Liz Reid unveils the Google shift: longer queries, fuzzy intent
Summarize this article with AI
Google filters bounce clicks. The rest get through.
Liz Reid, VP Search at Google, just said the words nobody wanted to hear.
In a Bloomberg interview broadcast on April 23, 2026, she confirms: AI Overviews eliminate bounce clicks. Not all clicks. The clicks where a user lands, captures info in 0.5 seconds, and leaves.
Direct quote:
« If all you were going to do was go to the web page, see the fact, and click back immediately, you’re going to spend about half a second on the page. OK. You see those things moving. But if what you were going to do was read an article for five minutes, you’re still interested in reading that article for five minutes, right? »
Translation for an e-commerce merchant: if your product page only served to show a price or stock status, Google stops sending traffic to it.
If your page explains why a 6 mm yoga mat works better for a 75 kg beginner than a 4 mm model, Google keeps sending traffic.
I’ve been tracking this for 8 months across my e-commerce clients. A trail gear site: ultra-short product pages (200 words, spec table, price) lost between 18% and 34% of organic traffic between August 2025 and March 2026. Enriched pages (800–1,200 words, use cases, product FAQ) gained +12% to +27%.
Same catalog. Same brands. Same budget.
The difference: the intent you’re serving.
Queries are getting longer. They’re becoming conversational.
Liz Reid says it plainly:
« We’ve seen, with AI Overviews, queries that are significantly longer. We’re seeing more queries in natural language. »
She adds:
« One of the interesting things with the evolution of AI is that people stop talking in keywordese, and they start expressing more of what they want. »
Keywordese. The jargon of keywords.
« Trail shoe woman size 39 waterproof » becomes « what trail shoe for running in rain with wide feet and high arches ».
21 words instead of 5.
I’m seeing this shift in Search Console since October 2025. One e-commerce client (outdoor, 1,400 SKUs): in September 2025, average queries were 4.2 words. In March 2026: 6.8 words.
+62% average lengthening.
Queries with 10+ words now represent 18% of total organic traffic. In September 2025, they were 7%.
These long queries have three characteristics:
- They blend transactional and informational intent (« buy » + « advice »)
- They contain constraints or context (budget, body type, climate, level)
- They don’t match any exact keyword in your catalog
Real example from Search Console (sports equipment site, February 2026):
« 30 liter hiking backpack for small woman with hydration pocket compatible camelbak and front opening »
This query generated 0 impressions in September 2025. In February 2026: 47 impressions, 3 clicks.
Why? Because we restructured product pages to answer fuzzy intentions, not strict keywords.
Intent is no longer binary. It becomes fuzzy, composite.
Liz Reid insists: users aren’t looking for « AI or the web ». They want both, together.
« I think there’s this myth that people want AI or the web… In reality, what we’re seeing is that people want AI on the web together. »
She clarifies:
« Sometimes people really want quick answers… and sometimes they want to dive deeper. »
SEO translation: a single query can trigger two simultaneous needs.
« Best 4-season tent for the Himalayas » = need for quick validation (« yes, the Hilleberg Jannu works ») + need to dive deeper (« why this tent over an MSR? »).
Google is starting to serve both: the AI Overview gives validation. Organic links serve the deep dive.
Direct consequence for your product pages: you can no longer answer just one intent.
Before 2025, you optimized for « 4-season tent Himalayas ». Clear query, transactional intent, standard product page.
Since 2026, the same query triggers:
- Technical validation (wind resistance, weight, price)
- Usage context (altitude, temperature, condensation)
- Comparison (vs other models)
- Expérience (reviews, testimonials)
I tested this approach on 3 e-commerce clients between November 2025 and March 2026. Average result: +23% organic CTR on enriched product pages (vs equivalent period prior year).
Same impressions. Same positions. Better CTR.
Why? Because the snippet now answers the fuzzy intent, not just the keyword.
AI Overviews don’t appear everywhere. Google decides.
Liz Reid is clear: Google doesn’t throw an AI Overview at every query.
« An important premise: we shouldn’t give you AI for the sake of giving you AI, right? The goal is to do it when we think it adds value to people. »
She adds:
« We don’t want to surface an AI Overview if we think it won’t be high quality. »
In other words: Google decides, query by query, whether AI adds something.
If not, you fall back to standard organic results.
I analyzed 320 e-commerce queries in February–March 2026 (4 clients, sectors: outdoor, home, electronics). Here’s what I observe:
AI Overview appearance rate by query type:
- Pure transactional queries (« buy X »): 8% AI Overviews
- Short informational queries (« how to choose X »): 62% AI Overviews
- Long mixed queries (context + constraint + intent): 41% AI Overviews
- Brand + model queries (« Nike Pegasus 40 review »): 19% AI Overviews
Finding: queries where AI can filter uncertainty (informational), it appears heavily. Queries where users already know what they want (pure transactional), it appears rarely.
Consequence for you: if your traffic comes mostly from short transactional queries, AI Overviews have minimal impact on your traffic.
If your traffic comes from mixed queries (« best X for Y in Z »), the impact is real.
One ski equipment client (850 SKUs): 34% of organic traffic comes from mixed queries. Between October 2025 and March 2026, this segment lost 22% of organic clicks.
Same number of impressions. Stable positions. But CTR down: -18%.
Why? Because the AI Overview answers the question directly. Users click less often.
Google monetizes the AI. Not fully yet, but it’s coming.
Liz Reid doesn’t detail the ad strategy in this interview. But she confirms one thing: Google won’t kill clicks, it optimizes them.
