This page collects internal linking statistics from primary sources only: crawl-scale measurements, Google's and Bing's own documentation, anchor text audits, orphan and broken link data, and the few ranking tests that were run with a control group. Every number links to its original publisher. Cite freely with a link.
Internal linking statistics at a glance
| Category | Statistic | Source |
|---|---|---|
| Link counts | The median inner page has 43 internal links on desktop and 39 on mobile | HTTP Archive Web Almanac, 2025 |
| Link counts | The top 1,000 sites' median mobile home page has 129 internal links; all sites, 41 (3.1x gap) | HTTP Archive Web Almanac, 2024 (computed) |
| Link counts | The median desktop inner page links internally 7.2 times as often as externally (43 vs 6) | HTTP Archive Web Almanac, 2025 (computed) |
| Qualifiers | nofollow appears on 32.3% of pages; sponsored and ugc on 0.5% each | HTTP Archive Web Almanac, 2025 |
| Anchor text | 46.3% of the top million home pages have at least one empty link | WebAIM Million, 2026 |
| Anchor text | 15.2% of home pages use ambiguous link text such as "click here" or "more" | WebAIM Million, 2026 |
| "Every page you care about should have a link from at least one other page on your site" | Google Search Central, 2025 | |
| Orphans | 57.3% of 100,000+ audited sites had a critical link failure in January to July 2026 | SEOmator, 2026 |
| Orphans | 0.72% of English Wikipedia's 7.2 million articles carry the orphan tag | Wikipedia, September 2026 (computed) |
| Broken links | 23% of news webpages and 21% of government webpages contain at least one broken link | Pew Research Center, 2024 |
| Ranking impact | Pages with at least one exact-match internal anchor had at least 5x more traffic (correlation) | Zyppy, 2022 |
| Ranking impact | Links to six nearby regional pages lifted organic traffic 7% in a controlled split test | SearchPilot, 2021 |
| Ranking impact | 80% of 15 internal link tests on one site were positive | SEOTesting via Screaming Frog, 2025 |
How many internal links does the average web page have?
The median inner page had 43 internal links on desktop and 39 on mobile in July 2025, according to the HTTP Archive Web Almanac, and the median home page carried more: 51 on desktop and 46 on mobile.
The Web Almanac is the largest public dataset that counts links on millions of real pages. HTTP Archive crawls every site in the Chrome UX Report once a month and publishes the queries, so anyone can rerun them. The 2025 edition uses the July 2025 crawl: 16,213,084 websites, of which 15.4 million were tested on mobile and 12.2 million on desktop (methodology). The 2024 and 2025 SEO chapters report home pages and one inner page per site separately; the 2021 and 2022 chapters used home pages only. The percentile table below is inner pages, per the Almanac's published data sheet.
| Percentile | Internal links, desktop inner pages | Internal links, mobile inner pages | Source |
|---|---|---|---|
| 10th | 7 | 6 | Web Almanac, 2025 |
| 25th | 19 | 17 | Web Almanac, 2025 |
| 50th (median) | 43 | 39 | Web Almanac, 2025 |
| 75th | 89 | 82 | Web Almanac, 2025 |
| 90th | 174 | 161 | Web Almanac, 2025 |
| Statistic | Source |
|---|---|
| Home pages in the July 2025 crawl carried a median of 51 internal links on desktop and 46 on mobile; at the 90th percentile, 213 and 197. | Web Almanac, 2025 |
| In the June 2024 crawl of 16.9 million websites, the median home page had 46 internal links on desktop and 41 on mobile; inner pages had 39 and 36. | Web Almanac, 2024 |
| In the June 2022 crawl of 8.36 million websites, where the SEO chapter used home pages only, the median desktop home page had 56 internal links and the median mobile home page 48. DATED. | Web Almanac, 2022 |
| In the 2021 edition, which covered home pages only, the median desktop home page had 64 internal links against 55 on mobile, a 16% gap. DATED. | Web Almanac, 2021 |
| Desktop home pages carried 11% more internal links than mobile home pages at the median in 2025 (51 vs 46), down from 16% to 17% in the 2021 and 2022 home-page data. Inner pages show a 10% gap (43 vs 39). | Computed from Web Almanac, 2021 to 2025 |
On desktop home pages, the all-sites median fell from 56 in 2022 to 46 in 2024 while the top 1,000 sites rose from 137 to 166. The page type is the same in both years, but the crawl doubled from 8.36 million to 16.9 million websites, and a bigger crawl adds mostly small sites, which pulls the all-sites median down. The Almanac authors describe internal linking as growing, citing the top 1,000 sites, and at the top of the ranking the data agrees.
