When Google referral traffic drops 25% in a single month, the most common response is to blame the most recent Google update. But in 2025, at least three distinct mechanisms can produce identical traffic decline symptoms: AI Overview expansion absorbing clicks, a core algorithm update changing rankings, or an indexation issue removing pages from the index. Each requires a fundamentally different remediation strategy, and misdiagnosis wastes weeks of effort on the wrong fix. Accurate differential diagnosis requires testing each hypothesis independently.
Step one: check indexation status to rule out the simplest and most fixable cause
Before investigating AI Overviews or algorithm changes, verify that pages remain indexed. Indexation loss is the most straightforward cause and the most directly fixable, making it the correct first diagnostic step.
Search Console’s Pages report (formerly Coverage report) shows indexed versus excluded pages with specific exclusion reasons. The URL Inspection Tool confirms individual page indexation status and identifies any issues preventing indexation. Site: searches in Google provide a quick spot-check of whether key pages appear in the index. The specific patterns that indicate indexation-driven decline include a sudden drop in indexed page count correlating with the traffic decline timeline, new “Excluded” entries in the Pages report coinciding with the decline start date, and robots.txt or noindex directives accidentally applied during a recent deployment.
Quantifying the traffic impact of indexation issues requires cross-referencing the deindexed URLs against their historical traffic in Google Analytics or Search Console. If the deindexed pages account for traffic volume approximately equal to the total decline, indexation is the primary cause. If deindexed pages account for only a fraction of the decline, indexation is a contributing factor alongside other causes. With increased use of JavaScript frameworks, rendering issues that prevent Googlebot from accessing content can cause indexation problems that mimic the symptoms of algorithm changes. Server-side rendering verification should be part of the indexation audit.
Step two: segment traffic by query type to isolate AI Overview-affected queries from algorithm-affected queries
AI Overview expansion affects informational queries disproportionately, while algorithm updates can affect queries across all intent categories. Segmenting traffic decline by query intent category reveals whether the drop concentrates in informational queries (suggesting AI Overview impact) or spans all categories (suggesting algorithm change).
The segmentation methodology uses Search Console’s Performance report filtered by query type. Group queries into informational (what, how, why, guide, tutorial), navigational (brand, specific site), transactional (buy, price, deal, review), and commercial investigation (best, compare, versus) categories. Calculate the click decline percentage within each category separately. If informational queries show 30-50% click decline while other categories remain stable, AI Overview expansion is the likely primary cause. If decline is distributed proportionally across all categories, an algorithm update is more probable.
The intent classification approach can be manual for small query sets (under 500 queries) or automated using keyword intent classifiers available in Semrush, Ahrefs, or custom classification scripts. The interpretation framework for segment-specific patterns includes: informational-only decline with stable impressions indicates AI Overview CTR suppression; decline across all segments with reduced impressions indicates algorithm-driven visibility loss; and decline concentrated in specific topic clusters regardless of intent indicates topic-specific algorithm impact.
Step three: compare ranking positions before and after the decline to distinguish visibility loss from click loss
If rankings remain stable but clicks decline, the cause is likely CTR suppression from AI Overviews or SERP feature changes. If rankings dropped, the cause is likely an algorithm update. This distinction is critical because the remediation strategies are entirely different.
The ranking-versus-click analysis uses Search Console’s Position metric alongside Click and Impression data for the same query set. Calculate the average position change and the click change independently. Queries where position remained within 0.5 positions of the pre-decline average but clicks dropped 20% or more are candidates for AI Overview CTR suppression. Ahrefs’ December 2025 data confirmed that AI Overviews reduce position-one organic CTR by 58%, meaning a page can maintain its ranking while losing more than half its click volume.
Query-level analysis is necessary for accurate attribution because the decline may be a mix of both causes. Export the full query performance data for the decline period and the equivalent pre-decline period. For each query, calculate position change and click change independently. Queries with stable position and declining clicks go into the AI Overview bucket. Queries with declining position and proportionally declining clicks go into the algorithm bucket. The resulting proportional attribution shows what percentage of the total decline each cause likely explains.
