Organic traffic dropped: is it AI Overviews or something else?
Organic traffic dropped? Use Search Console to tell AI Overviews from a ranking change, a technical fault or a tracking error, in five checks.
By Crawlito · Updated 10 min read
In short
- Search Console has a Generative AI performance report for AI Overviews and AI Mode, but Google's help page describes impressions only, so you can see where you appear but not a click count for those features.
- Rule out tracking and reporting faults first: Google says Search Console and Analytics count different things, and Search Console lists known logging errors on its Data anomalies page.
- Google's debugging guide says falling impressions and clicks together point to common causes such as technical faults, while steady impressions with falling clicks point to how your result is presented.
- Steady impressions and position with falling clicks is consistent with an AI Overview taking attention, but it is a hypothesis you test by searching the affected queries, not a diagnosis.
- Google advises against radical changes to pages that already perform well, and says that making changes does not guarantee a visible effect in search results.
On this page
- How to tell an AI Overviews effect from a ranking or technical drop
- Step 1: rule out data anomalies, tracking and consent problems
- Step 2: read impressions, clicks, CTR and position together
- Step 3: compare periods and segment by query, page, device and search type
- Step 4: check indexing, crawl errors, manual actions and known updates
- Step 5: check whether the lost queries are ones an AI answer can satisfy
- What to change and what not to change after a drop
- Frequently asked questions
- Check the technical side of your pages
If your organic traffic dropped, you cannot tell from one number whether AI Overviews are the cause. Search Console's Generative AI performance report shows impressions in AI Overviews and AI Mode, but Google's help page for it describes impressions only, with no clicks, so it cannot say how many clicks you lost to them. You work by elimination: rule out tracking and reporting faults, read impressions, clicks and position together, compare periods, check for technical faults and known updates, and only then ask whether the queries you lost are ones an AI answer can satisfy. This guide gives the steps in that order.
One caveat first. This process narrows the cause down. It does not prove one, and nothing here promises that a fix brings traffic back.
How to tell an AI Overviews effect from a ranking or technical drop
Each cause leaves a different pattern in Search Console's Performance report. Google's guide to debugging search traffic drops describes several of them, and the table below adds the AI Overviews case as a hypothesis, not a documented signature.
| What you see | A likely cause | Check next |
|---|---|---|
| Search Console is flat, but your analytics tool fell | Tracking, tag or consent problem | Step 1 |
| Impressions and clicks both fell sharply, site-wide | Technical fault, security issue or manual action | Step 4 |
| Impressions and clicks fell slowly on a group of pages | Something page-level, such as a misplaced noindex | Step 4 |
| Position and impressions fell together, around an update date | Ranking change or algorithm update | Step 4 |
| Impressions held, clicks fell | Snippet, rich result or results-page change, which could include an AI Overview | Steps 3 and 5 |
| The same dip appears every year | Seasonality | Step 3 |
Google's guide says that if impressions and clicks both fell, you should investigate the common causes, and that if impressions held while clicks dropped, your title and snippet may be less effective than they could be. It does not mention AI Overviews in that guide. The last row is where they come in, and why Step 5 exists.
Step 1: rule out data anomalies, tracking and consent problems
Before you investigate your site, check that the drop is real.
- Open the Data anomalies page in Search Console. Google lists known issues there that can explain dips or bumps in charts, mostly logging errors. At the time of writing, entries include logging errors that under-reported impressions, one of them for generative AI in Search. Check whether your dates overlap.
- Compare Search Console with your analytics tool, but not number for number. Google's guide to combining Search Console and Analytics says clicks and sessions are calculated differently, so totals will not match. It lists reasons for big gaps: pages missing the analytics tag, users who opt out of tracking, time zones (Search Console uses Pacific Time by default and you cannot change it), and canonical URLs.
- Match the filters. Apply the same country and device filters in both tools.