The desktop and mobile gap has narrowed on a like-for-like sample. On home pages, desktop carried 16% more internal links than mobile in 2021, 17% in 2022, 12% in 2024 and 11% in 2025. The site mix changed as the crawl grew, so treat the size of the drop as approximate. The Almanac attributes the difference to developers trimming menus and footers on small screens. Google indexes the mobile version of a page first, so any link that exists only on desktop is a link Google may never weigh.
If your template pages sit near 40 internal links, you are normal, and if a money page has six, it sits in the bottom tenth of the web. The mobile gap is the part I would check first on any audit. A mega-menu that collapses into three links on a phone quietly removes dozens of internal links from the version Google crawls.
A median hides how differently big and small sites behave, which is where the next table comes in.
Do popular sites use more internal links than small sites?
Popular sites use far more internal links: the median mobile home page on the top 1,000 sites carried 129 internal links in June 2024, 3.1 times the 41 on the median mobile home page across all sites, according to the Web Almanac 2024.
The Almanac groups sites by Chrome UX Report rank magnitude, from the top 1,000 by page views down to the top 10 million. In 2024 the link count falls at every step down the ranking, on desktop and on mobile.
| Rank group (home pages) | Desktop, 2024 | Mobile, 2024 | Desktop, 2022 | Mobile, 2022 | Source |
|---|---|---|---|---|---|
| Top 1,000 sites | 166 | 129 | 137 | 106 | Web Almanac 2024; 2022 |
| Top 10,000 | 150 | 122 | 139 | 117 | Web Almanac 2024; 2022 |
| Top 100,000 | 105 | 86 | 105 | 94 | Web Almanac 2024; 2022 |
| Top 1 million | 82 | 71 | 88 | 82 | Web Almanac 2024; 2022 |
| Top 10 million | 58 | 52 | 63 | 56 | Web Almanac 2024; 2022 |
| All sites | 46 | 41 | 56 | 48 | Web Almanac 2024; 2022 |
A note on the 2024 figures: the chapter's figure description lists the same six values, 129 down to 41, for desktop and mobile. The Almanac's published data sheet shows those six values are the mobile home-page medians, and the chapter text also calls 129 the mobile figure. The desktop values above come from the same data sheet, and on desktop the ratio is 166 / 46, or 3.6x.
Both years measure home pages, but the 2024 crawl covered twice as many websites, so read the columns for shape rather than growth. In both years the curve is steep at the top and flat at the bottom. On desktop, the top 10,000 sites carry more than twice the internal links of the top 10 million in 2022 (139 vs 63) and in 2024 (150 vs 58).
The 3.1x ratio is the most useful internal linking number on this page, and it is also the easiest to misuse. Big sites have more links because they have more to link to: more categories, more products, bigger mega-menus. Nobody gets into the top 1,000 by adding footer links. What the ratio does tell you is that the sites Google sends most traffic to are densely connected, so a thinly linked site is competing with one arm tied back.
Link density is one axis. Where the links point, inward or outward, is the next.
How do internal and external link counts compare?
