A client called me in a bit of a panic in June. Her practice's "how much does X cost" page — historically one of her best lead generators — had impressions up 18% quarter over quarter in Search Console. Same rankings, same position 3-4 average. But clicks were down almost a third, and the contact form submissions that used to come off that page had gone quiet. Her first read was that something had broken. Analytics, the form, the page itself. We checked all three. Nothing was broken.
What had actually happened was simpler and, in a way, worse: Google was answering her exact question directly in an AI Overview, sitting right above her result, built partly from her own page. People were getting the number they came for without ever clicking through. The page was still ranking. It just wasn't earning the visit anymore.
This is happening across a lot of accounts right now, and most owners are reading it wrong. They see a traffic dashboard trending down and assume they got outranked, or that an algorithm update hurt them, or that their SEO stopped working. Sometimes that's true. But increasingly, the honest diagnosis is narrower and more specific: the ranking held, the impressions held or even grew, and only the click behavior changed. That's a different problem with a different fix, and you cannot fix it if you're still treating it like a ranking problem.
If impressions are flat or rising for a query but clicks and CTR are falling while your average position hasn't gotten worse, an AI Overview or other answer-in-place SERP feature is very likely absorbing the click. Segment Search Console by query and date range to confirm it, then rewrite the page to win the citation, not just the ranking.
Why this is easy to misdiagnose
Most traffic dashboards — GA4's default acquisition overview, most SEO tool summaries — show you sessions and users over time. When that line drops, the instinct is to ask "what did we do wrong" or "what did Google change." Both are reasonable questions. But they only cover half of what can cause a decline.
The other half lives in Search Console, in a report almost nobody checks weekly: the Performance report's impressions and CTR columns, broken out by individual query, per Google's own Search Console documentation. Google explicitly separates impressions (your result was shown) from clicks (someone clicked it), and defines CTR as clicks divided by impressions. A query can have rising impressions and falling clicks at the same time — Search Console will show you that pattern plainly if you go look for it, but it will not summarize it for you. You have to build the comparison yourself.
That gap between "shown" and "clicked" is exactly where AI Overviews live. Google's own Search Central documentation on AI features confirms that a page can be cited as a source inside an AI Overview and that this still counts as an impression in Search Console when your link appears — the click behavior is what changes, because a meaningful share of users get their answer from the summary itself and never scroll to the organic results underneath it.
What we actually know, and what we don't
I want to be careful here, because a lot of the "AI Overviews are killing your traffic" content out there cites numbers that don't hold up under a source check. So here's what's actually documented, versus what's still an open question.
What's documented: Google has publicly confirmed AI Overviews are a live search feature rolling out broadly, and its Search Central AI features documentation describes how sources are selected and cited, and explicitly notes that the same ranking systems used for classic search results feed AI Overviews — meaning the same technical SEO, content quality, and E-E-A-T fundamentals still apply to whether you're eligible to be cited at all. Google has also stated in its own communications that clicks from pages featured in AI Overviews tend to be higher quality — the reasoning being that a user who clicks through after reading a summary is more informed and further along, though Google has not published a controlled study with numbers you can independently verify.
What's contested or unverified: You'll see specific percentage figures thrown around — "AI Overviews cut clicks by X%" — attributed to various SEO tools and industry studies. Some of that research is directionally useful, but methodologies vary wildly (different query sets, different industries, different time windows), and a lot of it gets repeated as gospel without anyone checking the original methodology. I'm not going to hand you a fabricated industry-wide number here, because the honest answer is that the size of the effect varies enormously by query type, industry, and even by day, and the only number that matters to you is the one sitting in your own Search Console account.
That's the whole point of this article, honestly. Stop looking for someone else's average. Go pull yours.
The GA4 and Search Console teardown
Here's the actual process we run when a client says "our traffic dropped." This takes about 30-40 minutes the first time, faster after that.
- Start in Search Console, not GA4. GA4 only sees people who clicked through. If the problem is upstream of the click, GA4 will just show you a symptom with no cause. Go to Performance > Search results.
- Set the date comparison to the maximum available window. Search Console retains roughly 16 months of data per Google's own documentation. Compare a recent 3-month block to the same 3-month block a year prior if you have the history, so you're not confusing this with a seasonal dip.