If Search Console is steady and only analytics fell, suspect a removed or broken tag, a consent banner change or a redirect that stripped campaign data. Google treats Search Console as the source for search performance and Analytics as the source for on-site behavior, so a fall in only one of them tells you which tool to look at.
Step 2: read impressions, clicks, CTR and position together
Open the Performance report and turn on all four metrics. Google defines them in its Performance report help. Two details matter for this diagnosis.
- Position is an average of each query's topmost position. Google says position values are averages and can mislead, so a small change is not evidence by itself. Google's debugging guide says impressions and clicks are the ultimate measure of success and cautions against over-focusing on position.
- AI Overviews occupy a single position. Per the same help page, an AI Overview takes one position in the results, all links in it share that position, and clicking an external link in it counts as a click. AI Mode links follow the same position method as a normal results page, and a follow-up question in AI Mode counts as a new query.
That second point is why position alone can hide an AI feature. Your link can keep a stable reported position while the page around it changes. Read the four metrics as a set:
- Impressions down, clicks down, position down: your visibility in the results fell. Look at ranking, indexing and updates.
- Impressions steady, position steady, clicks down: people see your result and click less. Look at your title, snippet and what else sits on the results page.
- Impressions down, position steady: fewer searches, or fewer results pages showing you. Check seasonality and demand.
Step 3: compare periods and segment by query, page, device and search type
Google's guide gives a concrete method. Set the date range to the last 16 months to see whether the dip repeats yearly. Then use the Compare tab, either the last three months against the previous period or year over year, and click through every tab: queries, pages, countries, devices and search appearances.
Segment in this order:
- Search type. Google says AI Overviews and AI Mode are reported in the Performance report under the Web search type, so a drop confined to Image or Video search points elsewhere. Google also added a Generative AI performance report in June 2026 and made it available to all sites on August 31, 2026. Its help page describes impressions by page, country, device and date for AI Overviews and AI Mode, and does not mention clicks, queries or position. It may not appear for a site with too few generative AI impressions. Check whether your impressions there fell or held, and treat that as one more input, not a verdict.
- Pages. Sort by clicks difference. A handful of pages carrying the loss points to something specific. Every page falling together points to something systemic.
- Queries. List the queries that lost the most clicks. You need this list for Step 5.
- Device and country. A drop in one country or on mobile only suggests a localized change, a rendering issue or a different results page there.
Google also recommends comparing with Google Trends to see whether interest in your topic fell, which separates a demand change from a visibility change. For history beyond 16 months, use the Search Analytics API or bulk data export.
Step 4: check indexing, crawl errors, manual actions and known updates
Google's guide separates technical issues, security issues, manual actions and algorithm updates, and says where to look for each.
- Technical faults. Google says site-wide problems show up as sharp drops, while page-level problems such as a misplaced noindex tag cause a slower decline. Check the Page indexing report and the Crawl stats report for spikes, then use URL Inspection on a sample of affected pages. Leftover staging settings are a common source, covered in staging noindex leaks.
- Site moves. Changing URLs causes fluctuation while Google recrawls. Google says a medium-sized site may take a few weeks to be noticed, and larger sites longer. See website migrations if you moved recently.
- Security issues and manual actions. Google's guide also points to the Security Issues and Manual Actions reports, so check both.
- Algorithm updates. Google's Search Status Dashboard lists incidents for crawling, indexing, ranking and serving, including core updates. Check whether your drop starts on an update date. Google's guide says a small move, such as position 2 to 4, can recover on its own, while a large one, such as position 4 to 29, calls for a self-assessment of the whole site for helpfulness and reliability. It also says changes can take a few days to several months to settle, so wait a few weeks before analyzing again.
If this step finds a cause, you probably do not need Step 5. AI Overviews are one possible explanation, not the default one.
Step 5: check whether the lost queries are ones an AI answer can satisfy
If Steps 1 to 4 found nothing, test the AI Overviews hypothesis directly. The Generative AI performance report shows impressions but, per its help page, not clicks, so you still check the results pages yourself.
