This Week’s Search Roundup: Core update turbulence, AI source labels and tighter Ads data limits

This Week in One Sentence

Across Search, Ads, and shopping surfaces, the direction is clear: platform visibility is becoming more dynamic during updates, more tightly coupled to enforcement and indexing signals, and more mediated by labelled AI and curated experiences. Measurement and control are shifting into platform-native reporting, feed attributes, and workflow tooling rather than relying on stable SERP patterns and long-term in-platform history.

This Week’s Exec Summary

• Ranking movement tied to the May 2026 core update remained turbulent, with reported “flip-flopping” behaviour and a rollout completion indicated around June 2.
• Google’s AI surfaces continued to formalise visibility signals, with a reported reduction in AI Overviews’ deindexing lag and expanded source labelling (“preferred sources”, “highly cited”).
• Merchant Center is being positioned as an AI shopping control surface via new conversational feed attributes and an AI Performance Insights report rolling out broadly.
• Google Ads is tightening operations at both ends: earlier policy feedback during ad creation, and a new 37-month ceiling for granular reporting data (starting June 1).
• Outside Google, ad tooling and image discovery UX both moved towards more explicit controls (budgeting/geo-targeting in beta; AI-curated image results with a user toggle).

Google Search: volatility and the May 2026 core update

Google Search May 2026 core update rollout and completion

Google launched a core update on May 21, 2026. Noticeable ranking volatility was observed from around May 23, with additional swings continuing into early June. Google indicated the rollout completed around June 2, while some observers reported the impact felt delayed after the announcement and included “flip-flopping” movements for certain sites.

• Google announced the May 2026 core update on May 21, 2026, and described it as a regular core update.
• The rollout was expected to take about two weeks (per Google’s stated guidance).
• Volatility was observed from around May 23, with additional swings through early June.
• Reported/observed: the rollout appeared to “start” in tools and analytics more strongly on the Saturday/Sunday after announcement, rather than immediately.
• Reported/observed: some sites saw major increases or drops, and some saw changes reversing (“flipping back and forth”).
• Memorial Day weekend was noted as a factor that can make analytics and tracking tools look inconsistent.

Why it matters: Treat rollout-period spikes as unstable inputs: pause reactive changes driven by single-day movement, then assess post-rollout trends across multiple time windows and page groups. Build your analysis to withstand short-term “flip” behaviour by tracking cohorts rather than isolated keywords.

Google AI surfaces: visibility alignment and source signalling

AI Overviews deindexing visibility lag appears to be removed

A third-party observer reported that there had been a lag where pages removed from standard search results (including via manual actions) could still appear in AI Overviews for some time. That lag now appears reduced or eliminated, meaning de-indexed or manually actioned content should no longer continue surfacing in AI Overviews. The same expectation was stated for AI Mode, and this was not described as a formal Google announcement.

• Reported/observed: AI Overviews previously could show content even when it no longer appeared in standard results.
• The lag was described as lasting for weeks in some cases.
• Reported/observed: the lag now seems to be gone or significantly narrowed.
• The change was also stated to apply to AI Mode.
• This was not described as a formal Google announcement; it is an observation.

Why it matters: Assume enforcement and indexing changes may now propagate faster into AI surfaces, and plan remediation workflows accordingly. Monitor whether removed or actioned pages still appear in AI citations and document any exceptions for follow-up.

Preferred source labels/carousels in AI Mode & AI Overviews; “highly cited” labels expanded

Google introduced preferred source labelling within AI Mode and AI Overviews, alongside two new carousels tied to preferred sources. Google also expanded “highly cited” labels beyond Top Stories to more web article links in the main results. These are presentation changes, but they shape which sources users are encouraged to trust and click.

• Preferred source labels now show in AI Mode and AI Overviews.
• If a publisher is a preferred source for a specific user, it is more likely to appear in AI citations for that user (as described).
• Two new preferred-source-related carousels were added, intended to promote social and perspectives content areas.
• The social and perspectives areas also display the preferred sources label.
• “Highly cited” labels, introduced for Top Stories in 2022, were expanded to additional web article links on the main results page.