The median inner page links to its own site about 7.2 times as often as to other sites: 43 internal links against 6 external links on desktop in the Web Almanac 2025, and 39 against 6 on mobile. Home pages show a similar ratio, 51 against 7 on desktop.
| Statistic | Source |
|---|---|
| The median inner page had 6 external links on both desktop and mobile in 2025; the 90th percentile had 25 on desktop and 24 on mobile. | Web Almanac, 2025 |
| The median desktop inner page links internally 7.2 times as often as externally (43 / 6); on mobile the ratio is 6.5. | Computed from Web Almanac, 2025 |
| The rel="nofollow" attribute appeared on 32.3% of pages in 2025, nearly the same as 2024. | Web Almanac, 2025 |
| The more specific rel="sponsored" and rel="ugc" values appeared on 0.5% of pages each in 2025. | Web Almanac, 2025 |
| The security values noopener and noreferrer appeared on 40.9% and 25% of pages respectively. | Web Almanac, 2025 |
Google added sponsored and ugc so that publishers could say why a link was qualified, yet only 0.5% of pages use either one. The web settled on two states, followed and nofollowed, and Google's own documentation now says to use nofollow "only when you don't trust the source".
Here is the practical point for internal links. Every one of those 32% of pages with a nofollow is a chance that someone nofollowed their own login, cart or filter links to "save PageRank", a trick Google stopped rewarding years ago. Internal links should almost always be plain followed links. If a page shouldn't be crawled, block it or noindex it; don't sabotage the path to it.
Counts and qualifiers describe the web as it is. Google's documentation describes what Google wants it to be.
What does Google say about internal links and anchor text?
Google says it uses links "as a signal when determining the relevancy of pages and to find new pages to crawl", and its Search Central documentation, updated December 2025, asks that every important page get at least one link from another page on the same site.
| Statistic | Source |
|---|---|
| "Every page you care about should have a link from at least one other page on your site." | Google Search Central, link best practices, updated December 2025 |
| Google "can only crawl your link if it's an `<a>` HTML element" with an href attribute. | Google Search Central, December 2025 |
| "Good anchor text is descriptive, reasonably concise, and relevant to the page"; Google discourages generic "click here" and "read more". | Google Search Central, December 2025 |
| A site of "about 500 pages or fewer" that is "comprehensively linked internally" may not need a sitemap at all. | Google Search Central, sitemaps overview, December 2025 |
| Google's crawl budget guide is written for sites with 1 million or more unique pages changing weekly, or 10,000 or more changing daily. | Google Search Central, crawl budget, updated July 2026 |
| Bing asks site owners to "ensure each important URL is reachable through crawlable internal links" with relevant anchor text, and says following its guidelines keeps URLs eligible for Copilot and grounding results. | Microsoft Bing Webmaster Guidelines, 2026 |
| In the original 1998 Google paper, a crawl of 24 million pages produced over 259 million indexed anchors, and PageRank's damping factor was usually set to 0.85. DATED. | Brin and Page, Stanford University, 1998 |
The 1998 paper is here for one reason. Google has indexed anchor text since its first crawl, and in 2025 it still tells site owners to write descriptive anchors. Twenty-seven years of the same advice is a signal in itself.
The crawl budget guide matters for the opposite reason. Google's thresholds, 1 million pages or 10,000 fast-changing pages, mean that crawl budget is not a problem for most sites. For a 300-page B2B site, the real constraint is whether internal links tell Google which of those 300 pages matter.
The 500-page line in the sitemap guide is the most underrated sentence Google has published on internal linking. It says that for most small sites, internal links are the discovery system and the sitemap is a backup. Bing's current guidelines go further and tie crawlable internal links to eligibility for Copilot and grounding results, which I read as: if a page can't be reached by a link, it doesn't exist for either engine's answers.
Google asks for descriptive anchors. The next section measures how many sites deliver them.
How descriptive is anchor text on the web?
