Google ranking-factor research points to a complicated picture: links still correlate with visibility, but content quality, technical experience, and search-result enhancements describe different parts of performance. The figures below come from large-scale studies and official guidance, with each measurement period kept explicit.
Contents
- How links correlate with rankings
- What content and on-page signals show
- What page speed data shows
- What Core Web Vitals data shows
- How structured data affects search performance
- The official Core Web Vitals thresholds
How links correlate with rankings
The clearest recent link statistics come from Ahrefs’ January 2025 study of the top 20 results for 1,000,000 high-volume US keywords. Its figures are correlations with Google ranking position, not proof that a specific link action causes a page to move. Even so, they show that link-related measures remain measurable signals in a very large sample.
In the Ahrefs analysis, Domain Rating had a Spearman correlation of 0.131 with ranking position. The number of backlinks had a stronger 0.248 correlation, while followed backlinks reached 0.242. Referring domains produced the highest result among the listed link measures at 0.255, and followed referring domains registered 0.250. The study also found weaker internal-link relationships: internal incoming links had a 0.117 correlation, internal outgoing links 0.093, and external outgoing links 0.083. These values describe association in the January 2025 dataset; they do not establish an independent ranking formula.
| Link or URL measure | Spearman correlation with ranking position | Study period |
|---|---|---|
| Referring domains | 0.255 | Ahrefs, January 2025 |
| Followed referring domains | 0.250 | Ahrefs, January 2025 |
| Backlinks | 0.248 | Ahrefs, January 2025 |
| Followed backlinks | 0.242 | Ahrefs, January 2025 |
| Domain Rating | 0.131 | Ahrefs, January 2025 |
| Internal incoming links | 0.117 | Ahrefs, January 2025 |
Backlinko’s 11.8-million-result study, reported in 2019 and updated in April 2025, provides a different way to view link prominence. The number-one Google result had an average of 3.8 times as many backlinks as positions two through ten. At the same time, approximately 95% of pages in the study’s results dataset had zero backlinks. A separate Backlinko-BuzzSumo analysis cited by the study found that 94% of content had zero backlinks.
The distribution is important: a first-page average can coexist with many pages that have no backlinks. Backlinko also found a median Ahrefs URL Rating of 12 for pages ranked in Google’s top six, compared with 11 for positions seven through ten. The average URL Rating for a first-page result was 11.2. Those are observational statistics from the study, not targets that guarantee a ranking.
What content and on-page signals show
Content length is one of the most frequently repeated SEO comparisons, but the available numbers do not support a simple “longer is better” rule. Backlinko measured a mean word count of 1,447 words for a Google first-page result. However, the analysis found no direct relationship between word count and ranking position among first-page results.
A content-quality measure showed a more useful relationship in the study’s top-30 subset. Increasing the Clearscope Content Grade by one point approximated a one-position improvement in Google ranking. This is an approximation within that analysis, not a universal conversion rate between a content score and a search position.
Keyword placement was common but not strongly differentiating at the top of the results. Google top-10 pages contained roughly 65% to 85% of the ranking keyword in their title tags. Backlinko’s model estimated only a 1% difference in title-keyword matching between the number-one and number-ten results. Results pages contained approximately 60% to 80% of the ranking keyword in their H1 tags.
URL structure also showed modest relationships. In Ahrefs’ January 2025 analysis, having the target keyword in the URL had a 0.034 Spearman correlation with ranking position. URL length had a -0.107 correlation. In Backlinko’s study, position-one URLs were an average of 9.2 characters shorter than position-ten URLs, the average URL length for a top-10 result was 66 characters, and most first-page URLs fell between 40 and 100 characters.
Taken together, these figures support a practical distinction. Relevance signals such as titles and headings are widespread among top results, so their presence alone may not separate first place from tenth. Content quality and usefulness can be associated with stronger positions, while word count by itself is not a ranking recipe. The study periods matter: Backlinko’s underlying results were reported in 2019, with the page updated in April 2025, while the URL and link correlations above come from Ahrefs’ January 2025 analysis.
What page speed data shows
Page speed statistics need to be separated into ranking correlation and user behavior. Backlinko found zero correlation between domain-level page speed and Google first-page rankings in its 1-million-domain speed subset. The same analysis measured an average loading speed of 1.65 seconds for a Google first-page result. These figures describe the observed relationship in that study; they do not mean speed is irrelevant to users or to every search system.
