SEO statistics and benchmarks are research figures used to compare search visibility, organic traffic, click behavior and website performance. This SEO statistics collection includes 10 benchmarks from original research: Ahrefs found zero estimated Google traffic for 96.55% of sampled pages in 2023, while Pew measured traditional-result clicks on 8% of AI-summary visits and 15% of visits without summaries in March 2025. Each benchmark includes its source, sample, measurement period and limitations.
Use the dataset to find relevant search statistics, check what each percentage measures and cite the underlying evidence. The 2026/27 edition combines available studies rather than claiming that every statistic was collected in 2026 or that 2027 results already exist.
Explore the statistics
10 observations from five original publications. Populations, methods and periods differ. The 2026/27 edition label does not imply that every study measures 2026 or that 2027 observations exist.
10 records
Traditional-result click rate with AI summary
8% 8% — Traditional-result click rate with AI summary. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 8
- Unit
- percent
- Denominator
- Visits with AI summary
- Collection Period
- 2025-03-01 to 2025-03-31; results reconstructed April 7–17
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- pew-ai-click
- Topic
- Search clicks
- Metric
- Traditional-result click rate with AI summary
- Numerator
- Visits followed by traditional-result click
- Sample
- 900 adults; 68,879 unique searches
- Geography
- United States
- Device
- Tracked devices; no device split reported
- Publication Date
- 2025-07-22
- Publisher
- Pew Research Center
- Method
- Panel browsing URLs matched to subsequently retrieved results
- Limitations
- Observational; April results may differ from March; overlapping result links count as AI clicks.
- Superseding Edition
- Not established
- Citation Summary
- Pew: 8% of AI-summary visits led to a traditional-result click (US panel, March 2025).
Traditional-result click rate without AI summary
15% 15% — Traditional-result click rate without AI summary. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 15
- Unit
- percent
- Denominator
- Visits without AI summary
- Collection Period
- 2025-03-01 to 2025-03-31; results reconstructed April 7–17
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- pew-no-ai-click
- Topic
- Search clicks
- Metric
- Traditional-result click rate without AI summary
- Numerator
- Visits followed by result click
- Sample
- 900 adults; 68,879 unique searches
- Geography
- United States
- Device
- Tracked devices; no device split reported
- Publication Date
- 2025-07-22
- Publisher
- Pew Research Center
- Method
- Panel browsing URLs matched to subsequently retrieved results
- Limitations
- Observational; April results may differ from March; overlapping result links count as AI clicks.
- Superseding Edition
- Not established
- Citation Summary
- Pew: 15% of visits without AI summaries led to a result click (US panel, March 2025).
Pages with zero estimated monthly Google organic visits
96.55% 96.55% — Pages with zero estimated monthly Google organic visits. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 96.55
- Unit
- percent
- Denominator
- Sampled Content Explorer pages
- Collection Period
- 2023 study snapshot; exact extraction dates unreported
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- ahrefs-zero
- Topic
- Organic traffic
- Metric
- Pages with zero estimated monthly Google organic visits
- Numerator
- Pages with zero estimated visits
- Sample
- Approximately 14 billion Content Explorer pages
- Geography
- Multi-country index; distribution unreported
- Device
- Not device-specific
- Publication Date
- 2023-12-01
- Publisher
- Ahrefs
- Method
- Vendor keyword-based monthly organic traffic estimates
- Limitations
- Selected index, not the whole web; rare queries may be missing.
- Superseding Edition
- Not established
- Citation Summary
- Ahrefs: 96.55% of sampled pages had zero estimated Google traffic in its 2023 study.
