistic matching
At install, we look for clicks with a matching device identifier. If there aren't any, we move on to clicks with matching device characteristics. If there are no eligible clicks, then the process repeats with impressions. Adjust always prioritizes clicks (the more deliberate user engagement) and device matching (the most reliable attribution method).
Device matching is Adjust's main attribution method. We collect a unique identifier from recorded engagements and installs, and if both IDs match, we can attribute that engagement to the install. With a 100% accuracy rate, click-based device matching is the most reliable attribution method.
Adjust uses these device identifiers for attribution:
- Advertising IDs like the IDFA (for iOS) and GPS ADID (for Android).
- Device IDs like the IDFV (for iOS) and Android ID.
- Adjust reftags, which are unique IDs we create for clicks/impressions.
Here's how Adjust uses device matching with advertising IDs and with the Android referrer:
The advertising ID matching workflow
Probabilistic matching is Adjust's secondary attribution method. We use device details to attribute installs to clicks and impressions with matching characteristics.
Adjust uses DeviceAtlas to enrich and verify device details like:
- Device type (e.g., phone)
- Device name (e.g., Samsung Galaxy S7)
- Operating system (e.g., Android)
- Operating system version (e.g., 7.1.2)
- IP address (e.g., 188.8.131.52)
- User agent (e.g., browser and operating system)
Since some of these details can change (e.g., IP address), probabilistic matching is most reliable over short time periods. Adjust only attributes a click or impression if all of these details match with the install.
Reattribution is when an inactive user engages with a retargeting campaign, returns to your app, and gets assigned a new attribution source in Adjust. We reattribute users through clicks, deeplink reattribution, and impressions.
Adjust records an impression when a user views your mobile ad. You can turn on (or turn off) impression-based attribution in your Adjust dashboard anytime. We only consider impressions for attribution when there are no eligible clicks.
After you turn on impression-based attribution in Adjust, you can share impression tracker URLs with your ad network. There are two ways to find your impression tracker URLs in Adjust:
- Finish creating a new tracker in the Campaign Wizard
- Go to your app's settings, select Tracker URLs, and select Settings (gear icon) next to any tracker
TV attribution is matching mobile app installs to TV impressions. To get started, turn on Adjust’s integration with any TV tracking partner in your app’s partner setup. Adjust is fully integrated with TV partners like Admo.TV, Teads, and TVSquared.
Note: TV tracking attribution is available with a Business Pro (or higher) pricing plan.
Only organic installs (i.e., installs that Adjust can't attribute to a click or impression) are eligible for TV attribution. Adjust sends your organic installs to any connected TV partners, and they have 7 days to make attributions based on their own criteria and resources. If the TV partner claims an install, we confirm that their attribution data meets our requirements, then update the attribution from organic to TV in your Adjust dashboard.
How do TV tracking partners make attributions?
TV attribution methods vary from partner to partner. Here are some of the ways TV partners make attributions:
- By comparing the volume of user engagement after a TV campaign to the typical volume of user engagement for that region, time, and/or day
- An algorithm that analyzes the probability that a specific user was influenced by a TV campaign
- Audio matching technology that identifies the audio track of a TV campaign