Link Building Statistics UK 2026

Statistics 11–20 — Link Authority & Relevance Statistics

Not every backlink carries the same contextual or competitive significance.

Modern backlink analysis increasingly considers factors including:

• the strength of the linking page;
• the diversity of referring domains;
• the authority of the wider domain;
• anchor text;
• topical context;
• and whether the linking page itself receives Search visibility.

The statistics below demonstrate why backlink quality cannot be reduced to a simple link count.

As with the previous section, these are primarily international and US Search-industry datasets rather than UK population statistics.

11. Domain Rating Recorded a 0.131 Correlation With Google Rankings

Ahrefs’ analysis of the top one million US keywords by Search volume found that its Domain Rating metric had a Spearman correlation of 0.131 with Google ranking position.

This was materially lower than the correlations recorded for page-level backlink and referring-domain metrics.

For comparison, the same study recorded:

• Referring domains: 0.255;
• Backlinks: 0.248;
• Domain Rating: 0.131.

Authority implication: The authority of the overall domain matters, but page-level external evidence may provide a more direct competitive signal for an individual ranking URL.

Source:
Ahrefs — One Million SERPs Link Study

12. Position #1 Pages Average More Than 200 Referring Domains

Semrush reports that pages ranking in Google’s #1 position have more than 200 referring domains on average.

By comparison, pages in approximately position #10 average fewer than 80 referring domains.

The difference does not establish causation, but it demonstrates how substantially backlink profiles can vary across positions within competitive Search results.

Authority implication: Link-gap analysis should focus on referring-domain breadth rather than merely comparing raw backlink totals.

Source:
Semrush — Referring Domains and SEO

13. 8 of Semrush’s Top 20 Ranking Correlation Factors Were Backlink-Related

Semrush reports that eight of the 20 factors with the strongest correlations to ranking in its 2024 Ranking Factors research were related to backlinks.

That means backlink-related measurements accounted for 40% of the top 20 factors identified within that particular correlation study.

This should not be interpreted as meaning that links account for 40% of Google’s ranking algorithm.

The research measures correlations within the Semrush dataset rather than Google’s internal weighting systems.

Authority implication: External authority remains strongly represented among observable characteristics associated with high-ranking pages.

Source:
Semrush — Backlink Analysis and Ranking Factors

14. The Median Number of Backlinks for Top-Ranking Pages Was Only 13

Although some high-ranking pages in Semrush’s ranking research had thousands of backlinks, the median number of backlinks among top-ranking pages was only 13.

This is an important counterweight to the assumption that strong Search performance always requires extremely large backlink volumes.

Backlink requirements vary substantially according to:

• query competitiveness;
• Search intent;
• existing Brand Authority;
• content quality;
• and the strength of competing pages.

Authority implication: Link building should be based on the competitive environment of the target topic rather than arbitrary backlink-volume targets.

Source:
Semrush — How to Find and Analyse Backlinks

15. Exact-Match Anchor Text Had Only a Weak Correlation With Rankings

Ahrefs studied the top 20 Search results across 19,840 keywords, covering 384,614 webpages, to examine anchor-text relationships.

The percentage of exact-match anchored links showed a Spearman correlation of:

• 0.1436 using the average;
• 0.1869 using the median.

Ahrefs characterised this as a relatively weak relationship.

Authority implication: Exact-match anchor text should not be treated as a shortcut to ranking performance. A natural backlink profile is more important than attempting to engineer rigid anchor-text percentages.

Source:
Ahrefs — Anchor Text Study

16. Phrase-Match Anchor Text Had an Even Weaker Ranking Correlation

Within the same Ahrefs study, phrase-match anchor text produced a Spearman correlation of 0.1057 using average values and 0.1393 using median values.

The measured relationship was weaker than for exact-match anchors.

Authority implication: Anchor wording may provide context, but the evidence does not support treating keyword-rich anchor ratios as a dominant ranking strategy.

Source:
Ahrefs — Anchor Text Study

17. Partial-Match Anchor Text Produced a 0.1076 Average Correlation

Partial-match anchors produced results almost identical to phrase-match anchors.

Ahrefs measured:

• average Spearman correlation: 0.1076;
• median correlation: 0.1393.

Again, this represents only a very weak association.

Authority implication: A backlink’s value cannot reasonably be assessed from anchor-text keyword matching alone.

Source:
Ahrefs — Anchor Text Study

18. Random Anchor Text Had Almost Zero Correlation With Rankings

Ahrefs also examined generic or random anchor text such as phrases comparable to “click here” or “this article”.

