What Cross-Device Attribution Is
Cross-device attribution links the different devices one person uses, so their marketing journey can be measured as one path. Someone sees a video ad on a smart TV, searches the brand on a work laptop, reads reviews on a tablet and finally installs the app on their phone. Measured device by device, that is four unrelated people. Measured across devices, it is one journey with four touchpoints.
For mobile growth teams the most common case is web to app: a user clicks an ad on desktop, signs up on the website, then installs the app on their phone days later.
How Devices Get Linked
Deterministic links use an identifier the user shares on both devices:
- a login or customer user ID, when the user signs in on the web and in the app,
- a hashed email or phone number captured on both sides,
- an account on a logged-in platform, which is how Meta and Google link devices inside their own systems.
These links are accurate when they exist, but they only exist for users who identify themselves.
Probabilistic links estimate that two devices belong to the same person from shared IP addresses, location and behavior patterns. They reach more users but make more mistakes, especially on shared networks, mobile carriers and offices. Privacy rules and platform policies have narrowed what is allowed here, and probabilistic device graphs are less common than they were.
Why It Matters
- Upper-funnel channels get credit. TV, video and desktop campaigns often start journeys that end on mobile. Without cross-device links they look like they drive nothing.
- Organic installs shrink to their real size. Some organics are paid journeys that switched devices.
- Web-to-app funnels become measurable. Products that sign users up on the web and activate them in the app can see the whole path. See web-to-app attribution.
Where It Stops
Cross-device attribution can only link what the data links. Users who never sign in, decline consent or use different emails on each device stay unlinked. On iOS, aggregate frameworks such as SKAN do not expose user-level data at all.
It also explains a common attribution discrepancy. Meta and Google can credit an ad click on one device for an install on another, using their own logged-in data. An MMP that matches signals on the same device will credit fewer of those installs, and the gap is expected.
Cross-Device Attribution in Linkrunner
Linkrunner's install attribution matches signals on the same device: the Google Play Install Referrer, ad click IDs, device IDs where the user allows it and Apple AdServices first, with IP and user-agent matching only when no deterministic signal exists.
Cross-device journeys are part of multi-touch attribution, available per project on request. They connect through your customer user ID and hashed email or phone, never IP, so linked journeys come from identifiers your users shared with you. The default lookback is 30 days with a 7-day revenue window, and reports show assists, journeys and channel synergy. The Web SDK, in beta, captures web touchpoints such as UTMs and ad click IDs so web visits can join the journey. See web-to-app and analytics.
Practical Steps
- Ask users to sign in on both web and app where it makes sense, and pass the same customer user ID to your measurement tools.
- Hash emails and phone numbers consistently on every surface, with consent.
- Report cross-device conversions separately from same-device ones, so you can see how much each adds.
- Expect logged-in platforms to report more conversions than your MMP, and reconcile by campaign rather than total.