- Sort queries by impressions, descending, then add the CTR and average position columns. Scan for queries where impressions are flat-to-rising, average position is flat-to-improving, and CTR has fallen. Those three conditions together are the signature you're looking for.
- Flag the query pattern, not just the query. AI Overviews trigger more often on informational, question-shaped searches ("how much does," "what is," "how do I") than on transactional or branded ones. If your flagged queries skew that way, the pattern fits.
- Cross-check in GA4. Go to Reports > Acquisition > Traffic acquisition, filter to Organic Search, and compare landing-page sessions for the same URL over the same window. If GA4 sessions for that URL are down roughly in line with the Search Console click drop, and not more than that, you've confirmed it's not a tracking or tagging issue on your end — the loss is happening before the click, not after it.
- Manually search the flagged queries yourself, logged out, in a normal browser. See if an AI Overview appears and whether your page is cited inside it. This is the only step that gives you direct visual confirmation, since Search Console doesn't label which impressions came with an AI Overview attached.
- Separate "cited but not clicked" from "not cited at all." These need different fixes. If you're cited and still losing clicks, the fix is making the surrounding page worth a visit beyond the summary. If you're not cited and a competitor is, the fix is closer to what we cover in why your competitor is cited and you aren't.
Do this per URL, not just site-wide. A site-wide average will hide the queries that are actually being hit hardest behind ones that aren't affected at all, and it'll lead you to the wrong fix.
What to do once you've confirmed it
The instinct is to rewrite the title tag and meta description to be more "clickable." That can help marginally, but it's treating the wrong layer. If someone already has their answer from the AI Overview, a punchier title isn't going to pull them past a summary that already satisfied them.
The fix that actually moves the number has two parts, and they work in sequence:
1. Win the citation on purpose
Put the direct, complete answer to the query in the first two sentences after your H1 — no throat-clearing, no "in this article we'll cover." That's the passage most likely to get lifted and cited, and it matters more now than ever, because if you're not going to win the click, you should at least win the attribution and the brand impression.
2. Give people a real reason to keep going
Right after that answer, the page needs to earn the click for the people who want more than a summary — specifics, your own data, a tool, a comparison, something an AI-generated paragraph can't replicate. If the rest of your page just restates the same shallow answer in more words, you've built a page an AI Overview can fully replace. If it goes somewhere the summary can't follow, you keep the visitors who actually convert.
This is also where the measurement mindset has to shift. Total organic traffic is no longer a clean scoreboard for whether your content strategy is working, because a growing share of "success" now looks like a citation and a brand impression with no click attached. Click volume alone increasingly undercounts your real reach — see winning the zero-click world for how to build reporting that accounts for it.
What this means for your reporting to leadership
If you're the one explaining a traffic chart to an owner or a client, this distinction is worth putting in plain language, because "clicks are down" and "we're losing visibility" are two very different problems that require two very different responses. One means something is broken. The other means the win moved to a different line item, and your job now is proving that it did.
Bring the Search Console query-level breakdown to that conversation, not just the top-line traffic graph. Show impressions holding, show position holding, show CTR falling on a specific, question-shaped set of queries. That's a defensible story grounded in your own data, not a guess borrowed from someone else's average.
Questions people ask
In Google Search Console, filter the Performance report to Search type: Web, then compare impressions to CTR by query over the last 16 months (the maximum Search Console retains). Queries where impressions hold steady or rise while CTR falls, especially informational, question-shaped queries, are the ones most likely to now trigger an AI Overview. Search Console does not label which queries triggered an AI Overview directly, so you are reading the shape of the data, not a tag Google hands you.
No. Ranking position drops, seasonal demand changes, a title or meta description that stopped matching intent, and increased competition can all lower CTR without any AI feature involved. Check average position for the query first. If position held steady or improved and CTR still fell, an AI Overview, featured snippet, or other SERP feature absorbing the click is the more likely explanation.
Do not just rewrite the title. Rewrite the page so the direct answer sits in the first two sentences after the H1, in language AI Overviews can lift as a citation, then use the rest of the page for the depth, specificity, and original data that a short AI summary cannot replace. The goal shifts from winning the click to winning the citation and the clicks that come from people who want more than the summary.
Want this done for you?
Want to know if AI is recommending you or your competitor? Get an AI visibility audit from Acromatico.
Get a free AI Visibility Audit →