The correlations were effectively zero:

• average: 0.0161;
• median: 0.0130.

This reinforces the wider finding that anchor-text ratios alone explain very little about ranking position.

Authority implication: Link acquisition should prioritise legitimate editorial references and useful placement rather than attempting to control every anchor phrase.

Source:
Ahrefs — Anchor Text Study

19. Only 20 of 44,589 SERPs Had Traffic-Bearing Links for Every Top-Ranking Page

Ahrefs examined 44,589 Search result pages to investigate whether backlinks from pages receiving organic Search traffic were particularly common among high-ranking results.

Only 20 SERPs had at least one traffic-bearing backlink pointing to every page in the top results.

That is approximately:

1 in every 2,229 SERPs

The result challenges simplistic assumptions that a backlink must come from a page receiving substantial organic traffic in order to have potential value.

Authority implication: Referring-page traffic can be a useful quality indicator, but it should not become an absolute requirement for evaluating every backlink.

Source:
Ahrefs — Links From Pages With Traffic Study

20. Around One-Fifth of SERPs Had No Top-Ranking Page With a Traffic-Bearing Backlink

In the same Ahrefs research, approximately one-fifth of the studied SERPs had no top-ranking page with a backlink from a referring page receiving organic Search traffic.

In other words, pages were capable of reaching Google’s first page even when none of their backlinks originated from pages meeting Ahrefs’ traffic threshold.

This does not mean referring-page traffic is irrelevant.

It means that link quality is multidimensional.

Potential considerations include:

• page-level authority;
• referring-domain strength;
• editorial relevance;
• topical context;
• placement;
• audience;
• and the broader link graph.

Authority implication: Link quality should be evaluated using several signals rather than a single metric such as referring-page traffic.

Source:
Ahrefs — Links From Pages With Traffic Study

Statistics 21–30 — Link Acquisition & Outreach Statistics

Link acquisition frequently depends on outreach, but the available evidence demonstrates how difficult cold outreach can be at scale.

Success is influenced by:

• prospect relevance;
• personalisation;
• subject-line construction;
• the number of contacts approached;
• follow-up strategy;
• timing;
• and the strength of the underlying reason for someone to link.

Several statistics in this section come from Backlinko and Pitchbox’s large analysis of 12 million outreach emails. Although this is an older benchmark dataset, its scale makes it useful for understanding the mechanics of outreach. More recent BuzzStream research covering over 51 million emails is included to provide a contemporary comparison.

21. Only 8.5% of Outreach Emails Received a Response

Backlinko and Pitchbox analysed 12 million outreach emails and found that only 8.5% received a response.

That means approximately:

1 response for every 12 outreach emails

The study highlights one of the central difficulties of traditional link acquisition: most unsolicited outreach does not generate an observable reply.

Outreach implication: Link-building forecasts should begin with realistic response assumptions rather than expecting every relevant publisher to engage.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

22. 91.5% of Outreach Emails Received No Response

The inverse of the 8.5% response rate is equally important.

Backlinko’s dataset found that 91.5% of outreach emails were ignored or otherwise received no response.

This demonstrates why increasing email volume alone can be an inefficient strategy.

If relevance and targeting remain poor, increasing the number of emails can simply increase the number of ignored messages.

Outreach implication: Prospect quality and pitch relevance should generally be improved before campaign volume is increased.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

23. Longer Outreach Subject Lines Produced 24.6% Higher Response Rates

Backlinko found that outreach emails with longer subject lines produced an average response rate 24.6% higher than emails with shorter subject lines.

A longer subject line may provide room to communicate additional context, relevance or specificity before the recipient opens the message.

The finding should not be interpreted as meaning that every subject line should simply be made longer.

Relevance and clarity remain more important than adding unnecessary words.

Outreach implication: Subject lines should communicate why the email matters to the recipient rather than relying only on short generic phrases.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

24. Personalised Subject Lines Increased Response Rates by 30.5%

Outreach emails with personalised subject lines received 30.5% higher response rates than emails without subject-line personalisation.

Personalisation can help indicate that the sender has identified the recipient intentionally rather than adding them to an indiscriminate bulk list.

However, meaningful personalisation requires more than automatically inserting a first name.

Useful personalisation can reference:

• a journalist’s recent coverage;
• a relevant article;
• a publication’s audience;
• a specific resource;
• or a genuine connection between the proposed asset and the recipient’s subject area.

Outreach implication: Relevance-based personalisation can materially improve the probability of engagement.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

25. Personalising the Email Body Increased Response Rates by 32.7%

Backlinko found that outreach emails with personalised body copy generated 32.7% higher response rates than emails without personalised message content.