Most anchor text on the web passes a basic descriptiveness test, but 46.3% of the top million home pages still contain at least one empty link, according to WebAIM's February 2026 audit.
| Statistic | Source |
|---|---|
| 46.3% of the top one million home pages had at least one empty link (a link with no text a screen reader can announce) in February 2026, up from 45.4% in 2025. | WebAIM Million, 2026 |
| 15.2% of home pages had ambiguous link text such as "click here", "more" or "continue", up from 13.7% in 2025 but down from 23.2% in 2020. Affected pages averaged 5.3 ambiguous links. | WebAIM Million, 2026 |
| In the Web Almanac 2025, 84.64% of desktop home pages and 86.64% of mobile home pages passed Lighthouse's descriptive link text test. | Web Almanac, 2025 |
| Inner pages passed more often: 92.42% on desktop and 93.45% on mobile in 2025, against 86% and 92% in 2024. | Web Almanac, 2025 |
| When scanning a list of links, users see "about 2 words" of each item, in a Nielsen Norman Group test of 80 participants and 20 real links. DATED. | Nielsen Norman Group, April 2009 |
The Nielsen Norman Group finding is from 2009 and I keep it because nothing has replaced it. If readers see about two words of a link, the first two words of an anchor carry most of its meaning for people. Google reads the whole anchor, but a link that starts with "click" or "learn more" wastes the words that humans scan.
Two audits, two answers. Lighthouse only flags a short list of generic phrases, so a pass rate of 85% to 93% says little about whether the anchor names the target. WebAIM checks the accessible name of each link, and nearly half of home pages fail somewhere, usually on icon links, logo links and social buttons with no text.
Empty links matter more for internal linking than anyone admits. An icon link to your pricing page with no text sends Google zero words about pricing. Home pages are where this happens most, and home pages hold the strongest internal links on the site. Fix the empty links in your header and footer before you rewrite a single blog anchor.
Anchor quality decides what a link says. Orphan pages are the pages no link points to at all.
How common are orphan pages and uncrawled pages?
Orphan pages are common and costly: SEOmator found that sites failing its crawlability checks carried a median 8% of their pages as orphans in its January to July 2026 audits of more than 100,000 sites.
An orphan page is a URL that exists (in a sitemap, in analytics, in server logs) but that no internal link points to. Vendors measure orphans on their own customers' crawls, so treat every number below as a sample of that vendor's users, not of the web.
| Statistic | Source |
|---|---|
| 57.3% of more than 100,000 sites audited between January and July 2026 had at least one critical failure in the Links category. The audits crawled 50 million+ pages. Vendor data. | SEOmator, 2026 |
| Sites that failed on crawlability carried a median 8% of their pages as orphans. Vendor data. | SEOmator, 2026 |
| On one large site analysed by Botify, more than 70% of the pages Googlebot explored were orphans: about 800,000 orphan pages against 300,000 in the site structure. DATED, single-site case. | Botify, May 2020 |
| On the same site, 95% of organic visits came from pages inside the structure, which were only 30% of the site's pages. Orphans produced 5% of visits. DATED. | Botify, May 2020 |
| English Wikipedia lists 52,447 articles in its "All orphaned articles" category, articles editors have tagged as having no incoming links from other articles, as of 24 September 2026. | Wikipedia, September 2026 |
| English Wikipedia had 7,244,074 articles on the same date, so 0.72% carry the orphan tag. | Computed from Wikipedia Special:Statistics |
Wikipedia is the best-linked large site on the web, with thousands of volunteers whose hobby is adding links, and it still has more than 52,000 tagged orphans. The tag is applied by editors, so 0.72% is a floor, not a census. SEOmator's 8% median describes a different population, sites that failed a vendor audit, so I show it as a separate fact rather than a ratio against Wikipedia.
Botify's 70/30 split is one site and six years old, and I still quote it, because every log file I have looked at since tells the same story in smaller numbers. Google keeps crawling orphan URLs it found once, while the pages that earn the traffic sit inside the navigation. The fix is boring: export orphans from your crawler, link the ones that deserve to live and redirect or remove the rest.
Google's documentation explains why orphans persist. A sitemap lets Google discover a URL without any internal link, so an orphan can be crawled and even indexed. What an orphan lacks is the context an internal link provides: the anchor text, the position in the site's structure and the signal that the site itself considers the page worth linking. Google's sitemap guide treats a site that is "comprehensively linked internally" as the normal case, and a sitemap as the fallback.
Orphans are pages nobody links to. Broken links are links that point nowhere.
How many internal links are broken?