Pages with 1–10 estimated monthly Google organic visits
1.94% 1.94% — Pages with 1–10 estimated monthly Google organic visits. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 1.94
- Unit
- percent
- Denominator
- Sampled Content Explorer pages
- Collection Period
- 2023 study snapshot; exact extraction dates unreported
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- ahrefs-low
- Topic
- Organic traffic
- Metric
- Pages with 1–10 estimated monthly Google organic visits
- Numerator
- Pages with 1–10 estimated visits
- Sample
- Approximately 14 billion Content Explorer pages
- Geography
- Multi-country index; distribution unreported
- Device
- Not device-specific
- Publication Date
- 2023-12-01
- Publisher
- Ahrefs
- Method
- Vendor keyword-based monthly organic traffic estimates
- Limitations
- Selected index, not the whole web; rare queries may be missing.
- Superseding Edition
- Not established
- Citation Summary
- Ahrefs: 1.94% of sampled pages had 1–10 estimated monthly visits in its 2023 study.
Googlebot share of HTML requests
4.5% 4.5% — Googlebot share of HTML requests. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 4.5
- Unit
- percent
- Denominator
- Observed HTML requests
- Collection Period
- 2025-01-01 to 2025-12-02
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- cf-googlebot
- Topic
- Crawling
- Metric
- Googlebot share of HTML requests
- Numerator
- Googlebot HTML requests
- Sample
- Customer HTML requests; exact count unreported
- Geography
- Global Cloudflare customer base
- Device
- All observed request clients
- Publication Date
- 2025-12-15
- Publisher
- Cloudflare
- Method
- Network request classification; average HTML request shares
- Limitations
- Cloudflare coverage only; HTML excludes other resource types; Googlebot separated.
- Superseding Edition
- Not established
- Citation Summary
- Cloudflare: Googlebot generated 4.5% of observed HTML requests, January 1–December 2, 2025.
Other AI bots share of HTML requests
4.2% 4.2% — Other AI bots share of HTML requests. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 4.2
- Unit
- percent
- Denominator
- Observed HTML requests
- Collection Period
- 2025-01-01 to 2025-12-02
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- cf-ai
- Topic
- Crawling
- Metric
- Other AI bots share of HTML requests
- Numerator
- AI-bot HTML requests excluding Googlebot
- Sample
- Customer HTML requests; exact count unreported
- Geography
- Global Cloudflare customer base
- Device
- All observed request clients
- Publication Date
- 2025-12-15
- Publisher
- Cloudflare
- Method
- Network request classification; average HTML request shares
- Limitations
- Cloudflare coverage only; HTML excludes other resource types; Googlebot separated.
- Superseding Edition
- Not established
- Citation Summary
- Cloudflare: other AI bots averaged 4.2% of HTML requests in its 2025 observation window.
Generative-AI referral growth to retail sites
- Value
- 1200
- Unit
- percent increase
- Denominator
- July 2024 AI-referred retail visits
- Collection Period
- 2025-02 compared with 2024-07
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- adobe-retail
- Topic
- AI referrals
- Metric
- Generative-AI referral growth to retail sites
- Numerator
- February minus July AI-referred retail visits
- Sample
- Retail research: over 1 trillion visits; travel count unreported
- Geography
- United States
- Device
- No device split reported
- Publication Date
- 2025-03-17
- Publisher
- Adobe Analytics
- Method
- First-party analytics aggregates of generative-AI-referred visits
- Limitations
- Selected customer sites; absolute referral counts not disclosed; historical comparison.
- Superseding Edition
- Not established
- Citation Summary
- Adobe: AI-referred US retail traffic grew 1,200% from July 2024 to February 2025.
Generative-AI referral growth to travel/leisure/hospitality sites
- Value
- 1700
- Unit
- percent increase
- Denominator
- July 2024 AI-referred travel visits
- Collection Period
- 2025-02 compared with 2024-07
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- adobe-travel
- Topic
- AI referrals
- Metric
- Generative-AI referral growth to travel/leisure/hospitality sites
- Numerator
- February minus July AI-referred travel visits
- Sample
- Retail research: over 1 trillion visits; travel count unreported
- Geography
- United States
- Device
- No device split reported
- Publication Date
- 2025-03-17
- Publisher
- Adobe Analytics
- Method
- First-party analytics aggregates of generative-AI-referred visits
- Limitations
- Selected customer sites; absolute referral counts not disclosed; historical comparison.
