Every confirmed Google ranking update since December 2025

Most arguments about a traffic drop are arguments about dates. This is the record, as Google publishes it, with nothing added: eight confirmed ranking updates between December 2025 and September 2026.

The confirmed record#

Update	Type	Started	Google’s stated duration
August 2026 spam update	Spam	18 Aug 2026	2 days, 16 hours
June 2026 spam update	Spam	24 Jun 2026	2 days, 1 hour
May 2026 core update	Core	21 May 2026	11 days, 21 hours
March 2026 core update	Core	27 Mar 2026	12 days, 4 hours
March 2026 spam update	Spam	24 Mar 2026	19 hours, 30 minutes
February 2026 Discover update	Discover	5 Feb 2026	21 days, 17 hours
December 2025 core update	Core	11 Dec 2025	18 days, 2 hours
August 2025 spam update	Spam	26 Aug 2025	26 days, 15 hours

The same table is published as JSON with the durations expressed in hours, so you can join it against your own analytics export instead of retyping it.

What the durations tell you#

Core updates cluster. The three in this record took 11 days 21 hours, 12 days 4 hours and 18 days 2 hours, an average of about 14 days. When one starts, assume a fortnight of movement rather than guessing.

Spam updates do not cluster at all. March 2026 finished in under 20 hours; August 2025 took more than 26 days. The 8-day average describes nothing you would plan around - spam updates vary with what is being enforced, not with a fixed process.

The longest rollout is the most misread one. February 2026’s Discover update ran 21 days 17 hours and touched Discover only. A site whose Discover traffic moved that month while web rankings held steady was not experiencing a partial core update.

The gap nobody has explained#

No core update has been confirmed since 21 May 2026. As of 10 September that is the longest interval between confirmed core updates in this record, filled with the usual industry conversation about unconfirmed volatility.

Google confirms what it confirms. Volatility trackers measure something real, but a proxy, and attaching “unconfirmed update” to a week of movement is a hypothesis, not a finding. If your traffic changed outside every window in the table above, the first suspect is your site or your market, not an update nobody can name.

Why we care#

Export daily clicks and impressions from Search Console for a window starting at least a week before an update began and ending a week after it completed, then check which surface moved. A shift confined to Discover, to news carousels or to a single country is evidence about that surface, and the table above tells you which updates were even capable of producing it - the difference between diagnosing core update damage and assuming it.