Traffic Down but Rankings the Same? How to Find the Real Cause

If your traffic is down but your rankings look the same, one of three things happened: fewer people clicked, fewer people searched, or your analytics stopped counting them. Search Console can separate those in about an hour. Most of the time the answer is lower click-through, and the fix depends on which queries lost it.

Most guides on this stop at “check your CTR and look for AI Overviews.” That misses two problems. Search Console’s own impression data was wrong for almost a year, which breaks the comparison most people run first. And the “position” number can look healthy after your normal listing has slipped.

Why can’t you trust a year-over-year comparison right now?

Google’s data anomalies page says a logging error prevented accurate impression reporting from May 13, 2025 through April 27, 2026. Google says it affected impressions, CTR and average position, but not clicks. It was a logging problem only, so it says nothing about how your site actually performed.

So as of today, a year-over-year comparison pits clean recent data against a faulty baseline. Any impression, CTR or position change in that report is suspect. Clicks are fine.

There’s an earlier step change too. In September 2025, Search Engine Journal reported sharp desktop impression drops and better-looking average positions after Google’s num=100 results parameter stopped working. Google hadn’t commented when that piece ran. One community theory was that SEO and AI tools loading 100-result pages had been inflating impressions.

Our rule for this diagnosis:

• Year over year: use clicks only, plus Google Trends for demand.
• Impressions, CTR and position: compare two periods that both start after April 27, 2026. “Last 3 months vs previous period” clears that date as of November 2026.
• Leave out the last 2 to 3 days. Google says data is usually available after that lag.

How do you split the loss into causes?

Open Performance, Search results, then Compare. Filter to Web search and export clicks and impressions for the two periods. Then do this sum, once for the site and again for each important page group:

• Demand effect = change in impressions × old CTR
• CTR effect = new impressions × change in CTR

The two add up to your total change in clicks. Here’s a hypothetical example, made up to show the arithmetic:

Clicks fell by 1,250, or 39%. Demand effect: −2,000 × 4.0% = −80 clicks. CTR effect: 78,000 × −1.5% = −1,170 clicks. That’s −1,250 in total, and 94% of the loss is CTR.

This owner shouldn’t be buying links or rewriting pages to win rankings they already hold. If 2% of those visits became leads at $800 each (again, invented numbers), the loss is 25 leads and $20,000 a quarter. Put your own numbers in, because that figure tells you how much this investigation is worth paying for.

What does each pattern mean?

Those last checks come straight from Google’s guide to debugging traffic drops, which also suggests a “Last 16 months” view to rule out a drop that recurs every year. Do that before anything else if you’re seasonal. For a drop that lines up with a new site or new URLs, use our checks for traffic loss after a redesign instead.

Tracking is the cheapest cause to rule out, and it’s a separate system. Google notes that some tools, such as Analytics, only count users with JavaScript enabled, and the two products will never match exactly. A gap isn’t the signal. A gap that opened suddenly is.

How can position stay flat when your listing actually dropped?

Average position is easy to misread for three reasons.

AI Overview links share one position. Google says an AI Overview occupies a single position and its links share it, and that average position uses your topmost position for each query. In our reading, a citation inside an AI Overview can hold your reported position up even if your normal blue link has moved down the page. If a query shows position 1 and CTR has collapsed, look at the real results page.

Averages hide mix changes. Lose hundreds of low-ranking long-tail queries and your average can improve while your traffic gets worse. That’s why the query-level view matters. Note the table shows at most 1,000 rows and leaves out rare queries, so on bigger sites filter by folder or page.

Your rank tracker may not be measuring what you think. It usually checks one location and device on a schedule. Split the Search Console data by Country and Device before you accept “rankings are the same.”

How much traffic do AI Overviews really take?

Enough to explain a large part of many drops, though the published numbers vary and none of them will be your number.

