Skip to content
Naveed Ganatra

AI Search

Ranking #1 No Longer Guarantees Clicks — Here's the Data

AI Overviews appear on 59% of searches and position-one click-through has fallen from 1.76% to 0.61%. Here is what that means for your top pages, and the four fixes to make now.

Naveed Ganatra6 min read

Prefer to read? The full written version is below.

If your site ranks first and your traffic still fell, nothing is broken. The result page changed underneath you.

This is the shift I spend most of my time explaining to clients right now, so here it is with the numbers attached.

What we used to call strategy

Let me be honest about the era most of us came from, myself included.

We picked a keyword. We put it in the title, the meta description, the URL, the H1. If the article ran 800 words, we made sure keyword density hit at least 2.5%, spread across the body. Then we submitted the URL — Webmaster Tools back then — built links in bulk, and watched it rank.

We called that SEO strategy. It was not strategy. It was keyword matching. There was no thinking in it. Take a string, distribute it, submit, done. It worked because the result page rewarded it, and everyone did the same thing.

Every query has an intention behind it

Here is what that approach could never account for.

Someone types "running shoes." Who are they?

They might be a first-time runner. They might be experienced, coming back from an injury, looking for more support. They might be a parent buying for a child. The string is identical in all three cases and the useful answer is completely different.

Google formalised this back in 2015, classifying search intent into four buckets. Applied to that same runner:

  • Want to know — "how does knee pain from running start"
  • Want to go — "running shoe store near me"
  • Want to do — "how do I fix knee pain when I run"
  • Want to buy — "best running shoes for flat feet"

Same subject, four different pages required. Keyword matching cannot see any of this. It sees one string.

How people actually search now

Type "dietary supplement" into Google and you get very little. It is too ambiguous to answer.

Type the way people genuinely type now — a full situation, in a sentence — and you get a direct, synthesised answer with the key point highlighted.

That is the behavioural change. Search used to be a pile of blue links you filtered yourself, so you typed three or four words and did the work. Now it answers, so people describe their actual circumstances: "I have been running for three weeks, my knees hurt and I have flat feet, what should a beginner get?"

This matters for a specific technical reason: query length predicts whether an AI answer triggers at all. Short one-to-three-word queries frequently do not trigger one. Queries of six or more words very often do. As user behaviour shifts towards longer, conversational queries, the share of searches with an AI answer above them keeps climbing.

The numbers

Before AI Overviews were fully rolled out, they already appeared on around 50% of searches. Post-rollout, that reached 59%.

The click-through consequence for the top organic position: from 1.76% down to 0.61%. Roughly a 60% drop in organic clicks, on the same ranking.

Watch what happens on a real search. Type a question, read the AI Overview, get the answer. I did not scroll once. Whoever ranks first organically was never seen. The AI Overview covered the entire first screen, and if I ask a follow-up it opens AI Mode and I never reach the organic results at all.

The position did not change. The value of the position did.

Ranking first does not get you cited

Here is the part that surprises people most.

If you assume that ranking well means you will be pulled into the AI Overview, that assumption is wrong.

  • Ranking first: roughly a 31.4% chance of appearing in the AI Overview.
  • Ranking fourth: that collapses to about 2.6%.

And in the other direction: around 90% of pages cited in AI Overviews rank beyond position 20 — second page, third page, deeper.

Google has also quietly reduced how much it favours top results here. Where a top-ranking page previously had roughly a 76% chance of being the recommended source, that has fallen to about 38%.

Read those two facts together and the conclusion is unavoidable: AI Overview citation runs on different parameters than organic ranking. It runs on relevance to the specific question. A page sitting at position 25 that answers the question precisely can be cited over a page at position 2 that answers it vaguely.

Which is why sites that never rank organically are being cited, and sites that rank first are being skipped.

Answer position inside the page decides everything

This is the most actionable finding of the lot.

Suppose your page runs a thousand words and the answer a user actually needs is in the tenth paragraph. Google will not dig that far. It scans, takes chunks, finds something relevant or moves on.

So: lead with the answer.

If the page exists to solve a problem, the solution goes at the top — not in the middle, not at the bottom. Everything else can follow. The supporting detail, the context, the nuance: all of it can come after, and it should. But the answer goes first.

This runs against how most people write. We build up. We give background. We arrive at the point. That structure is now costing you citations.

Four things to change this week

1. Rewrite titles as questions

A page title does not need to be a keyword cluster. It needs to match the question being asked. If the page answers a question, phrase it as one — an answer is for a question, a solution is for a problem, and matching that shape helps Google map the page to the query.

2. Move your answers to the top

Take your five highest-traffic pages from Search Console and Analytics. On each one, find the answer the user came for, and move it above the fold. That single edit is the highest-return change available on most sites right now.

3. Add an FAQ section to every important page

Not for the schema — Google's emphasis on FAQ rich results has softened. For the coverage.

An FAQ section lets you address the edge cases around a topic in short, self-contained answers. Short answers in question-and-answer format are exactly the shape AI systems extract and cite. It is the cheapest way to multiply the number of queries a page can be pulled into.

I keep repeating this because it keeps being ignored.

Whatever you are spending on backlinks, move part of it to brand mentions. A mention in a community discussion, a review platform, a third-party roundup — these get crawled, and Google associates your brand name with the subject.

Nobody needs ten thousand links. Ten good ones, plus consistent mentions on relevant sites, will do more. The more your brand is cited, the more it gets recommended, and visibility is what converts into leads now.

The purpose has not changed

Worth ending on this, because the noise around AI search gets overwhelming.

SEO's mechanics changed. Its purpose did not. There have always been two parties: the site owner and the end user. The user is still looking for a product, a service, help with something, or information. Your job is still to be the best available answer.

What changed is where that answer gets displayed and what earns the display. Adapt to that and the channel keeps working. Keep optimising for a result page that no longer exists and it will not.

Tags

  • ai overview
  • click through rate
  • search intent
  • zero click
  • seo strategy
Naveed Ganatra

Written by Naveed Ganatra

SEO Consultant · Karachi, Pakistan

Google-certified SEO consultant with 10+ years building topical authority for logistics, healthcare, legal and e-commerce brands. I publish free SEO training on YouTube.

Get a free audit of your site →

Read next

Free audit

Let’s find out what’s capping your rankings

Send me your domain and I’ll reply with the three issues costing you the most organic traffic right now. No pitch deck, no obligation.

You work with me directly — no account managers, no offshore handoff, no junior taking over after the sales call.