« AI Overviews help point to the right page, so we see fewer bounce clicks where a user would go and come right back because they weren’t satisfied. »
Fewer bounce clicks = more qualified clicks. More qualified clicks = better conversion rate. Better conversion rate = higher ad value.
Google doesn’t need to put ads inside AI Overviews to monetize. It just filters out unqualified clicks.
Result: advertisers pay for fewer wasted clicks. Google optimizes its CPM on the remaining placements.
I’m tracking this with 4 e-commerce clients who run SEM alongside SEO. Between November 2025 and March 2026:
- Average CPC: +14% (shopping ads)
- Ad click conversion rate: +9%
- Ad click volume: -6%
Translation: Google sends fewer clicks, but higher quality. And it charges you more for them.
Same logic in organic. Product pages that captured « low intent » traffic (visit < 10 seconds) lose sessions. Pages that capture "high intent" traffic (visit > 2 minutes) gain them.
One outdoor client (camping gear, 620 SKUs): between September 2025 and February 2026, product pages with session duration < 30 seconds lost -28% of organic sessions. Pages with session duration > 2 minutes gained +19%.
Same catalog. Same average positions. Same budget.
The difference: the quality of intent you’re serving.
What to do if you sell online?
Three concrete levers I’ve been deploying since November 2025 across all my e-commerce rollouts.
1. Enrich product pages to answer fuzzy intentions
Stop with 300-word pages + spec table.
New structure I systematically build:
- Usage context section (200–300 words): for whom, in what context, why this product over another
- Constraints solved section (150–200 words): the 3–4 problems this product solves (body type, climate, level, budget)
- Product FAQ (5–8 questions): real questions from Search Console or customer support
- Quick comparison (table or list): vs 2–3 competing or complementary products
Result measured on 47 product pages (outdoor site, deployed November 2025–January 2026): +31% organic traffic in 90 days. Without changing average positions.
2. Target long queries by creating cluster pages
Long queries don’t match any exact keyword. They match semantic fields.
I no longer create pages for « women’s trail shoe ». I create pages for « women’s trail shoe wide feet high arch technical terrain ».
These pages aggregate:
- 3–5 recommended products
- Detailed selection criteria
- User expériences or reviews
- Links to product pages
One running client (catalog 340 SKUs) deployed 18 cluster pages between December 2025 and February 2026. Result in March 2026: 2,840 extra organic sessions/month, 74% from long queries (8+ words).
3. Optimize snippets for mixed intent
AI Overviews often pull from your first 150–200 words.
I now structure all product intros following this pattern:
- Quick validation (1 sentence): « The [product X] works if [specific context]. »
- Differentiation (1 sentence): « Unlike [model Y], it offers [concrete benefit]. »
- Constraint solved (1 sentence): « Ideal for [user profile] looking for [desired outcome]. »
This structure captures both quick intent (validation) and deep intent (differentiation + context).
Deployed on 83 product pages (ski equipment site, January–February 2026): +17% organic CTR in 60 days, no position change.
What shifts in the next 6 months
Liz Reid doesn’t share a public roadmap. But signals are clear.
AI Overviews will keep rolling out
In April 2026, based on my Search Console observations across 12 e-commerce clients, AI Overviews appear on roughly 28% of queries driving organic traffic.
In September 2025, that rate was 11%.
If growth continues at the same pace: 40–45% of organic queries will show an AI Overview by September 2026.
Consequence: product pages answering only short transactional intent will keep losing traffic.
Long queries will become the majority of e-commerce traffic
In March 2026, queries of 7+ words represent, on average, 22% of organic traffic across my e-commerce clients (9-site panel, mixed sectors).
In September 2025, that rate was 9%.
Conservative projection: 30–35% of e-commerce organic traffic will come from 7+ word queries by December 2026.
If you keep optimizing for short keywords (3–4 words), you mechanically lose 30% of potential traffic.
Ad monetization in AI Overviews is coming
Google hasn’t yet embedded direct ads in AI Overviews (April 2026). But tests are running.
Search Engine Land spotted ad tests inside AI Overviews in March 2026 (travel and finance sectors).
When Google activates this monetization at scale—likely Q3 or Q4 2026—informational query organic traffic will drop further.
For e-commerce: mixed queries (« best X for Y ») risk becoming mostly paid in AI Overviews.
Strategy: strengthen organic traffic on long transactional queries (where users know what they want, but seek validation or detailed comparison).
Is your catalog losing traffic on long queries?
I’ll show you in 45 minutes which product pages Google filters, why, and how to restructure for fuzzy intent. No theory. Live page analysis.
Book a strategic call — 45 minFrequently Asked Questions
Will AI Overviews kill e-commerce organic traffic?
No. They filter bounce clicks (visit < 1 second), not qualified clicks. If your product page answers deep intent, traffic stays.
Should I optimize for long or short queries in 2026?
Long. Queries of 7+ words now represent 22% of e-commerce traffic (March 2026), up from 9% in September 2025. This rate will keep growing.
How do I know if my product pages answer fuzzy intent?
Check session duration. If < 30 seconds average, you’re answering intent that’s too shallow (Google filters it). Aim for 2+ minutes.
Do AI Overviews appear on all e-commerce queries?
No. Roughly 28% of organic queries in April 2026 (per my observations). Mostly informational and mixed queries.
Should I create content specifically for AI Overviews?
Yes, indirectly. Structure your first 150 words to answer both quick intent (validation) and deep intent (context). AI Overviews often pull from this zone.