• Ahrefs compared 300,000 keywords using December 2023 and December 2025 data and found a 58% lower CTR for the top result on keywords with an AI Overview, 46.4% at position 3 and 19.4% at position 10. Ahrefs describes this as a correlation, and it’s desktop data from one month.
• Pew Research tracked 900 US adults in March 2025. They clicked a traditional result on 8% of visits to pages with an AI summary and 15% without one. About 1% clicked a link inside the summary.

Note what those studies don’t say. They don’t tell you which of your queries lost clicks, or whether your commercial pages were touched at all. To find that, take your 20 biggest losing queries and search each one on a phone in a clean browser. Write down what sits above you: AI Overview, ads, map pack, videos, shopping. It’s dull and it’s the best hour you’ll spend on this.

If your property has Google’s generative AI reports, announced in June 2026 for a subset of sites, use them to see which pages appear in AI features. As described at launch they showed impressions, not clicks, so they tell you about exposure and not about loss.

Not sure whether your drop is CTR, demand or tracking? That’s the kind of diagnosis our team does. Talk to SEOIndexr about a traffic drop review.

What should you do once you know the cause?

If an AI Overview now answers the query: decide whether the query ever made you money. If it’s a definition or a basic how-to that never led anywhere, in our view you should accept the loss and stop reporting on it. If it’s close to a purchase, it’s worth working on citations and on pages that give a reason to click, like pricing, comparisons and real examples. Our guide on optimizing for AI Overviews covers the basics, and our generative engine optimization services cover it if you want it handled.

If the listing itself is the issue: this is the case Google’s guide flags, where impressions hold and clicks fall because your snippet is weak or competitors have richer results. Rewrite titles and descriptions on the top losers only, then compare after a few weeks. Don’t touch pages that hold their CTR.

If a rich result vanished: check Search appearance first. Google’s anomalies page says FAQ rich results stopped appearing in Search from May 7, 2026. If you lost clicks around then on FAQ-heavy pages, that’s why, and no amount of markup work will bring it back.

If demand fell: confirm it in Trends. If the whole market is down, SEO can’t fix that. Move effort to adjacent queries that are growing, or to other channels.

If tracking broke: fix it, annotate the date and stop the panic meeting.

What should you ask your agency?

If someone sends you a report saying “rankings are stable, it’s AI,” send this back. A good agency can answer it in a day. It’s also worth having reporting terms in your contract that make this data yours.

Hi, before we accept AI as the explanation, please send me: (1) the click loss split into demand effect and CTR effect, using periods that both start after April 27, 2026; (2) the 20 queries and 10 pages that lost the most clicks, split by device; (3) what appears above us on those 20 queries today, with screenshots; (4) how many of them drive leads or sales; (5) Search Console clicks against analytics sessions for the same dates. Thanks.

FAQ

Can traffic really fall if rankings haven’t changed?

Yes. Rankings measure where you appear, while traffic depends on how many people search and how many click. Either can fall with no change in position, and AI Overviews, ads or a weaker snippet can cut clicks on their own.

Why do my Search Console impressions look so different from last year?

Partly because of Google’s logging error from May 13, 2025 to April 27, 2026. Don’t read those year-over-year impression changes as real gains or losses. Clicks weren’t affected, so use those for long comparisons.

How long should we wait before acting on a drop?

Check tracking and technical reports the same day, since those are the urgent causes. For the rest, in our view you need enough data to separate a trend from noise, and Google notes small position changes can happen at any time. Two to four weeks of a sustained drop is a sensible trigger for a full review.

Should we block Google’s AI features to get our clicks back?

We wouldn’t start there. You’d give up visibility on the off chance of winning clicks back, and zero-click searches aren’t only an AI feature problem. Do the query-level diagnosis first, and treat any blocking decision as a separate business call.

Sources

• Google: Data anomalies in Search Console
• Google Search Central: Debugging drops in Google Search traffic
• Google: What are impressions, position, and clicks?
• Google: Performance report, troubleshooting data discrepancies
• Ahrefs: Update: AI Overviews Reduce Clicks by 58%
• Pew Research Center: Google users are less likely to click on links when an AI summary appears
• Search Engine Journal: Google modifies search results parameter
• PPC Land: Search Console generative AI reports