Written by: Arjun Karnik, Growth Marketing Specialist

Key Takeaways

  • Google AI Overviews create a “scissors pattern” where impressions rise while clicks fall, with a 47% CTR drop when summaries appear.
  • Being cited inside an AI Overview drives 120% more clicks than appearing on the same SERP without a citation, so citation becomes the key competitive metric.
  • Query fan-out means each page must answer dozens of hidden sub-questions, so traditional keyword rankings no longer guarantee AI citations.
  • Content freshness now drives visibility. Most cited pages were updated within the last year, and some pages lose 78–99% of impressions in two months without updates.
  • Work with Arjun Karnik to map your scissors pattern and see which queries are losing clicks to AI Overviews.

Why AI Overviews Push Impressions Up and Clicks Down

The scissors pattern reflects a shift in how search results deliver value, not a collapse in rankings. Your content still powers AI answers. It just sends fewer visitors back to your site.

Pew Research Center tracked 68,879 real Google searches from 900 U.S. adults in March 2025 and found that users clicked a traditional result in 8% of visits when an AI summary appeared, compared with 15% when no summary appeared. That is a 47% relative CTR drop. Users clicked links inside the AI summary only 1% of the time.

Bar chart comparing click-through rate on a traditional search result, 15 percent with no AI summary shown and 8 percent when an AI summary is shown. Source: Pew Research Center, July 2025, 900 US adults across 68,879 Google searches.
The click roughly halves when an AI summary appears above the result. Pew also found only 1 percent of users clicked a link inside the summary itself.

The zero-click problem extends beyond AI Overviews. Similarweb clickstream data shows the zero-click rate for Google searches reached 68.01% in January through April 2026, up from 60.45% in 2024, with only 276 out of every 1,000 Google searches resulting in a click to the open web. AI Overviews now appear in approximately 48% of Google search results as of 2026.

Seer Interactive’s April 2026 study of 53 brands, 5.47 million queries, and 2.43 billion impressions found that organic CTR on queries showing AI Overviews rebounded from 1.3% in December 2025 to 2.4% in February 2026. The more important finding is that being cited inside an AI Overview produced +120% more organic clicks per impression than appearing on an AIO-present SERP without a citation. Citation now drives the incremental clicks that remain.

Ahrefs’ December 2025 analysis of 300,000 keywords found a 58% drop in position-1 organic CTR when an AI Overview is present. Rankings hold. Clicks fall. The scissors open.

Line chart showing the scissors pattern over twelve months, with an impressions line rising while a clicks line falls away from it. Illustrative shape of the pattern, not data from a specific account.
Both lines start together. The content keeps getting read so impressions rise, the answer gets delivered on the results page so the click never happens. Most owners see only the falling line.

The scissors pattern explains what is happening to your traffic. The next step is understanding why your content is not being cited even when rankings stay strong. That answer lives in how AI systems retrieve and assemble information.

Why Strong Rankings Still Miss AI Citations

Fan-out queries explain why AI often skips your site. A single buyer prompt rarely triggers a single lookup. It triggers dozens of hidden retrieval sub-queries, and the AI assembles its answer from everything those sub-queries return.

Query fan-out has replaced one-keyword-one-page ranking logic with topical-breadth evaluation, so pages that rank #1 for a head term but fail to address related sub-questions are frequently passed over for citation.

32.9% of citations come exclusively from fan-out sub-queries that have zero search volume in traditional keyword tools. Keyword tools can show strong rankings while ChatGPT still names a competitor.

In my own test lab, I extracted fan-out queries directly from ChatGPT for my target topics. I then rewrote URLs, titles, H1s, and H2s on a set of pages to match that language exactly and held a matched set of control pages back. The rewritten pages earned citations. The controls did not. The only consistent difference was fan-out alignment.

Ahrefs analysis of 863,000 keywords showed that the share of AI Overview citations coming from pages ranking in Google’s top 10 dropped from 76% in July 2025 to 38% in early 2026. Traditional rankings and AI citations now move on separate tracks. Optimizing for the visible keyword while ignoring fan-out targets the wrong surface.

Query-passage semantic alignment is 7.3× more predictive of citation than domain authority. The system matches questions to specific answers, not to a seniority list.

How Quickly Content Decays in AI Search

Content now decays far faster in AI search than most teams expect. In my own decay tracking on my test lab site, pages dropped 78% to 99% in two months without updates. That is a cliff, not a gradual slide, and it often completes before a monthly report runs.

Independent research points in the same direction. Seer Interactive analyzed 47,097 AI citations across 7,683 pages in ChatGPT, Gemini, and Perplexity between March and June 2026 and found that 75% of cited pages had been updated within the last year, with pages cited consistently across all four months averaging under six months since their last update.

