May 2026 core update

What Google said

Google used its standard core update wording: a regular update designed to better surface relevant, satisfying content for searchers from all types of sites, with a rollout that might take up to two weeks. It applied globally, in all regions and all languages. No new guidance accompanied it, which is normal. Google also repeated its standing position that a negative ranking movement in a core update does not necessarily mean anything is wrong with your pages.

What the data showed

Three patterns came through consistently in community analysis and in the sites I reviewed.

• Topical authority carried more weight. Why: Sites with fewer, deeply researched articles on a narrow subject generally outperformed high-volume operations trying to cover every keyword variant.
• First-hand experience became decisive. Why: Pages containing real examples, original testing or case data gained against generic rewrites and AI-assisted summaries of what already ranked.
• AI Overview citation share diverged from organic position. Why: Being cited in an AI answer and ranking well organically stopped moving together, which means they now need to be tracked as two separate metrics.

The rollout was also unusually front-loaded. Impact was visible by May 23, only two days in, where the March update took considerably longer to land. That matters for diagnosis, because a drop on May 23 is inside this rollout even though it feels too early to be.

One data caveat for this rollout

A confirmed Search Console reporting bug began on May 21, the same day the update launched, and caused the Links report to show zero or dramatically reduced link counts for many sites. If you were reading link data during this window, treat it as unreliable. It was a reporting fault, not a loss of links, and it had nothing to do with the core update.

How to measure your own impact

Google advises waiting a full week after a rollout completes before drawing conclusions, because mid-rollout rankings are unstable. For this update that means the earliest reliable analysis window opened around June 9.

Step	What to compare	Measurement (what it tells you)
Set the baseline	May 14 to 20, 2026	Your pre-update performance, before any rollout noise
Set the after window	June 3 to 11, 2026 or later	Post-rollout performance once rankings stabilised
Check scope	Click loss grouped by page template	Site-wide points at core, one group points elsewhere
Check displacement	Clicks against impressions and position	Flat impressions with falling clicks is not a core drop
Check Discover separately	Search type set to Discover	Discover losses may trace to the February 2026 Discover update

What to do if you dropped

A core update drop is a reassessment, not a penalty. Nothing was violated and there is no reconsideration process. The work is raising the quality bar on the pages that lost ground, and the reassessment comes at the next core update, which is why how long recovery takes matters for planning.

• Ask whether a competitor page now answers the query better. Why: That reframe points at the actual fix. Asking what you broke usually does not, because often nothing broke.
• Consolidate thin near-duplicates rather than rewriting everything. Why: Depth beat breadth in this update. The content pruning strategy covers how to decide what to merge.
• Add first-hand evidence to the pages that lost most. Why: Original testing, real examples and named authorship were the clearest differentiators between winners and losers.
• Track AI citation share separately from position. Why: If your rank held and clicks fell, the AI Overview click recovery method applies instead of quality work.

The core update recovery service covers the full process, and how the ranking systems fit together explains why core updates behave differently from spam and product updates.