Google adds links to AI Search without click data: e-commerce impact
Summarize this article with AI
The expansion of links in AI Search: what happened?
A client calls me on a Tuesday morning. An e-commerce site with 1,200 product SKUs. Organic traffic down 22% in three months. He changed nothing in his SEO. Nothing. Zero. Invisible.
Yet in Search Console, no alert. Total clicks are dropping, but impossible to know if AI Overviews are responsible. Because Google doesn’t provide the data.
I looked closer. According to Search Engine Journal, Google quietly added five new link surfaces to its AI Search in April 2025, without offering any dashboard for publishers. Rich cards, suggestions, contextual links. All of this multiplies opportunities for sites to appear—but with zero traceability.
Reminder: AI Search is the synthetic answer Google places above blue results. It aggregates sources, summarizes, answers. And sometimes it tacks on links. These are the links e-commerce sites care about.
The problem? No way to know how many clicks these new placements bring. Google counts them in overall organic traffic without isolating them. Total opacity. Undetectable.
Yet studies are multiplying. Pew Research Center analyzed 68,000 queries across more than 900 adults in May 2025. Result: users click on an organic result in 8% of cases when an AI Overview is present, versus 15% without it. Only 1% of clicks go to a link inside the AI answer itself.
You see it? 8% instead of 15%. The difference is massive. And nobody can tell you whether you’re losing or winning.
Click data in silence: Google’s denial
Google shifted its message at each wave of criticism. At the AI Overviews launch in May 2024, publishers immediately raised alarms. But the giant replied: « No data to share. »
« We have no data to share. » — Google Marketing Live, May 2025.
Then, facing DMG Media studies revealing CTR drops as high as 89%, Google moved the debate. Words from Liz Reid, VP Search: « The clicks that remain are higher quality. » Without numbers to back it up.
The term « bounce clicks » even appeared in October 2025 to describe clicks that AI Search replaced, supposedly with no value. But again, no metrics were published.
This semantic dance has a name: commercial denial. Google doesn’t want to show that its AI Search cannibalizes free traffic, because that could threaten its advertising ecosystem. For an e-commerce site, it’s a black box.
I witnessed it with a motorcycle parts client. His product page traffic dropped 31% in six months. Impossible to link the decline to AI Search via Search Console. Yet « buying guide » queries displayed AI Overviews almost everywhere. Coincidence? None.
The result? Blind decisions. You invest less in content, when actually, now is the time to restructure it.
Quantified impact on e-commerce sites
I compiled available data:
- Pew Research Center (May 2025): 8% clicks with AI Overviews vs 15% without. A 47% drop in overall click-through rate.
- DMG Media (CMA report): CTR drops as high as 89% on certain commercial queries.
- Digital Content Next (2026): median 10% decline in organic traffic across 19 major publishers.
- Reuters Institute (2025): publishers anticipate search traffic drops exceeding 40%.
These numbers are not estimates: they are field measurements. Yet impossible to link them to your site. No report. Zero.
For e-commerce, impact is even harsher. Picture a product page for « best cordless drill » that was driving 400 sessions a month. Once the AI Overview lands, Google summarizes the comparison in three lines and places a carousel of links. Your page, now in 12th position, sees only 30 sessions. Pure loss: 93%.
But you cannot measure it directly. Making diagnosis nearly impossible. And therefore, impossible to fix.
Measuring the invisible: 3 methods beyond Search Console
Not being able to measure doesn’t mean you can’t act. I used three approaches with my clients to trace lost clicks.
1. Server log analysis. Every visit leaves a trace. I search for entries with the google.com referrer and missing query parameters from classic data, then cross-reference with AI Overview display times (which I detect manually via SERP monitoring tools). By overlaying, I identify clusters of pages suddenly getting less traffic. A organic cosmetics client discovered this way that 11 SKUs lost 60% of their traffic.
2. Fine-grained time segmentation in GSC. I compare total clicks across identical periods before and after major updates, accounting for seasonality. Even if AI clicks aren’t isolated, the drop reads in the trend. In 2026, I see e-commerce sites losing between 12 and 22% of organic clicks with no other explanation.
3. Revenue tracking by source. The most pragmatic approach: stop looking at clicks, start looking at revenue. If organic revenue doesn’t drop while traffic falls, remaining clicks convert better. I helped Julien, a motorcycle parts client, shift to this tracking. He found his average order value rose 14%, offsetting the volume decline.