Bar chart showing 75 percent of pages cited by AI assistants were updated within the last year and 25 percent were older. Source: Seer Interactive, July 2026, 7,683 pages and 47,097 citations across ChatGPT, Gemini and Perplexity.
Three quarters of cited pages were updated inside a year, and the consistently cited ones averaged under six months. The page you refresh beats the page you write.

76.4% of pages cited by ChatGPT were updated within the prior 30 days. Content freshness accounts for 40% of Perplexity’s ranking signal, and pages under 30 days old receive 3.2× more citations than older content.

Approximately 50% of sources cited for a given prompt will change within 13 weeks. The source set resets every few weeks. A fixed content library decays in place. The page you refreshed outperforms the page you only wrote once.

On my own site, I run impression-decay tripwires via AI Growth Agent (I am a partner and disclose it) that auto-queue updates when a page’s performance drops past a set threshold. The library repairs itself on a loop instead of waiting for a quarterly audit that arrives after positions disappear.

SEO versus GEO: Metrics That Actually Drive Citations

The table below shows how SEO and GEO differ across what they optimize for, which authority signals matter most, how success is measured, and how freshness works in practice.

Dimension SEO GEO Source
Optimizes for Human-ranked lists and domain authority Machine retrieval and citation inside AI answers Machine Relations, 2026
Authority signal Backlinks (correlation with AI Overview inclusion: 0.218) Branded web mentions (correlation: 0.664), entity density (pages with 15+ Knowledge Graph entities show 4.8× higher selection probability) Ahrefs 75,000-brand study; Machine Relations, 2026
Primary success metric Rank position and organic CTR Citation rate, share of answer, AI referrer traffic Birdeye State of AI Search 2026
Freshness requirement Periodic updates improve rankings incrementally Most cited pages are recent (see the 75% figure above), and stale content loses citations at 3× the normal rate past the three-month threshold Seer Interactive, July 2026; OptimizeGEO, 2026
Bar chart comparing correlation with AI Overview visibility, branded search volume at 0.392 against backlinks at 0.218. Source: Ahrefs study of 75,000 brands.
Branded search correlates with AI Overview visibility almost twice as strongly as backlinks do. The authority model that governed SEO is not the one governing this.

Citation selection depends more on structural properties such as entity density, content structure, and query-passage alignment than on traditional authority signals like backlinks (r² = 0.038) or traffic (r² = 0.05). The ranking game and the citation game now run on different engines.

Technical Plumbing That GEO Needs Before Anything Else Works

Technical structure comes first because AI search engines test reachability, parseability, trust, verification, and freshness before they consider citation. AI search engines check whether a page can be reached, parsed, trusted, verified, and dated before deciding whether to cite it. Any content investment that sits on broken plumbing fails quietly.

The technical checklist, in order of dependency:

  1. Unblock AI crawlers. Check robots.txt for blocks on GPTBot, Google-Extended, PerplexityBot, and ClaudeBot. This is the most common silent blocker and the first fix, because nothing else matters if crawlers cannot reach your pages.
  2. Implement schema markup on every page. Once crawlers can access your content, help them interpret it. Pages with HowTo, FAQ, or Article schema markup are included in AI Overviews at a rate 37% higher than equivalent unstructured pages. FAQPage, Article, and HowTo schema are the priority types. Pages with stacked FAQPage + Article + HowTo schema see up to a 1.8× improvement in AI citation frequency.
  3. Make pages machine-parseable. After structure comes clarity. Use answer-first H2s, short declarative sentences, tables, and numbered lists. Tables increase citation likelihood 2.5× while FAQPage schema correlates with 2.4–3.2× higher appearance or roughly 44%+ citation lift.
  4. Align query language in URLs, titles, and H1s. Once pages are parseable, match the language the machine retrieves against instead of internal jargon. The slug, title tag, H1, and H2s should mirror buyer phrasing. In my own buyer-language test, relabelling a page titled “What is GEO” to “How to Get Your Business Recommended by AI Search” produced citations within weeks of that specific change.
  5. Audit what AI currently says about your brand. When the plumbing works, check the output. A wrong AI answer hurts more than no answer, so run the defensive audit before any growth work.

Measurement Shifts: From Rankings to Citations and Share of Answer

Measurement must follow buyer behavior. Rankings track a surface many buyers now skip. Citations track whether your business appears in the answer buyers actually read.

The measurement stack that matches this new channel includes four layers that work together:

  • Citation monitoring across ChatGPT, Perplexity, Gemini, and Google AI Overviews. Start with a fixed set of buyer questions and test them at least monthly. Record the exact prompt, engine, date, and cited sources for each run. AI Growth Agent clients average more than 12,000 additional AI citations and mentions, over 100,000 additional bot visits, and a 20% or greater lift in impressions across the first twelve weeks.
  • AI referrer tracking in analytics. Segment chatgpt.com and equivalent referrers as a distinct traffic class. This traffic behaves like referral traffic because the assistant recommended you before the visit.
  • Impression and decay curves in Google Search Console. Use the scissors chart as a diagnostic. Impressions rising while clicks fall on informational queries, with average position holding, signals AI Overview impact.
  • Share of answer as the headline metric. Track what percentage of target prompts return your brand as a named, cited source across all four surfaces. This metric ties directly to visibility inside answers.