Why it matters: Build reporting that captures label presence (not just ranking) and segment performance by AI surfaces versus standard results, because the UI now explicitly favours certain sources. Track whether labelled placements correlate with changes in referral quality as well as volume.

Shopping: product feeds for AI-driven results

Google Merchant Center introduces conversational attributes for AI-driven surfaces

Google Merchant Center documentation described new conversational attributes that merchants can add to product feeds. The intent is to help products match conversational queries and appear in AI-driven responses such as AI Mode and AI Overviews. This shifts part of optimisation from landing pages to structured product data.

• Merchant Center includes new conversational attributes for “AI-driven surfaces”.
• The documentation frames this as aligning product content with conversational searches.
• Conversational attributes can be added to feeds to help products appear in AI responses.
• AI Mode and AI Overviews were named as example surfaces for these AI responses.

Why it matters: Treat feed completeness as an AI visibility input, not a background task. Align any conversational attributes with the landing page and keep claims consistent, so eligibility and user expectations do not drift apart.

Google Merchant Center rolls out AI Performance Insights to all users

Google Merchant Center released an AI Performance Insights report intended to show performance on Google AI services and provide guidance for optimising product data for AI-powered experiences. Google indicated the report is rolling out broadly to all users, with multiple insight views intended to connect feed quality with AI-surface outcomes.

• Merchant Center introduced an AI Performance Insights report.
• It is positioned as a view into performance on Google AI services.
• It also aims to guide optimisation of product data for AI-powered experiences.
• Included insight areas listed: share of voice, shopping funnel performance, product term insights, and product attributes insights.
• Google indicated the report is rolling out to everybody.

Why it matters: Use this report as your baseline for AI-surface measurement, then tie feed edits to changes in impressions and downstream conversion quality. Operationally, it becomes the reference point for prioritising product data work alongside on-site optimisation.

Google Ads: workflow, policy, and measurement constraints

Real-time policy reviews during responsive search ad creation

Google Ads introduced a real-time policy review experience that runs while ads are being created. It is currently available for responsive search ads and is intended to help ads go live sooner, reducing post-submission disapprovals for straightforward issues.

• Google Ads now runs policy reviews during the ad creation process.
• The feature is currently available for responsive search ads.
• The stated aim is to help ads go live sooner.
• The review happens while the ad is being built, rather than only after submission.

Why it matters: Shift quality checks earlier: resolve policy prompts during build and update internal sign-off steps so launch timelines do not assume an approval delay that may no longer apply. This also reduces wasted iteration cycles when testing new creative.

Google Ads to delete hourly/daily/weekly performance data after 37 months

Google Ads is changing its data retention policy, deleting shorter-term performance reporting data (hourly, daily, weekly) within 37 months. Higher-level aggregations (monthly, quarterly, yearly) will be retained for up to 11 years. The change was described as starting June 1.

• Google Ads will begin deleting shorter-term performance reporting data starting June 1.
• Hourly, daily, and weekly reporting data will be deleted within 37 months.
• Monthly, quarterly, and yearly data will be retained for 11 years.
• The change affects reporting granularity availability over time, not necessarily billing records.

Why it matters: Treat 37 months as a hard ceiling for in-platform granular history and build or update your exports now if you depend on multi-year daily data for modelling, seasonality, pacing, or audits. Make sure dashboards and alerts do not silently degrade as older granularity disappears.

Search Console reporting: Discover data gap

Discover report logging error causes missing May 21 data with no fix

Google Search Console reported a logging error that caused decreased clicks and impressions in the Discover performance report for May 21, 2026. Google stated it is a reporting issue rather than a ranking issue, and that the missing data will not be fixed or backfilled. It was described as separate from earlier Discover reporting issues on May 7 and May 8.

• A logging error impacted Discover clicks and impressions reporting for May 21, 2026.
• Google said the issue is limited to reporting, not rankings.
• Google stated the missing data will not be fixed or backfilled.
• The May 21 issue was described as separate from Discover report bugs on May 7 and May 8.

Why it matters: Annotate May 21 in any Discover reporting and adjust automated comparisons so they do not misread the drop as performance. Where possible, corroborate trends using on-site analytics and owned-channel metrics to avoid making decisions on a known reporting gap.