To remember: you cannot track native AI clicks. But you can isolate their impact by exclusion. It’s tedious, but it’s the only way.
What nobody says: AI Search doesn’t kill clicks, it shifts conversion
Google repeats that remaining clicks are « higher quality. » Logically, that makes sense: a user who clicks after reading a synthetic answer probably has stronger intent. But without data, it’s an unverifiable claim.
Yet on the ground, I observe signals that confirm this hypothesis. A connected watch shop saw organic traffic drop 18% between May and September 2025, but conversion rate jumped from 1.9% to 2.8%. Its monthly revenue stayed stable. So the lost clicks were perhaps just noise.
Another client in apparel even saw a 9% revenue jump thanks to higher average order value on remaining traffic. Why? AI Overviews tend to answer transactional questions upfront, leaving committed buyers to click product pages.
That’s where it pinches: e-commerce sites optimizing product pages for traffic (keywords, technical specs) see their pages demoted in favor of synthetic results. Those investing in answer-first content (guides, comparisons, semantic cocoons) get cited in AI Search more often and benefit from qualified residual traffic.
The key isn’t mourning lost clicks—it’s understanding the game has shifted: real contact surface is no longer the blue link, but mention in the answer. That’s exposure influencing without users leaving the SERP. And that doesn’t count as clicks. It counts as revenue.
Reshaping architecture: semantic cocoon in 2026
To be cited in AI Search, you need to answer precise questions. And those questions cluster around themes. Enter the semantic cocoon.
I’m not selling magic. But I’ve applied the DOSE framework (taught by Guillaume Attias at BMO Academy) since 2016 to structure client content. The principle: each page answers one specific intent, all interconnected by relevant links. The ensemble covers a need without internal competition.
Concrete example: for a supplement site, I built a cocoon around « magnesium. » A pillar page answering why to take it, then articles on each form (bisglycinate, citrate), detailed product pages, buying guides. All linked. AI Search today pulls excerpts from these pages.
Why works? Because Google’s algorithm hunts for authoritative sources to build its answer. A site with coherent semantic structure signals « I own this topic. » And when an AI Overview needs three links, it often picks from these cocoons.
Result? For this site, organic traffic fell 14% after AI Overviews landed, but mentions in answers exploded, boosting awareness and indirectly driving sales through other channels. E-commerce revenue grew 22% over 12 months.
Reshaping architecture isn’t optimizing for clicks. It’s optimizing to be the cited source. And that takes holistic vision, not keyword-by-keyword thinking.
The question every e-commerce site should ask
If tomorrow Google deleted all click data from AI Search, how would you steer growth?
That’s the question I ask every client in audit. The answer determines survival. Those clinging to traditional reporting go blind. Those accepting invisibility and building presence as a conversational window keep selling.
E-commerce SEO in 2026 isn’t won on CTR. It’s won on citation. On the ability to be the trusted source Google displays in its answer. Direct traffic, word-of-mouth, trust—all flow from this silent exposure.
So stop measuring what doesn’t exist. Measure what matters: revenue, conversion rate, average order value. And build a site worthy of being cited.
I don’t sell method. I show pages.
And you? How many of your pages are ready to be cited by AI?
Audit your AI Search impact
In 45 minutes, I analyze your traffic, product pages, and cocoons live. You leave with the method to measure the invisible and restart growth.
Book a strategic call — 45 minFrequently Asked Questions
Why won’t Google share AI Search click data?
Likely to hide negative impact on free CTR while keeping control of its ad ecosystem. Lost clicks threaten the editorial model, and Google has zero incentive to expose them.
How do I know if my traffic is impacted by AI Overviews?
Compare organic clicks before/after launch in Search Console, cross-reference with server logs, and track revenue by source. Even without isolated data, the drop reads in the trend.
Are the remaining clicks really higher quality?
Google claims yes but offers no proof. In the field, some shops see conversion rates rise, suggesting lost clicks were low-intent. Verify case by case.
Does semantic cocoon help you show up in AI Search?
Yes, because it structures pages around precise answers and authority. Google often cites sources organized in thematic silos when building synthetic responses.
Should I still invest in e-commerce SEO if clicks keep dropping?
Absolutely, but shift the goal: from blue-link clicks to AI answer citations. A well-structured site drives qualified indirect traffic and reinforces brand authority.
