Challenges and Concepts

As users interact with your product and marketing activities across multiple platforms and devices, cross-device attribution aims to tie it all together, measuring the contribution of each campaign as well as the true lifetime value of your users.

There are obvious challenges: users may still be anonymous when interacting with ads and marketing campaigns, while changes in privacy regulation, cookies, and advertising IDs make device graphs almost obsolete unless you're Google or Facebook.

However, privacy laws have also created an environment wherein many products, users are authenticating very early, for example in order to provide consent. This allows advertisers to establish a clear identity link across devices that is also privacy-friendly by design. 

Singular's Cross-Device Methodology

Users vs. Devices

Standard attribution tools for mobile are based on advertising IDs - IDFA on iOS and Google Advertising ID on Android - with some additional identifiers used occasionally. These tools often define conversion as the app install (or first app open), and attribute every in-app activity to the marketing campaign that resulted in the install. This model is problematic when users are active on multiple devices, as well as when the truly meaningful conversion only takes place after the install - for example, when a user registers or makes a purchase.

Nonetheless, device attribution is still meaningful: partners are mostly platform-specific and require events to fire back with advertising IDs. And campaigns that engage an existing user on a new device may still be important to track for that user's retention and LTV.

This is why Singular's attribution system distinguishes between device attribution and user attribution, and tracks them in parallel. As a marketer, you can see how many new devices your mobile campaign yielded, while also tracking how many new users these ended up producing. Both measurements are valuable as these users will continue to interact with your product across every device they own.

Establishing Identity

Regarding the ownership of identity, Singular defers to the advertiser. Singular does not mix device data across customers and the advertiser is responsible for determining which device is associated with which user and whether that user can be tracked.

When the advertiser recognizes users through sign-up and sign-in events, the advertiser reports the events to Singular to provide a pairing of advertising ID + user ID. If the user ID is new to Singular, our attribution platform recognizes a new user and attributes the user's activity to the campaign that led to the sign-up. At the same time, Singular keeps tracking device attributions as always. These two parallel streams of device attributions and user attributions are at the core of our cross-device product.

Cross-Device Reporting

The cross-device report provides the following metrics:

1. App Metrics:
   • App Installs - New device installs (devices recognized for the first time)
   • App Re-Engagement - Existing devices re-engaging with a new campaign
   • Total App Conversions - The sum of installs and re-engagements
     Note that App Metrics include all app-based conversions (e.g. both iOS and Android)
2. Web Metrics:
   • New Visitors - These are sessions from browsers recognized for the first time.
   • Re-Engaged Visitors - These are sessions from known browsers re-engaging with a new campaign, or re-engaging with the website after being inactive (see more about this in the Web Attribution FAQ).
   • Total Web Conversions - The sum of new and re-engaged visitors.
3. Cross-Device Metrics:
   • Total Conversions - Total conversions across all platforms.
   • User Attributions - New users that have signed up (i.e., for which a Custom User ID has been generated).

Cohort Reporting

Mobile attribution providers have traditionally calculated cohorts based on the date of the app install, while web attribution providers have looked at the click (which leads to a page visit) as the point of conversion, and have calculated cohorts based on the click date.

Singular's cross-device attribution defines the conversion event as the time when a user signs up. Consequently, User-Based Cohorts are calculated based on the sign-up date, regardless of which platform it occurred on. The user may then sign in on multiple platforms, but their identity has already been established, and all events will be tracked based on the initial sign-up/registration.

Registration-based cohorts provide tremendous value as LTV and other metrics can be calculated based on users across all platforms. Marketers can now measure the real yield of every marketing activity accounting for every platform users interact with.

To view cohorts by user registration, use the toggle in the cross-device report:

Device-based cohorts are still provided for both backward compatibility as well as understanding what partners are receiving, given that partners are mostly platform-specific.