One caveat applies across this stack. Buyers often copy an answer and type a brand name directly into a browser, which lands in analytics as direct traffic. Whatever you measure is a floor, not a ceiling. The bot visit and impression lift mentioned earlier provides a useful proxy when direct attribution is incomplete.

Bar chart showing 2.5 percent of downstream brand visits after an AI mention carry a trackable referral parameter while 97.5 percent arrive untraceable. Source: Profound, analysis of more than 2 million AI conversations, January to June 2026.
Buyers read an answer, then type your name into a browser. That visit lands as direct or branded search, so whatever you measure here is a floor and never a ceiling.

Cross-engine citation overlap sits near 11%, meaning an analysis of 680 million citations found only about 11% of domains cited by both ChatGPT and Perplexity. Monitoring one surface and assuming it represents the others produces a false picture, so each surface requires its own tracking.

Action Plan: Fan-out Mapping, AI Audits, and Decay Tripwires

Execution works best in a specific order. Fix technical plumbing before content. Run a defensive audit before growth work. Map fan-out queries before publishing. Build a freshness loop before calling the system complete.

  1. Run a baseline visibility audit. Query your brand and your target topics across ChatGPT, Gemini, Perplexity, and Google AI Overviews. Record what the assistants currently say, where competitors appear instead, and where the gaps sit. This becomes the control group for every later result.
  2. Fix technical plumbing. Unblock AI crawlers, implement schema, and make pages machine-parseable. No downstream work performs until the retrieval layer can read the site.
  3. Correct wrong AI answers. If the audit surfaces inaccurate information about your brand, correct it before any growth work begins. This protects reputation while you expand reach.
  4. Extract fan-out queries from ChatGPT directly. Pull these queries from the assistant instead of inferring them from keyword tools. The target is the machine’s internal questions, not only the human’s typed query. Build your production queue from this map.
  5. Rewrite URLs, titles, H1s, and H2s to match buyer language. Replace practitioner jargon at the exact points where the machine matches questions to answers. Clear buyer language improves both retrieval and conversion.
  6. Publish structured pages at machine cadence. Via AI Growth Agent, my own site runs 5 to 8 autonomous actions per day, mixing new articles with updates. New articles reached thousands of monthly Google impressions within weeks, and the GEO subfolder became the only source of new impressions on the domain in 60 days.
  7. Set impression-decay tripwires. Wire automated triggers to Search Console signals that queue content updates when performance drops. In my tests, pages lost 78% to 99% in two months without maintenance, and tripwires caught the drop before positions vanished.
  8. Shift the reporting dashboard. Replace rank position as the headline metric with citation rate, share of answer, and AI referrer traffic. Report on the channel buyers actually use.

Frequently Asked Questions

Does appearing in AI Overviews actually drive more clicks than not appearing at all?

Being cited inside an AI Overview produces significantly more clicks than appearing on the same SERP without a citation, although it still produces fewer clicks than queries where no AI Overview appears. In practice, citation becomes the new competitive divide. On an AI Overview-present SERP, the gap between cited and uncited sources functions like the old page-one versus page-two split. Businesses that earn citations recover a meaningful share of the click volume that zero-click behavior removes, while uncited businesses absorb the full CTR collapse. The measurement target therefore shifts to citation rate, not rank position, because rank alone no longer predicts the outcome.

How is GEO different from SEO, and do I have to choose between them?

SEO and GEO serve different retrieval mechanics. SEO builds authority through backlinks and domain strength to rank on a human-readable list. GEO builds authority through topical coverage, structured content, and freshness to earn citation inside a machine-generated answer. The authority signals overlap but do not match. Branded web mentions correlate with AI Overview inclusion at nearly three times the strength of traditional backlinks. The technical fundamentals, such as clear structure and relevant content, support both. What changes is the target you optimize toward and the metric you report on. You do not stop SEO. You expand your framework so the same content can serve both retrieval layers.

How long does it take to see results from GEO work?

Coverage and impressions often appear within weeks of publishing structured, fan-out-aligned content. Citations across ChatGPT, Perplexity, Gemini, and Google AI Overviews generally follow within one to three months of consistent publishing and refreshing. Compounding, where topical authority accumulates and citation rates rise across head terms, usually begins after month three. On my own site, new articles reached thousands of monthly Google impressions within weeks, and the GEO subfolder became the only source of new impressions on the entire domain within 60 days. These results came from a system running at machine cadence with structured content, schema on every page, and impression-decay tripwires maintaining freshness. A single article published once and left to decay will not produce the same outcome, because the freshness loop sustains whatever the initial publishing achieves.