Platform shifts beyond Google

OpenAI Ads Manager beta adds daily/lifetime budgets and granular US geo-targeting

OpenAI expanded its Ads Manager beta with daily and lifetime budget options for new campaigns and introduced more granular US location targeting. Targeting can be set by state, designated market area (DMA), and zip code, and can be adjusted after setup. Daily budgets were described as limited to newly created campaigns.

• OpenAI expanded Ads Manager beta features for budgeting and geo-targeting.
• New campaigns can choose either daily or lifetime budgets.
• Daily budgets are currently limited to newly created campaigns.
• US targeting can be set by state, DMA, and zip code.
• Location targeting can be configured during setup or adjusted later in campaign settings.

Why it matters: Treat the added controls as the minimum needed for clean experiments: align geo settings with your existing reporting regions and keep budget types consistent across test and control campaigns. That discipline makes cross-channel comparisons more reliable.

Bing officially releases AI-curated image results with on/off toggle

Microsoft officially rolled out AI-curated image results in Bing after earlier testing. The experience groups images into categories and includes a toggle so users can turn the AI-curated view on or off. This is a product UX change that can alter how users browse and discover image content.

• Microsoft announced the official release of AI-curated image results in Bing.
• The feature was previously spotted as a test (reported as March testing).
• Users can toggle the AI-curated experience on or off.
• Image results are organised into grouped categories.
• The feature is positioned as reducing overwhelm while browsing images.

Why it matters: Review image discovery performance with the expectation that category-grouped browsing can concentrate attention into fewer paths. Track shifts in traffic patterns over time and check how your imagery and metadata align with common category intent.

Policy and legal: reported enforcement pressure and ongoing appeals

EU reportedly considering a fine over alleged self-preferencing in Google Search

A report claims the European Union may impose a fine on Google over allegations that Google favours its own services in search results despite DMA-related changes. The potential fine was described as high triple-digit million euros, with one mention putting it as potentially up to 999 million euros. Timing was suggested as possibly before the summer, but this was not confirmed.

• Reported: the EU is considering action under the Digital Markets Act (DMA) related to alleged self-preferencing.
• Reported: the concern is favouring Google’s own services in results (examples mentioned include Maps and flights).
• Reported: a possible fine was described as high triple-digit million euros, potentially up to 999 million euros.
• Suggested (unconfirmed): a decision could come out before the summer.
• Prior EU fines against Google were referenced in discussion (including 2017’s 2.7 billion euros and a more recent 3.5 billion euros), but those are separate matters.

Why it matters: Treat this as a live risk factor for European SERP layout stability: be ready to re-baseline click distribution on queries where modules blend organic results with platform-owned destinations. Keep regional reporting segmented so any change does not get diluted in global averages.

Google appeals the ruling that found it to be an illegal search monopoly

Google filed an appeal challenging a prior ruling that found it to be an illegal search monopoly. The appeal was described as a 111-page filing, and the discussion referenced remedies issued after the ruling, said to have come out in September 2025. Some details (such as the exact ruling date and jurisdiction) were referenced in discussion but not fully specified in the notes.

• Judge Mehta ruled that Google was an illegal search monopoly (as described).
• Remedies were issued after the ruling; they were said to have come out in September 2025.
• Google submitted an appeal filing on Friday (relative to the reporting week).
• The appeal was described as a 111-page document.
• Some timing and case-detail references were discussed but not fully substantiated in the provided notes.

Why it matters: Plan for extended uncertainty: appeals can lengthen the window in which remedy-driven distribution or default-placement changes remain unresolved. Keep acquisition forecasts flexible where “default search” placements materially affect channel mix.

What to watch next week

• Whether May 2026 core update volatility truly settles post-rollout, and which site types show sustained winners/losers after the “flip” behaviour.
• Whether the reported AI Overviews deindexing/manual-action alignment holds consistently across more examples (and whether any lag returns).
• Merchant Center adoption: early signals from AI Performance Insights and whether conversational feed attributes correlate with measurable lift.
• Operational readiness for Ads data retention: confirm export/storage plans before June 1 if you rely on granular history.