The finding is particularly relevant to scaled link building.

Automated templates may improve efficiency, but excessive standardisation can remove the context that explains why the recipient was approached.

Outreach implication: Automation can support outreach operations, but the underlying message still needs to demonstrate specific relevance to the recipient.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

26. Multiple Outreach Attempts Generated Twice as Many Responses

The Backlinko/Pitchbox analysis found that contacting the same prospect multiple times generated approximately twice as many responses as sending only one outreach email.

Recipients may miss an initial email because of:

• inbox volume;
• timing;
• other priorities;
• holidays;
• or simply not seeing the message.

A relevant follow-up can therefore recover opportunities that would otherwise have been lost.

Outreach implication: A campaign should not automatically be judged as unsuccessful after the first email, although follow-ups should remain proportionate and relevant.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

27. Contacting Multiple People at the Same Organisation Increased Responses by 93%

Sending outreach to more than one relevant contact at an organisation produced a response rate 93% higher than contacting only one person.

The effect was particularly relevant for larger organisations where responsibility for:

• content;
• editorial decisions;
• PR;
• partnerships;
• or website management

may be distributed across several people.

Backlinko also found diminishing returns after approximately five contacts.

Outreach implication: Identifying the correct person can be as important as identifying the correct website.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

28. Multiple Contacts Combined With Multiple Messages Increased Responses by 160%

When Backlinko compared a single email to one person with campaigns using multiple messages and multiple contacts, the latter generated a 160% higher response rate.

This was one of the strongest differences identified in the 12-million-email study.

It demonstrates how outreach performance can depend on the campaign process rather than merely the wording of one email.

Outreach implication: Structured outreach sequences can outperform one-off mass emailing when the contacts are genuinely relevant.

Source:
Backlinko / Pitchbox — 12 Million Outreach Email Study

29. A 2025 Study Found a 3.59% Reply Rate Across Link-Building Outreach

BuzzStream analysed a dataset containing more than 51 million emails and reported an overall reply rate of 3.59% for link-building outreach.

That equates to approximately:

1 reply for every 28 link-building emails

The result is materially below the older 8.5% Backlinko/Pitchbox benchmark.

The studies use different datasets, populations and time periods, so the difference should not automatically be interpreted as proof that outreach effectiveness has declined by a specific amount.

It does demonstrate that response-rate assumptions can vary considerably according to campaign type and dataset.

Outreach implication: Organisations should establish their own baseline reply and placement rates rather than applying one industry benchmark universally.

Source:
BuzzStream — Analysis of More Than 51 Million Outreach Emails

30. 57% of Link-Building Replies Arrived Within Six Hours

BuzzStream’s 2025 outreach analysis found that 57% of link-building replies arrived within six hours of the original email.

Almost 90% arrived within two days.

The average response time was 12.85 hours, with slower responses creating a longer tail.

This provides useful context for follow-up timing.

An immediate second email may be unnecessary when a significant proportion of interested recipients respond during the first day.

Outreach implication: Outreach teams should measure response timing as well as response rate and avoid judging campaign performance too quickly.

Source:
BuzzStream — How Long Does It Take to Get Email Outreach Replies?

Statistics 31–40 — Digital PR Statistics

Digital PR has become one of the principal ways organisations attempt to earn authoritative editorial coverage and backlinks.

Unlike traditional link acquisition, Digital PR normally begins with something intended to be genuinely newsworthy or useful to journalists, including:

• original data;
• research;
• expert commentary;
• newsjacking;
• creative campaigns;
• press releases;
• and proprietary insight.

The statistics below combine BuzzStream’s 2026 State of Digital PR research involving more than 150 Digital PR professionals with Muck Rack’s research involving more than 1,500 journalists.

Together, the datasets provide perspectives from both sides of the media relationship: the professionals seeking coverage and the journalists deciding what is worth covering.

31. 85.8% of Digital PR Professionals Say Backlink Building Is an Effective Outcome of Digital PR

BuzzStream’s 2026 State of Digital PR research found that 85.8% of respondents identified building backlinks as an area where Digital PR is effective.

Backlinks remained the most frequently identified outcome in the study, ahead of other effects including Brand Awareness and organic Search performance.

This is important because Digital PR is increasingly discussed as a broader Brand Authority discipline, but earned links remain a major part of how practitioners evaluate its value.

Digital PR implication: Editorial link acquisition remains one of the clearest measurable connections between PR activity and Search Authority.

