← All terms·Digital attribution
Reference library · Digital attribution
TV-to-App Attribution
TV-to-App Attribution is the practice of linking a household's or device's exposure to a specific TV ad airing with a mobile app install or in-app event (registration, purchase) that follows it within a defined time window. Because TV exposure and mobile app events sit on different IDs, the link is probabilistic: airing logs or ACR/panel data are matched against app event timestamps by time, region, or device graph rather than a single deterministic user ID.
Worked example
A brand airs a 20" spot at 20:15 on a national channel in Poland, media cost €18,000. Baseline app installs run 45 per 10-minute slot; in the 30 minutes after airtime, installs total 380 against an expected baseline of 135 (3 x 45). Incremental installs = 380 - 135 = 245, giving a cost per incremental install of €18,000 / 245 ≈ €73.5 — the plan can now compare that CPI against the brand's other acquisition channels for the same market.
How it is used
Buyers run this as spike detection: ACR or set-top-box exposure logs are time-stamped against MMP postbacks, using a 15-60 minute attribution window and a matched control region or pre-campaign baseline to isolate the lift. Because there is no deterministic device-level ID connecting the TV set to the phone, results are read at DMA or regional aggregate level, not per-household. The recurring mistake is crediting every install inside the window to the spot without subtracting organic baseline volume, which inflates attributed installs and understates true CPI.
The common mistake
Always net out the organic baseline (via a control region or pre-airing period) before crediting a post-spot install spike to the TV exposure, or the attributed volume and CPI will be overstated.