- Superseding Edition
- Not established
- Citation Summary
- Adobe: AI-referred US travel traffic grew 1,700% from July 2024 to February 2025.
Google searches ending without a click
68.01% 68.01% — Google searches ending without a click. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 68.01
- Unit
- percent
- Denominator
- Weighted observed Google searches
- Collection Period
- 2026-01 to 2026-04
- Source Url
- https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- spark-zero
- Topic
- Search clicks
- Metric
- Google searches ending without a click
- Numerator
- Searches without a click
- Sample
- Desktop/mobile clickstream panel; exact count unreported in article text
- Geography
- United States
- Device
- Browser desktop and mobile; Google search app excluded
- Publication Date
- 2026-06-08
- Publisher
- SparkToro / Similarweb
- Method
- Panel; two-thirds mobile weighting; 10-second mobile inactivity cutoff
- Limitations
- Weighted estimate; older editions use different panels; no causal inference.
- Superseding Edition
- Not established
- Citation Summary
- SparkToro/Similarweb: 68.01% of US browser Google searches were zero-click, January–April 2026.
Google searches proceeding to AI Mode
0.34% 0.34% — Google searches proceeding to AI Mode. Read the population, period and limitations below; this is not a cross-study comparison. - Value
- 0.34
- Unit
- percent
- Denominator
- Observed Google searches
- Collection Period
- 2026-01 to 2026-04
- Source Url
- https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/
- Checked Date
- 2026-09-11
Full record, verification and limitations
- Id
- spark-ai-mode
- Topic
- Search clicks
- Metric
- Google searches proceeding to AI Mode
- Numerator
- Searches proceeding to AI Mode
- Sample
- Desktop/mobile clickstream panel; exact count unreported in article text
- Geography
- United States
- Device
- Browser desktop and mobile; Google search app excluded
- Publication Date
- 2026-06-08
- Publisher
- SparkToro / Similarweb
- Method
- Panel; two-thirds mobile weighting; 10-second mobile inactivity cutoff
- Limitations
- Weighted estimate; older editions use different panels; no causal inference.
- Superseding Edition
- Not established
- Citation Summary
- SparkToro/Similarweb: 0.34% of searches proceeded to AI Mode, January–April 2026.
No matching records. Try a broader search or reset the filters.
Dataset change log
- September 11, 2026 — initial sourced dataset assembled for editorial review; not published.
Which SEO statistics belong in a benchmark?
A useful SEO benchmark measures a defined outcome in a population that resembles the audience or system you are evaluating. A large sample is helpful, but it cannot repair a mismatched denominator.
- Search behavior measures what people do after a query.
- Organic traffic estimates describe modeled visits to pages.
- Crawl shares describe requests made by automated clients.
- Referral growth compares one traffic source across periods.
The dataset keeps those families separate. It includes original studies and platform reports, not statistics copied from marketing roundups. This is a selected reference collection, not a complete census of SEO performance.
How often do Google searches end without a click?
SparkToro’s analysis of Similarweb data reports a 68.01% zero-click share for US browser searches during January–April 2026. Its original study also reports that 0.34% of searches proceeded to AI Mode.
- The panel covers desktop and mobile browsers.
- The Google mobile search application is excluded.
- Mobile sessions use a 10-second inactivity cutoff.
The researchers weight mobile searches at two-thirds of the total. Earlier editions used different panels, so a chart joining them is not a controlled longitudinal experiment. A zero-click search also does not identify whether the searcher achieved their goal.
Do AI summaries change search-result clicks?
Pew observed traditional-result clicks on 8% of visits with an AI summary and 15% without one. Its US browsing-panel analysis covers March 2025 activity from 900 adults.
- Search results were reconstructed in April.