Source:
BuzzStream — State of Digital PR Report 2026

32. 68.2% Say Digital PR Is More Effective Than It Was 12 Months Earlier

More than two-thirds of Digital PR professionals surveyed by BuzzStream said Digital PR had become more effective during the previous 12 months.

BuzzStream reported the figure at 68.2%.

This is a practitioner perception rather than an independently measured performance outcome, but it provides evidence of confidence within the Digital PR industry during 2026.

Digital PR implication: Digital PR practitioners increasingly see earned media as relevant to a wider visibility environment rather than solely conventional media exposure.

Source:
BuzzStream — State of Digital PR Report 2026

33. 85.2% Say Digital PR Produces Measurable Results Within Six Months

BuzzStream found that 85.2% of surveyed Digital PR professionals said measurable campaign results typically appear within six months.

This helps establish an important distinction between:

• the time required to secure the first piece of coverage;
• the time required to build a portfolio of coverage and links;
• and the longer period potentially required to observe Search or commercial impact.

Digital PR implication: Campaign evaluation should not be limited to whether a story receives coverage immediately after launch.

Source:
BuzzStream — State of Digital PR Report 2026

34. 51.4% Report That Measurable Digital PR Results Typically Take Three to Six Months

More specifically, 51.4% of respondents said it normally takes approximately three to six months for a Digital PR campaign to generate measurable results.

BuzzStream reported that this proportion was around five percentage points higher than in its previous study.

The finding supports a longer-term view of Digital PR rather than treating each campaign as an isolated short-term link acquisition exercise.

Digital PR implication: Sustainable authority programmes need sufficient time to accumulate coverage, links, mentions and associated visibility signals.

Source:
BuzzStream — State of Digital PR Report 2026

35. 81% of Digital PR Professionals Secure Their First Coverage Within One Week

Although overall campaign effects may take months to assess, initial editorial coverage can occur much faster.

BuzzStream found that 81% of Digital PR professionals reported securing first coverage within one week of pitching.

The most commonly reported individual timeframe was three to five days, selected by 29.9% of respondents.

Digital PR implication: Initial media pickup and longer-term campaign performance should be measured separately.

Source:
BuzzStream — State of Digital PR Report 2026

36. 85.1% Use Quality Links as a Digital PR Success Metric

The most commonly reported Digital PR success metric in BuzzStream’s 2026 study was the number of quality links.

It was tracked by 85.1% of respondents.

Other commonly monitored measures included:

• total mentions: 72.3%;
• total links: 70.9%;
• organic traffic and rankings: 63.5%.

The distinction between total links and quality links is significant.

Digital PR implication: Digital PR teams increasingly recognise that the source and quality of editorial coverage matter more than accumulating the largest possible raw placement count.

Source:
BuzzStream — State of Digital PR Report 2026

37. 32.5% Say One Digital PR Professional Can Generate 31+ Links Per Month

BuzzStream found that 32.5% of respondents believed a single Digital PR team member could generate an average of 31 or more links per month.

That equates to roughly one link per day.

The study notes that the figure does not automatically distinguish every link according to:

• authority;
• relevance;
• syndication;
• follow attributes;
• or editorial value.

Digital PR implication: Link volume can vary considerably between campaigns and sectors, making quality and relevance necessary alongside output targets.

Source:
BuzzStream — State of Digital PR Report 2026

38. 84% of Journalists Say Stories Often Begin With PR Pitches

Muck Rack’s 2025 State of Journalism research, based on responses from more than 1,500 journalists, found that 84% said stories often begin with pitches from PR professionals.

The finding demonstrates that journalist pitching remains capable of contributing directly to editorial discovery.

However, this does not mean journalists welcome indiscriminate outreach.

Digital PR implication: A strong, relevant pitch can provide journalists with a legitimate starting point for editorial coverage.

Source:
Muck Rack — State of Journalism 2025

39. 86% of Journalists Ignore Off-Topic PR Pitches

Muck Rack found that 86% of journalists ignore pitches that are outside their area of coverage.

This is one of the clearest quantitative arguments against indiscriminate media-list building.

A campaign can have:

• strong data;
• a compelling headline;
• good creative assets;
• and an experienced PR team

and still underperform if the story is sent to journalists who do not cover the subject.

Digital PR implication: Journalist relevance is not a minor optimisation. It is a fundamental part of campaign distribution.

Source:
Muck Rack — State of Journalism 2025

40. 69% of Journalists Prefer PR Pitches Under 200 Words

Muck Rack’s journalism research found that 69% of journalists prefer pitches containing fewer than 200 words.

This matters because journalists operate in exceptionally crowded inbox environments.