- Take your top 10 to 20 queries by lost clicks from Step 3.
- Search each one in a logged-out or private window, on the device type where the drop happened. Note whether an AI Overview appears and whether your page is linked in it.
- Record the result in a simple table: query, AI Overview present, your page linked, and what the rest of the results page looks like.
Interpret it carefully:
- Many lost queries show an AI Overview that does not link you, and impressions held: consistent with fewer clicks because the answer is on the results page. It is still a hypothesis, because other results-page changes can have the same effect.
- Few lost queries show an AI Overview: AI Overviews are unlikely to explain most of the loss. Google says they are shown only when its systems determine they are additive to classic Search, and that they often do not trigger.
- Results differ from search to search: Google notes that personal and location factors can make a reported position differ from what an individual sees, so treat your checks as a sample, not a measurement.
As a working heuristic, not a Google rule, queries a short direct answer can satisfy, such as definitions, simple how-tos and quick facts, are more exposed than queries where the visitor needs your product, your inventory or your specific judgment. Google says clicks from results pages with AI Overviews tend to be higher quality, with visitors more likely to spend more time on the site. That is Google's statement about quality, not about volume, so judge it against your own conversions.
What to change and what not to change after a drop
Google's guidance is cautious, and it is worth following. It advises against radical changes to pages that already perform well, and it says that making site changes does not guarantee a visible effect in search results.
What Google says about AI search in Succeeding in AI search (May 2025) is the same advice as for classic search: create unique, useful content, provide a good page experience, make sure Googlebot is not blocked and pages return HTTP 200 with indexable content, and match structured data to visible content. Google's AI features documentation also says no special files, markup or optimizations are needed for AI Overviews or AI Mode. The requirements guide lists the three conditions to verify.
In practice:
- Fix what Steps 1 to 4 found, then wait. Rankings and indexing can take days to months to respond.
- Do not rewrite pages in response to a hunch. If your evidence is only "impressions held, clicks fell", you do not yet know what to rewrite.
- Change what you measure. Google recommends measuring the overall value of visits, such as sales, signups and engagement, and not raw click counts alone. Put those next to clicks in your reporting dashboard.
- Keep AI referral traffic separate. Visits from assistants such as ChatGPT are a different source from Google clicks. AI citations vs referral traffic explains why the numbers answer different questions.
Frequently asked questions
Can I see AI Overviews traffic separately in Search Console?
Partly. Google's AI features documentation says sites shown in AI features are included in overall search traffic and reported in the Performance report under the Web search type. Google also offers a Generative AI performance report, launched for some sites in June 2026 and available to all sites since August 31, 2026, but its help page describes impressions only, covering AI Overviews and AI Mode, with no clicks, queries or position. At the time of writing you can see impressions there, not a click total for those features.
If clicks dropped but impressions did not, is it AI Overviews?
It is consistent with that, but it is not proof. Google's debugging guide points to title, snippet and rich result changes for that pattern. Search your lost queries and see whether an AI Overview appears and whether it links to you.
How long should I wait before deciding a change worked?
Google says changes can take a few days to several months to take effect, and suggests waiting a few weeks before analyzing again after an update. Use the same compare view before and after, with the same filters.
Do I need llms.txt or special markup to appear in AI Overviews?
Google says no. Its documentation says AI Overviews and AI Mode need no extra technical requirements, special optimizations, AI text files or special schema.org markup. A page must be indexed and eligible to show in Search with a snippet.
Check the technical side of your pages
If Step 4 raised a doubt about access or indexing, the free check reads one page without an account: robots.txt rules for named AI crawlers, the robots meta tag and headers, and whether the main text is in the raw HTML. It does not read your Search Console data or measure traffic. For a whole site, an audit runs the checks across up to 500 pages with evidence and prioritized fixes.
Filed under Strategy and audits. Topics: google, ai-overviews, analytics, troubleshooting