Broken internal links hit 36.1% of the sites SEOmator audited in the first half of 2026, with a median of 14 broken internal links per affected site, and broken links in general are common across the web.
| Statistic | Source |
|---|---|
| Broken internal links were found on 36.1% of audited sites, with a median of 14 per site, in January to July 2026. Vendor data. | SEOmator, 2026 |
| 23% of news webpages contained at least one broken link, and 5% of all links on news pages pointed to inaccessible pages, across 500,000 pages from 2,063 news sites and more than 14 million links. DATED. | Pew Research Center, May 2024 |
| 21% of government webpages contained at least one broken link, and 6% of 42 million links on those pages were broken. DATED. | Pew Research Center, May 2024 |
| 54% of English Wikipedia pages had at least one broken reference link, and 11% of reference links were dead, in a sample of 50,000 pages. DATED. | Pew Research Center, May 2024 |
| At least 66.5% of links pointing to 2,062,173 websites since 2013 are dead. These are backlinks from other sites, not internal links. DATED. | Ahrefs, February 2024 |
| The Internet Archive's bot had redirected more than 9 million broken Wikipedia links to Wayback Machine copies across 22 language editions by October 2018. DATED. | Internet Archive, October 2018 |
Pew counted all links on each page, internal and external, using Common Crawl snapshots from March and April 2023, and it found no real difference between high-traffic and low-traffic news sites. Popularity does not protect a site from link rot. The Ahrefs figure measures a different thing, backlinks that other sites once pointed at yours and later lost, but it shows the same decay from the receiving end: two thirds of the links earned since 2013 no longer exist. Pages move, content decays and links nobody maintains break; I track the same effect for rankings on the content decay statistics page.
A broken internal link is the cheapest SEO fix that exists, because you own both ends of it. External link rot is somebody else's problem; internal link rot is a to-do list your crawler writes for you. Fourteen broken links per site sounds small until you notice they cluster on old blog posts that still carry backlinks.
Clean links and connected pages are hygiene. The question everyone really asks is whether internal links move rankings.
Do internal links improve rankings and traffic?
Internal links correlate strongly with Google traffic, and the few controlled tests show modest gains: Zyppy's 2022 study of 23 million internal links found URLs with 40 to 44 internal links got four times the clicks of URLs with 0 to 4, and SearchPilot split tests found lifts of 5% to 7%.
Keep correlation and cause apart here. Zyppy's study is correlation. SearchPilot's tests are split tests with a control group. The SEOTesting results are before and after measurements on one site.
| Statistic | Source |
|---|---|
| Zyppy analysed 23 million internal links across 1,800 websites and about 520,000 URLs, matched to Google Search Console clicks. Published July 2022, updated February 2026. DATED. | Zyppy, 2022 |
| URLs with 0 to 4 internal links averaged 2 clicks from Google Search; URLs with 40 to 44 internal links averaged 4 times that. Correlation. | Zyppy, 2022 |
| After a URL received about 45 to 50 internal links, Google traffic began to decline. Correlation. | Zyppy, 2022 |
| Pages with at least one exact-match internal anchor had at least five times more traffic than pages without one. Correlation. | Zyppy, 2022 |
| Pages with URL anchors from internal links saw almost 50% more traffic; empty anchor text made no measurable difference. Correlation. | Zyppy, 2022 |
| Adding links from each of about 8,000 US regional pages to its six nearest neighbours produced a 7% organic traffic uplift in a SearchPilot split test. DATED. | SearchPilot, January 2021 |
| Expanding a home page footer from 30 to more than 100 links lifted organic traffic to linked pages by 5% overall (below 95% confidence) and 10% on desktop (at 95% confidence). DATED. | SearchPilot, September 2021 |
| Of 15 internal link tests run on SEOTesting's own site, 80% were positive and 20% negative, with average changes of +173.5% in clicks, +264.5% in impressions and 4.87 positions of improvement. Before and after, no control group. | SEOTesting via Screaming Frog blog, May 2025 |
Zyppy's own caveat is worth repeating: "correlation isn't causation". Pages that already rank well attract internal links because editors link to what works. The exact-match finding is the strongest in the study and also the most likely to be reversed cause: you write exact-match anchors to pages you consider important, and those pages were getting traffic before the link.