- Query mix can differ between the two groups.
- The comparison does not isolate a causal effect.
These are search-visit outcomes, not the click-through rate of an individual publisher. Keep that distinction when discussing the findings with editors or setting expectations for a new resource.
How much content receives estimated organic traffic?
Ahrefs classified 96.55% of its sampled pages as receiving zero estimated monthly Google traffic, with 1.94% receiving 1–10 visits. The 2023 Content Explorer study examined approximately 14 billion pages.
- The sample is a selected index, not the entire web.
- Traffic is estimated from the vendor’s keyword data.
- Rare queries may produce visits the model misses.
A zero estimate does not prove that nobody visited a page. For your own publication, use actual reporting alongside estimates and document the differences in your measurement baseline.
How much crawling comes from AI bots?
Cloudflare reports Googlebot at 4.5% of observed HTML requests and other AI bots at an average 4.2%. Its 2025 network review covers January 1 through December 2.
- The denominator includes HTML requests only.
- Googlebot is separated from the other AI-bot group.
- The observations reflect Cloudflare’s customer network.
A request is not a reader, referral or conversion. Use crawl evidence to investigate delivery and access decisions; use human traffic reporting to judge audience response. Review the difference between crawling and indexing controls before changing access rules.
How quickly are AI referrals growing?
Adobe reported AI-referred US retail traffic growth of 1,200% and travel traffic growth of 1,700% between July 2024 and February 2025. The original analytics report describes those historical comparisons.
- Growth is measured against the earlier referral volume.
- Growth percentage is not share of total traffic.
- Neither figure supplies a publisher conversion benchmark.
Large relative growth can begin from a small base. Preserve the baseline month when citing either number, and do not substitute a later period merely because the resource has a newer edition date.
How should publishers compare these benchmarks?
Publishers should compare these benchmarks by matching the measured event, population and period to their own reporting. Use the five checks below to decide whether an external figure is a useful reference for your publication.
- Name the decision: content investment, crawl policy or referral measurement.
- Choose the relevant metric family in this dataset.
- Check geography, device coverage and observation dates.
- Record missing information before making a comparison.
- Test the resulting hypothesis against your own evidence.
Do not average the percentages across rows. They describe different events and use different denominators. Even two traffic studies can disagree because one estimates page visits while another observes browser journeys.
How do you cite an SEO statistic correctly?
Cite the original publisher and retain the metric’s scope. Each record includes a short citation summary you can adapt without removing its period or population.
- Link to the source that collected or analyzed the data.
- Name the measurement period, not just the publication year.
- Carry material limitations into the surrounding explanation.
Use the downloadable dataset to keep these fields together. If a source changes its methodology, treat the revised series as a new comparison until you establish that the definitions still match.
How is this statistics resource maintained?
This statistics resource is maintained through quarterly source checks and documented corrections by Approve SEO Research Desk. Maintenance covers the original links, metric definitions, observation periods and any replacement editions. The checked date records that review, not a new research measurement.
Change log: the September 11, 2026 first edition adds 10 observations from five source publications. Later evidence will be added with its own period and supersession note. Unavailable fields remain explicitly unreported.
Are these all 2026 or 2027 statistics?
These are not all 2026 or 2027 statistics: this edition includes source observations from 2023 through 2026. The dataset has no 2027 observations; the edition label describes the resource rather than the age of every statistic.
Can these figures predict my website traffic?
These figures cannot predict your website traffic. They describe sampled pages, search journeys and platform activity, while your traffic also reflects your audience, query mix and publishing history. Use them to frame a question, then measure your own results.
Why are some sample details marked unknown?
Some sample details are marked unknown because the original publisher does not disclose the count or device breakdown. We preserve that reporting gap instead of estimating an undocumented sample size.
Build a comparable measurement record
Choose one relevant observation and write down what would make your own results comparable. Then use our publication analytics guide to connect the question to a measurement method you can maintain.