Google to stop scanning Gmail email content to target ads, will use other user info it has for ad targeting
Mark Bergen / Bloomberg :
Context & Ripple Effects
Gmail content scanning has been Google's longest-running ad-privacy flashpoint: a December 2016 privacy suit settlement already confined scanning to mail after it arrived in the inbox, a concession framed at the time as having little practical effect. Ending content-based ad targeting outright goes further, and the substitution is the real story — Google is not giving up the targeting, it is moving it off message text and onto the other user information it already holds across its services.
That substitution follows a path Google laid out back in 2015, when it pushed advertisers to supply customer details like email addresses for search ad targeting, mirroring Facebook's Custom Audiences. The direction of travel since has been consistent: away from reading what users write, toward profile-level and advertiser-supplied signals.
First-order effects
- Advertisers who bought against Gmail message content lose that inventory signal immediately, and must re-buy against Google's broader logged-in user profile instead.
- Google gains a cleaner privacy story for Gmail — the consumer-facing product most associated with the scanning practice — without shrinking its overall ad-targeting capability.
Second-order effects
- The shift strengthens the case for advertiser-supplied first-party data matches of the kind Google sought in 2015, pushing more of the targeting burden and cost onto brands rather than Google's own crawling.
- Rivals can no longer use Gmail scanning as a competitive contrast, which raises the bar for privacy positioning across free webmail generally.
Third-order effects
- This fits the sequence visible in the coverage — the 2016 settlement, the 2018 Ad Settings overhaul, the 2019 halt on sharing page-content signals with auction buyers after EU scrutiny, and the 2020 ban on demographic and zip-code targeting for housing, employment, and credit ads: ad targeting steadily migrating from raw content inspection toward consented, profile-based, and category-restricted signals.
- If that pattern holds, the durable advantage accrues to platforms with large cross-service login footprints, since they alone can replace content signals with their own user data — a structural consolidation pressure on smaller ad-supported services.
The trend: Webmail-era content scanning is being retired piece by piece under legal and regulatory pressure, with ad targeting rebuilt on platform-held user profiles and advertiser-supplied data instead.