The footer test deserves a closer read. SearchPilot expanded a footer from 30 links to more than 100 and got a significant 10% lift on desktop, but the overall result fell short of 95% confidence. The SearchPilot team added its own warning against packing footers with links. Both SearchPilot tests date from 2021, before AI Overviews changed click behaviour, so treat them as evidence of direction rather than a current benchmark.
If I had to bet a budget on one number here, it would be SearchPilot's 7%, not Zyppy's 5x. A controlled 5% to 10% lift is what internal linking realistically buys on an established site, and on 8,000 pages that is real money. The SEOTesting averages are inflated by small starting numbers; a page going from 10 clicks to 40 is +300%. For how internal work compares with other SEO spend, see the SEO ROI statistics.
The four numbers computed for this page are laid out below so you can check them.
How we calculated the original numbers
Four figures on this page do not appear in any source. Each one uses a single publisher's data, so no numbers from different populations are mixed.
- 3.1x: internal links on top 1,000 sites vs all sites. The Web Almanac 2024 (June 2024 crawl) gives the median mobile home page on the top 1,000 sites 129 internal links and the median across all sites 41. Both figures share an edition, client, page type and percentile; only the rank group differs. 129 / 41 = 3.15, rounded down to 3.1x. On desktop home pages the same data sheet gives 166 / 46 = 3.6x. The 2022 edition, also home pages, gives 106 / 48 = 2.2x on mobile, but the 2024 crawl covered twice as many sites, so I don't present the change as a trend.
- 7.2x: internal vs external links per page. The Web Almanac 2025 (July 2025 crawl) gives the median desktop inner page 43 internal and 6 external links. 43 / 6 = 7.17. On mobile inner pages, 39 / 6 = 6.5; on desktop home pages, 51 / 7 = 7.3. Both are medians taken separately, so the ratio describes the typical page, not a per-page average.
- 11%: the desktop over mobile gap. Home pages only, so every year uses the same page type: 2025 (51 − 46) / 46 = 10.9%; 2024 (46 − 41) / 41 = 12.2%; 2022 (56 − 48) / 48 = 16.7%; 2021 (64 − 55) / 55 = 16.4%. Inner pages in 2025: (43 − 39) / 39 = 10.3%. The crawl grew from 8.36 million sites in 2022 to 16.2 million in 2025, so part of the narrowing may come from the changing site mix.
- 0.72%: orphan share of English Wikipedia. On 24 September 2026 the MediaWiki API returned 52,447 pages in Category:All orphaned articles and 7,244,074 articles in total (rechecked at 20:46 UTC; the article count moves by a few an hour). 52,447 / 7,244,074 = 0.724%. The category depends on editors tagging orphans, so the true share is higher.
What do the 2026 internal linking numbers say?
Normal is 40 to 50 internal links per page, and the winners carry three times that. The Web Almanac's all-sites medians have held between 36 and 64 across four editions, on home and inner pages, desktop and mobile. The rank split is the stable part: every step up the traffic ranking adds links. A site with 10 links per template page is disconnected, whatever its designers call it.
Anchor text is where most sites still leave value on the table. Nearly half of top home pages have an empty link, and those links sit in the header and footer, the most repeated internal links a site has. Google has asked for descriptive anchors since 1998. Fixing icon links is a one-afternoon job that improves accessibility and internal linking at once.
Internal linking helps, by a modest and measurable amount. Controlled tests put the gain at 5% to 10% on the pages that receive links. That is smaller than the correlation studies imply and larger than doing nothing, and it compounds on large sites. The honest pitch to a CFO is 7%, not 5x. For more benchmarks on search behaviour, the full statistics hub collects every page in this series.
FAQ
How many internal links should a page have?
The median inner page has 43 internal links on desktop and 39 on mobile, and the median home page 51 and 46, according to the HTTP Archive Web Almanac 2025. Zyppy's 2022 correlation study found Google traffic rising up to about 40 to 44 incoming internal links per URL and declining after 45 to 50. Neither is a rule, but a desktop inner page with fewer than 19 internal links sits in the bottom quarter of the web.
