Ranking First and Getting Zero Clicks

You rank number one. Your impressions are flat or climbing. Your clicks fell off a cliff. Nothing is broken on your site, and no, you probably did not get hit by an update.

What changed is that position one is no longer the top of the page. It sits under an AI Overview, a map pack, a people-also-ask block, a video carousel, sometimes all four. The ranking is real. The click is gone. And the fix is not more ranking work. It is deciding, query by query, which answers you are happy to give away without a click and which ones you need to recapture with a different page type or a different intent.

Why your rankings held and your clicks collapsed

Position one is now a slot inside a page built to answer, not to route. The AI Overview resolves the question above you. The map pack, the people-also-ask block and the video carousel each take a share of the attention that used to land on the blue links. Your ranking did not move. The page around it did.

And the click decline is not confined to the queries where an AI Overview appears. Click behavior has shifted across the board, which matters because it changes the fix.

That kills the most common advice you are being given right now. "Get cited in the AI Overview" is not a strategy. If clicks are falling on queries where the AI Overview never appears, optimizing for the AI Overview alone cannot restore your traffic. The click is not coming back in the volume you had. Plan around that.

How much of your traffic is actually at risk

Size the exposure before you spend a dollar fixing it.

AI Overviews are not everywhere. They trigger on a subset of queries, concentrated in particular intents, and published trigger rates vary by the dataset and the month they were measured.

Zero-click rates depend entirely on who is counting and what they count. Panels, devices, countries and query mixes all move the number, which is why the published figures disagree with each other. Do not pick the scariest one and build a slide out of it. Measure your own.

An afternoon in Search Console gets you the real answer for your business:

• Export the last 16 months of query-level data.
• Split queries by intent: informational, commercial, navigational, local.
• Chart impressions against clicks for each group, not for the site as a whole.
• Flag every query where impressions held or rose and clicks fell by more than a third.

That flagged list is your actual exposure. It is almost always smaller than the panic, and it is almost never evenly spread.

The queries you should be happy to lose

Intent decides who clicks. Informational and lookup queries resolve on the results page far more often than commercial ones, because the answer is short and the page can hold it. Commercial queries still send traffic, because the decision needs more than a summary.

Most of what you are grieving is informational traffic you never monetized. Definitions. "What is." Conversion calculators. Glossary pages. Top-of-funnel explainers written to catch volume. That traffic rarely converted, it inflated dashboards, and now the answer engine serves it directly.

Let it go uncited and stop paying to defend it. There is a real consolation: being named in the answer still earns attention. People read the answer, recognize the source, and then search the brand. That is what an impression is worth now.

Local is its own category, and it is not a content problem. Map pack and business profile results are built to resolve on the results page. Hours, directions, phone number, reviews, photos. If you are a service business losing clicks there, the work is profile completeness and review velocity, not another blog post.

The queries you need to win back

Keep the list short and commercial. These are the ones where a click is the product.

Comparison and alternative queries. "X versus Y," "best for," "alternatives to." The user wants a position, not a definition. Build a structured comparison page with an explicit recommendation, named tradeoffs and a table the reader actually needs to see in full.

Pricing and cost queries. Publish real numbers or real ranges. A page that answers the cost question honestly is the page a summary cannot replace, because the summary has nothing specific to repeat.

Proof queries. Case studies with the client, the constraint, the method and the measured outcome. First-party data cannot be summarized away, because it does not exist anywhere else.

Tool, calculator and template queries. Interactive beats extractable. The answer requires the user's input, so the answer requires the visit.

Deep implementation queries. The "how do I actually do this in my stack" questions where a four-sentence overview is obviously insufficient.

Notice the pattern. Every page type on that list is one where the AI answer is a trailer and your page is the film. If a 60-word summary fully satisfies the query, that query was never going to be a click.

Getting cited is not the same as ranking

Two different systems, and the distinction matters more than anything else in this article.

Inside Google's AI Overviews, ranking still carries weight. The pages Google cites in its own answers skew heavily toward pages already ranking well in its own results. So inside Google, strong rankings remain a strong input.

Outside Google, the relationship breaks. The overlap between what ranks first on Google and what AI engines actually cite has collapsed from roughly 70% to as low as 17%. ChatGPT, Perplexity, Gemini and Claude are different audiences, different signals and a different optimization discipline. Anyone showing you one number for both is conflating them.

This is why we run AI citation visibility audits across ChatGPT, Perplexity, Gemini and Claude as a separate exercise from search ranking work. The inputs that earn a citation are structural and reputational: schema, structured comparisons, FAQ architecture, and earned third-party citations that these engines trust and repeat.

What to measure now

Re-baseline your reporting or you will keep optimizing against a metric that no longer describes anything.

• Impressions are a brand metric. Treat them like share of voice, not demand capture.
• Clicks matter on commercial-intent queries only. Report that segment separately and let informational clicks fall where they fall.
• Branded search volume is your downstream conversion signal. If the answer engine is doing the awareness work, branded search is where that work shows up.
• Citation share by engine. Track how often you are cited across ChatGPT, Perplexity, Gemini and Claude, monthly, by query cluster.
• Revenue per session, not sessions. Fewer, better-qualified visits is the shape of this market. Judge it on money.

The 30-day triage

Week one, pull 16 months of query data and segment by intent. Week two, classify every flagged query into let it go, defend with a different page type, or recapture with different intent. Week three, rebuild the top five commercial queries as comparison, pricing or proof pages with schema and structured answers. Week four, baseline your AI citation share by engine and fix local profiles if local matters to you.

Ranking first and getting no clicks is not a failure of your search work. It is the end of one scoreboard and the start of another. The brands that will own the next five years are the ones being cited in the answer, not just listed underneath it.

That is the work we do at Payani Media: separating the queries worth defending from the ones worth conceding, then building the pages and the citation footprint that make you the source the answer engines quote.

Frequently asked questions

Why do my impressions keep rising while my clicks fall?

Because position one now sits under AI Overviews, map packs and people-also-ask blocks, so you are still being shown and far less often clicked. Ahrefs found the presence of an AI Overview correlates with a 58% lower click-through rate for the top-ranking page. Treat impressions as a brand and visibility metric, and judge clicks only on commercial-intent queries.

If I get cited in the AI Overview, do my clicks come back?

Partially, not fully. Pages included in AI Overviews get 3.2 times as many clicks as excluded pages, so citation is worth having. But Seer Interactive found organic click-through rate fell 41% even on queries with no AI Overview at all, which means user behavior changed beyond the layout and citation alone will not restore your old traffic.

Which queries are actually worth fighting for now?

Commercial ones. A UK study of 87.6 million monthly queries found 62% of informational searches ended with no click, while commercial searches showed a 21% zero-click rate. Concentrate on comparison and alternative pages, pricing pages, case studies with first-party data, calculators and deep implementation guides, the pages a short summary cannot replace.

Does ranking on Google still help me get cited by ChatGPT or Perplexity?

Inside Google's AI Overviews, yes. Ahrefs found 76.10% of cited pages rank in the top 10, with a median rank of 3. Outside Google it is a different story: the overlap between what ranks first on Google and what AI engines cite has collapsed from roughly 70% to as low as 17%, so those engines need their own optimization discipline.

I run a local service business and my map pack clicks dropped. What do I fix?

Not your content. Map pack and business profile results are designed to resolve on the results page itself, with hours, directions, phone and reviews answered in place. The work is profile completeness and review velocity.