Google May 2026 core update: what to measure now

Google began rolling out the May 2026 core update on 21 May, with a two-week completion window. It is the second core update of 2026 and arrived 48 hours after Google unveiled an AI-first Search redesign at I/O. The two events together move both organic rankings and AI Overview citations at once. Most marketing teams watch Search Console alone. That covers one of those surfaces, not both. Here’s what is genuinely happening and what to do about it.

The news most recaps are missing: two algorithm shifts in 48 hours

The May 2026 core update is the second core update of the year, following the March 2026 rollout that ran from 27 March to 8 April. That much is in every recap. What most of them aren’t connecting: this core update came 48 hours after Google used I/O 2026 to unveil the biggest Search redesign it has shipped in over a decade.

On 19 and 20 May, Google announced three Search changes:

• An intelligent search box powered by Gemini 3.5 Flash
• Agent-driven experiences that prioritise AI-generated interfaces over traditional blue links
• Information agents that track web changes continuously in the background

AI Overviews now serve 2.5 billion monthly users, with AI Mode reaching 1 billion. The new search box rolls out this week, with generative UI arriving later this summer.

Two days after that announcement, the core update started rolling out. Two foundational changes, 48 hours apart.

That timing puts the May 2026 core update in direct tension with what Google announced at I/O. Marketing teams trying to read ranking shifts this week aren’t seeing a clean test of a single algorithm change. They’re seeing the combined effect of a core update affecting traditional rankings and a Search redesign affecting how those rankings are presented, surfaced and synthesised into AI-generated answers.

Why a core update now moves two algorithms, not one

Google’s position is that AI Overviews and AI Mode are powered by the same underlying ranking systems as organic Search. Foundationally, that’s true. In practice, the outputs of those two surfaces have diverged sharply.

Late 2024 research showed around 75% of AI Overview citations came from pages ranking in the organic top 12. BrightEdge research from October 2025 put the overlap at 54%. Two more recent 2026 studies, including independent analysis from Ahrefs and ALM Corp, found that only 17% to 38% of AI Overview citations now come from top-10 organic results.

The gap has widened because Google uses a fan-out query process for AI-generated answers. Each user query is broken into multiple sub-queries, and the system cites pages that perform well across that wider cluster rather than just the original keyword. A page ranking sixth for one query might be cited because it ranks first for a related sub-query the user didn’t explicitly type.

Selection criteria differ too. Citation rewards structured, well-sourced content from credentialed domains with strong schema implementation. Conventional SEO levers like page recency, load speed and reading-grade level don’t predict citation. That’s the territory of Generative Engine Optimisation (GEO) and Answer Engine Optimisation (AEO), which operate on different evidence patterns than traditional ranking.

One site finishes the May 2026 rollout holding its #1 organic position but missing its AI Overview citation. Another finishes with two organic positions lost and a citation gained that it didn’t have before. Same algorithm change. Different outcomes. Those two surfaces now overlap in their outputs by only around a third, so the same shift in underlying signals produces different shifts on each surface.

There’s a third surface that doesn’t get mentioned in core update coverage at all. ChatGPT, Perplexity and Copilot use independent training data and retrieval systems. A Google core update tells you nothing about visibility in those tools, even though they’re now part of how qualified prospects research vendors.

What Search Console can and cannot tell you this time

Search Console is excellent at what it was designed for: measuring organic search appearance, clicks, impressions, average position and CTR for the standard blue-link results that Google still serves. That covers maybe half of how prospective customers now reach a B2B or public sector website.

What Search Console doesn’t cleanly cover:

• AI Overview citation impressions and clicks (Google added AI search filtering during 2025 but coverage is incomplete and inconsistent)
• Impressions or citations within AI Mode
• Any visibility data from ChatGPT, Perplexity, Copilot or other LLM-powered search tools

For most of 2024 and 2025, the gap mattered less because organic and AI Overview citation tracked closely. A site rising in organic also rose in citation. Sites dropping in organic dropped in both. The data from 2026 shows that pattern has broken.

After the May 2026 core update, some sites will see Search Console show organic traffic holding steady while AI Overview citation share drops by half. Others will see organic traffic fall while AI Overview citation goes up. Both are real outcomes. Neither is visible if you only watch one dashboard.

The discipline of the next two weeks

Google’s recommendation for every core update is the same: wait at least one full week after the rollout completes before drawing conclusions from Search Console. For the May 2026 update, that means analysis from around 11 June at the earliest. Comparisons should measure the two or three weeks before 21 May against the week after the rollout completes, not mid-rollout data, which is noise dressed up as signal.

Beyond that base discipline, a few things matter more this time.

Treat your site as a whole, not page by page. Core updates re-evaluate the relative quality of entire sites, not individual URLs. Page-level positions move because the site-level evaluation has shifted. Trying to fix a single page that dropped is rarely the right diagnosis.

Separate Web Search, Images, Video and News performance in Search Console. Each tab has its own dynamic during a core update. Mixing the data together obscures what’s actually happening.

Don’t make structural changes during the rollout. Deletion is a last resort. Sites that delete sections in panic during week one regularly find those sections were valuable signals once the dust settles. Restructuring content around perceived signal changes during an unstable rollout adds more variables to data that is already hard to interpret.

The same discipline applies to AI surfaces, with one extra wrinkle. AI Overview citation has higher week-to-week variance than organic position at baseline. During a core update rollout, citation patterns can fluctuate enough that any conclusion drawn from a single week of data is too thin to act on. Treat AI surface data with the same patience as organic data.

A measurement framework for the AI-era

The May 2026 core update is the right moment to extend the measurement stack rather than run another tactical recovery audit. Three sources of data, watched alongside each other, give a fuller picture of how the update has affected real visibility.

Search Console remains the baseline for organic position, impressions and clicks. Pull a 90-day view that brackets the rollout cleanly, then sub-divide by tab (Web, Image, Video, News) and by content cluster rather than by individual page.

AI Overview citation tracking sits on top of the organic baseline. Google’s AI search filter inside Search Console gives a partial picture. Third-party tools that track citation appearance per query for your site against your competitors fill the gap. Track citation share for your priority query cluster on a week-on-week basis rather than aggregate citation counts.

AI Share of Voice is the equivalent metric for LLM-powered search tools. It measures how often a brand is referenced inside ChatGPT, Perplexity, Copilot and Gemini answers for a defined set of priority queries, compared with named competitors. For a Google core update, AI Share of Voice changes will lag organic shifts by days or weeks because each LLM refreshes its indexed data on its own schedule. That lag is informative: it tells you which surface is moving and when.

Watched together, these three views answer questions that Search Console alone can’t. Did organic move because of the core update, the AI redesign or both? Is AI Overview citation moving with organic or against it? Has LLM visibility shifted yet? If not, when?

In two weeks the May 2026 core update will be done. The measurement gap it has exposed between organic ranking and AI search visibility will not. Marketing teams running organic SEO and AI search together will move through this rollout with a clearer picture than teams running either in isolation.

Priority Pixels works with mid-market and public sector teams that want both surfaces measured properly. If you want a structured view of how the May 2026 core update has affected your organic and AI search visibility together, get in touch.