Do internal links help SEO rankings?
Controlled tests show gains of 5% to 10%: SearchPilot measured a 7% organic traffic uplift from adding nearby-location links in 2021, and a 10% desktop lift from expanded footer links. Google itself says it uses links as a relevancy signal and to find new pages (Google Search Central, 2025).
Do popular websites use more internal links?
Yes, 3.1 times as many at the median: mobile home pages on the top 1,000 sites carried 129 internal links in June 2024 against 41 across all sites, computed from the Web Almanac 2024. On desktop home pages the gap is 3.6x (166 vs 46), and the count falls at every rank step.
Does exact-match anchor text help internal links?
Zyppy found pages with at least one exact-match internal anchor had at least five times more Google traffic, in a 2022 study of 23 million internal links across 1,800 sites. That is a correlation, and part of it runs the other way: sites give exact anchors to pages that already perform.
How common are orphan pages?
SEOmator found a median of 8% orphan pages on sites that failed its crawlability checks, in audits of more than 100,000 sites in January to July 2026. Even English Wikipedia has 52,447 tagged orphan articles, 0.72% of its 7.2 million articles (September 2026).
How many websites have broken internal links?
36.1% of sites audited by SEOmator in the first half of 2026 had broken internal links, with a median of 14 per site. Pew Research Center found at least one broken link on 23% of news pages and 21% of government pages (May 2024).
Sources
- HTTP Archive, Web Almanac 2025, SEO chapter (January 2026, updated June 2026)
- HTTP Archive, Web Almanac 2025, Methodology (January 2026)
- HTTP Archive, Web Almanac 2024, SEO chapter (December 2024)
- HTTP Archive, Web Almanac 2024, Methodology (December 2024)
- HTTP Archive, Web Almanac 2022, SEO chapter (September 2022)
- HTTP Archive, Web Almanac 2022, Methodology (September 2022)
- HTTP Archive, Web Almanac 2021, SEO chapter (2021)
- Google Search Central, SEO link best practices for Google (updated December 2025)
- Google Search Central, Learn about sitemaps (updated December 2025)
- Google Search Central, Crawl budget management for large sites (updated July 2026)
- Microsoft Bing, Bing Webmaster Guidelines (accessed September 2026)
- Sergey Brin and Lawrence Page, Stanford University, The Anatomy of a Large-Scale Hypertextual Web Search Engine (1998)
- WebAIM, The WebAIM Million 2026 (2026)
- Nielsen Norman Group, First 2 Words: A Signal for the Scanning Eye (April 2009)
- SEOmator, Internal Link Checker and 2026 audit data (2026)
- Botify, Orphan Pages and SEO (May 2020)
- Wikipedia, Category: All orphaned articles (retrieved 24 September 2026)
- Wikipedia, Special:Statistics (retrieved 24 September 2026)
- Pew Research Center, When Online Content Disappears (May 2024)
- Ahrefs, Link Rot Study (February 2024)
- Internet Archive, More than 9 million broken links on Wikipedia are now rescued (October 2018)
- Zyppy, 23 Million Internal Links: SEO Case Study (July 2022, updated February 2026)
- SearchPilot, Internal linking case study: links to nearby location pages (January 2021)
- SearchPilot, Should you add more internal links to your footer for SEO? (September 2021)
- Ryan Jones (SEOTesting) on the Screaming Frog blog, Finding and Testing Internal Link Changes (May 2025)
Methodology. Numbers were collected on 24 September 2026 and checked against the original publisher's page, not against other statistics roundups. Web Almanac figures come from the chapter text and each figure's published data sheet, which states the client and page type (home or inner page), with crawl month and sample size taken from each edition's methodology page. Vendor audit data (SEOmator, Botify, SEOTesting) describes each vendor's own crawls or site and is labelled as such. Correlation studies are marked as correlation. Numbers from different sources or populations are shown as separate facts and never combined into one figure. Last updated 24 September 2026.