Overview

Approach

CTV attribution works differently than mobile attribution; smart TVs do not use mobile device IDs like IDFA or GAID. Measurement platforms often rely on household IP addresses to connect a TV ad view to a mobile app install. For example, if someone sees a TV ad and then installs the app on a phone using the same home internet connection, the install may be credited to the CTV campaign.

Fraudsters used shared residential IP addresses to make fake installs appear as if they came from real households exposed to FeedMob’s Samsung or LG ads. In other words, while our ads were being served, another media partner was fraudulently hijacking credit from those real impressions.

Solution

FeedMob worked with the client’s MMP to identify and eliminate fraudulent CTV installs, restoring the campaign to a healthy performance level. The team reviewed abnormal view-through install rates, install timing, suspicious device clusters, and IP-level activity before blocking suspicious IP ranges, rejecting unrealistically fast installs, filtering fraudulent devices, and removing the fraudulent DSP from the media plan.

As a result, fraudulent installs dropped to zero, the direct fraud rate reached 0%, and view-through install ratios normalized from 95% to a healthy 60% benchmark. The campaign resumed at full scale on verified Samsung and LG OEM Direct inventory, reinforcing the value of direct OEM supply partners for transparency, audit rights, and device-level visibility.

Ready to protect your CTV investment?