Google’s Mobile Site Speed Playbook reported an average full load time of 22 seconds for 900,000 mobile sites tested globally in research reported in 2017. Google also reported that 53% of people leave a mobile site when it takes more than three seconds to load. A SOASTA study cited in the same Google playbook found that mobile pages loading one second faster saw up to a 27% increase in conversion rates.
The Google-cited mobile research reported an average mobile bounce rate of 13% for loads under three seconds, with bounce rates approaching 60% after load times exceeded nine seconds. These are user-experience and conversion measurements from the cited research, not direct Google ranking coefficients.
The playbook also reviewed 316 European travel, retail, and finance sites. The fastest top retail sites averaged 3.8 seconds to load, and only 14 of the 316 sites loaded in two seconds or less. Two case studies illustrate how performance changes can accompany business changes. Money.co.uk reduced load time from 11.7 seconds to 3.5 seconds after a Google audit; its on-site click-through rate rose from 12% to 32%, while bounce rate fell from 80% to 52%. Kempinski Hotels reported an 11% conversion-rate increase after using Google’s speed and usability recommendations. The playbook reports these outcomes from 2017 research and case studies, so they should not be read as current forecasts.
What Core Web Vitals data shows
The 2024 Web Almanac SEO chapter provides a more recent view of field performance. It found that 48% of mobile sites passed the combined Core Web Vitals assessment. On mobile, 59% had good Largest Contentful Paint (LCP), 74% had good Interaction to Next Paint (INP), and 79% had good Cumulative Layout Shift (CLS).
Desktop results were stronger for the combined assessment: 54% of desktop sites passed. The desktop metric results were 72% good LCP, 97% good INP, and 71% good CLS. The gap between individual metric rates and the combined pass rate shows why a site can perform well on one measure while failing the overall assessment.
These Web Almanac percentages are population measurements for 2024, not forecasts of an individual site’s ranking. They are useful as a benchmark for technical SEO work: mobile combined passage at 48% means a pass is achieved by less than half of the measured mobile sites, while desktop passage at 54% is still not universal.
How structured data affects search performance
Structured data has two different questions attached to it: whether pages rank higher, and whether a qualifying page earns a more visible search presentation. Backlinko found structured data on 72.6% of first-page Google pages in its 11.8-million-result study, but found no relationship between structured-data presence and ranking position. That result does not negate potential click or interaction benefits from enhanced search features.
Google’s structured-data case studies report several performance comparisons. Rotten Tomatoes saw a 25% higher click-through rate on 100,000 pages enhanced with structured data. Food Network converted 80% of its pages to enable search features and reported a 35% increase in visits. Rakuten reported that users spent 1.5 times more time on pages with structured data than on pages without it.
Rakuten also reported a 3.6-times higher interaction rate on AMP pages with search features than on non-feature AMP pages. Nestlé measured an 82% higher click-through rate for pages appearing as rich results than for its non-rich-result pages. Google does not state a case-study period for these examples on its structured data introduction, so the comparisons should be treated as reported case-study outcomes rather than a dated industry-wide estimate.
The distinction is central for SEO planning. The available ranking study did not find a ranking-position relationship, while Google’s case studies measured higher clicks, visits, time, or interactions when search features were present. Structured data can therefore be evaluated as a search-result eligibility and presentation tactic, with the reported outcomes understood as organization-specific measurements.
The official Core Web Vitals thresholds
Google’s Core Web Vitals threshold guidance defines “good” at the 75th percentile for each metric. The good LCP target is 2,500 milliseconds or less, the good INP target is 200 milliseconds or less, and the good CLS target is 0.1 or less. Google classifies a page or site as good for a Core Web Vital when at least 75% of page views meet that metric’s good threshold.
| Metric | Good threshold at the 75th percentile |
|---|---|
| Largest Contentful Paint (LCP) | 2,500 milliseconds or less |
| Interaction to Next Paint (INP) | 200 milliseconds or less |
| Cumulative Layout Shift (CLS) | 0.1 or less |
The thresholds are measurement standards, not promises of a particular Google position. The 2024 Web Almanac results show the operational challenge: 48% of mobile sites and 54% of desktop sites passed all three together, despite higher pass rates for several individual metrics. For a site team, the most defensible use of these statistics is to measure real page experience against the official thresholds, then interpret ranking, click, and conversion evidence separately.