Step four: cross-reference timing with known AI Overview rollout dates and confirmed algorithm updates
Correlating the traffic decline timeline with documented events provides circumstantial evidence for attribution. Timeline correlation is not definitive proof, but it strengthens or weakens each hypothesis.
The data sources for timeline correlation include the Google Search Status Dashboard (search.google.com/status) for confirmed algorithm updates with official start and end dates, Semrush Sensor and other SERP volatility tools for detecting unannounced algorithm changes, and AI Overview prevalence tracking data from Semrush’s studies showing AIO prevalence fluctuation across 2025. The Google December 2025 Core Update, for example, directly evaluated AI content authenticity and expanded E-E-A-T requirements beyond YMYL topics, creating ranking changes that coincided with continued AI Overview expansion.
Unannounced updates also occur. The November 2025 period showed real ranking volatility despite no official core update announcement from Google. Cross-referencing with third-party SERP volatility sensors provides coverage for these undocumented changes. The confidence levels achievable through temporal analysis alone are moderate: strong correlation with a known event increases attribution confidence to 70-80%, while no temporal correlation reduces confidence below 50%.
The differential diagnosis limitation: multiple causes frequently overlap, making clean attribution impossible
In practice, AI Overview expansion and algorithm updates often occur during overlapping timeframes, making it impossible to attribute traffic decline cleanly to a single cause. This attribution overlap is the norm rather than the exception in 2025.
Google’s AI Overview prevalence expanded from 6.49% in January 2025 to approximately 25% in July, while multiple core updates rolled out during the same period. Any traffic decline during these months could reflect either or both causes. Digiday’s analysis found that AI Overviews were linked to a 25% drop in publisher referral traffic, but this occurred alongside algorithm changes that independently affected publisher visibility.
The analytical approaches for estimating proportional attribution when causes overlap include the segment isolation method (using query-type segmentation to separate AI-affected and algorithm-affected queries), the position-CTR decomposition method (attributing position-stable CTR declines to AI Overviews and position declines to algorithms), and the control group method (identifying a query set with no AI Overview presence as a control to measure algorithm-only impact). The strategic implication is that remediation strategies should address both causes simultaneously rather than waiting for definitive attribution that may never arrive.
How does query intent segmentation reveal whether AI Overviews or algorithm changes caused a traffic drop?
Segment Search Console data by query intent: informational, navigational, transactional, and commercial investigation. If informational queries show 30-50% click decline while other categories remain stable, AI Overview expansion is the likely cause. If decline distributes proportionally across all categories, an algorithm update is more probable. Informational-only decline with stable impressions specifically indicates AI Overview CTR suppression rather than ranking loss.
What is the position-CTR decomposition method for attributing traffic decline?
Compare ranking positions and click volumes independently for each query. Queries where position remained within 0.5 positions of the pre-decline average but clicks dropped 20% or more are candidates for AI Overview CTR suppression. Queries with declining position and proportionally declining clicks indicate algorithm-driven loss. Ahrefs confirmed AI Overviews reduce position-one CTR by 58%, meaning a page can maintain its ranking while losing more than half its click volume.
Why is clean single-cause attribution for traffic declines typically impossible in 2025?
AI Overview expansion and algorithm updates often occur during overlapping timeframes. Google’s AI Overview prevalence expanded from 6.49% in January 2025 to approximately 25% in July, while multiple core updates rolled out simultaneously. The practical response is addressing both causes in parallel using the segment isolation method, position-CTR decomposition, and control group analysis with queries that have no AI Overview presence, rather than waiting for definitive attribution that may never arrive.
Sources
- https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/
- https://www.semrush.com/blog/semrush-ai-overviews-study/
- https://searchengineland.com/guide/google-core-updates
- https://digiday.com/media/google-ai-overviews-linked-to-25-drop-in-publisher-referral-traffic-new-data-shows/
- https://almcorp.com/blog/google-algorithm-december-2025-core-update-guide/
