Google Discover France: manual spam action removes nine domains

Abondance reported on 6 November that nine domains were removed from Google Discover France in three separate waves, between 21:30 on 3 November and 11:00 on 4 November. According to the publication, the domains together represented around 2% of the articles served by Discover in France. Affected publishers received a Search Console message titled with a term the report does not quote in full on 5 November.

What the report establishes, and what it leaves open

The sequence matters more than the number. Three waves within roughly fourteen hours point to a manual enforcement decision applied in batches rather than an algorithmic recalculation. The Search Console notification, arriving a day after the removals, is consistent with the standard pattern for manual actions: the visibility change comes first, the explanation follows.

The report does not name the domains, does not describe the nature of the spam policy breach, and does not say whether the action is site-wide or limited to Discover surfaces. It also does not indicate whether an appeal or a reconsideration request has been filed, or what share of each publisher's total traffic the lost Discover impressions represented. Those gaps are the ones that matter most for anyone trying to read the event.

Why Discover exposure is a fragile channel

Discover is a recommendation surface, not a search results page. Traffic arrives without a query, often on mobile, often in a feed context, and it is notoriously volatile: a single story can produce a spike that disappears within days. That makes it attractive for reach and awkward for planning. Teams that build revenue models on Discover volume are effectively renting distribution from a system that can revoke it without a grace period.

General industry context is useful here. Google's spam policies apply across its surfaces, and manual actions are human-reviewed interventions rather than automated adjustments. Publishers who depend on recommended feeds typically discover their exposure only when it is removed.

What a team can check this week

• Pull the last 90 days of Discover performance in Search Console and calculate what share of total sessions and revenue depends on it.
• Verify that Search Console notifications reach a monitored inbox, and that someone owns the response if a manual action appears.
• Review content produced specifically for feed distribution against Google's spam policies, particularly around misleading previews and scaled content.

The broader lesson is about concentration. A channel that can drop roughly 2% of a national feed's inventory overnight is a channel worth measuring separately from search, and worth capping as a share of revenue. The report leaves the cause unspecified; the operational question for publishers does not depend on knowing it.