Separate Muck Rack findings show that on a normal day:

• 55% of journalists receive more than five PR pitches;
• 17% receive between 11 and 20 pitches.

A pitch therefore needs to communicate the essential story quickly.

Digital PR implication: The strongest pitch is usually not the longest explanation of the campaign. It is the clearest explanation of why the story matters to that journalist and their audience.

Source:
Muck Rack — State of Journalism Pitching Preferences

Statistics 41–50 — Content That Earns Links: Research, Data, Statistics, Tools & Linkable Assets

The majority of published content earns very few external links.

This makes the type of content an organisation creates increasingly important.

Large-scale content studies consistently suggest that certain assets are more naturally referenceable than ordinary promotional or informational content.

These can include:

• original research;
• statistics;
• new datasets;
• industry reports;
• reference resources;
• infographics;
• tools;
• and authoritative answers to important questions.

Several of the largest quantitative studies in this area are older than the 2026 publication date of this research. They are included because they analyse exceptionally large content datasets and continue to provide useful benchmarks. Their dates should be considered when interpreting the findings.

41. 94% of 912 Million Blog Posts Had Zero External Links

Backlinko and BuzzSumo analysed 912 million blog posts and found that 94% had received zero external links.

Only approximately 6% of the content in the dataset had earned at least one external backlink.

The scale of the study illustrates how unusual naturally linked content actually is.

Content implication: Simply publishing content does not make it linkable. A resource needs a reason for another publisher to cite or reference it.

Source:
Backlinko / BuzzSumo — Analysis of 912 Million Blog Posts

42. Only 2.2% of Content Earned Links From More Than One Website

Within the same 912-million-post dataset, only 2.2% of content generated backlinks from more than one website.

This is a more demanding benchmark than simply earning one backlink.

It measures whether multiple independent publishers considered the resource worth referencing.

Content implication: Assets capable of earning links from multiple independent domains are relatively rare and potentially much more valuable for building referring-domain diversity.

Source:
Backlinko / BuzzSumo — Analysis of 912 Million Blog Posts

43. Content Over 3,000 Words Earned 77.2% More Referring-Domain Links Than Content Under 1,000 Words

Backlinko and BuzzSumo found that content exceeding 3,000 words received an average of 77.2% more links from referring domains than content containing fewer than 1,000 words.

The finding does not mean that adding unnecessary words causes backlinks.

Longer resources may have greater potential to provide:

• more evidence;
• more original analysis;
• greater topic coverage;
• more reference points;
• and more reasons for other publishers to cite them.

Content implication: Comprehensive resources can have greater link-earning potential when additional depth adds genuine informational value.

Source:
Backlinko / BuzzSumo — Analysis of 912 Million Blog Posts

44. “Why”, “What” and Infographic Content Earned 25.8% More Referring-Domain Links Than How-To Posts and Videos

The 912-million-post study found that “Why” posts, “What” posts and infographics received an average of 25.8% more referring-domain links than how-to articles and videos.

These formats often provide information that can be incorporated into another publisher’s work as:

• an explanation;
• a definition;
• a supporting reference;
• or a visual citation.

Content implication: Content designed to become a reference resource may have different characteristics from content designed primarily to generate immediate visits or social engagement.

Source:
Backlinko / BuzzSumo — Analysis of 912 Million Blog Posts

45. Social Shares and Backlinks Had a Correlation of Only 0.078

Backlinko and BuzzSumo found almost no relationship between the number of social shares received by an article and the number of backlinks it earned.

The measured Pearson correlation coefficient was only 0.078.

This is important because highly shareable content is sometimes assumed to be highly linkable.

The evidence suggests that people may share and cite content for different reasons.

Content implication: Social engagement and link acquisition should be measured as different content outcomes rather than being treated as interchangeable indicators of success.

Source:
Backlinko / BuzzSumo — Analysis of 912 Million Blog Posts

46. 93% of B2B Content Received Zero External Links

Backlinko and BuzzSumo separately examined the B2B subset of their dataset.

They found that 93% of B2B content received no external backlinks.

This was close to the 94% recorded across the complete dataset.

The result is particularly relevant to professional-services, SaaS, technology, manufacturing and other B2B organisations publishing large volumes of informational content.

Content implication: Producing B2B articles at scale does not automatically create external authority. Content needs distinctive informational value if it is expected to earn citations.

Source:
Backlinko / BuzzSumo — B2B Content Analysis

47. Only 3% of B2B Content Earned Links From More Than One Website

Within the B2B subset, only 3% of content received links from more than one website.

The result closely matched the 2.2% level recorded across the broader dataset.

This demonstrates how difficult it is for ordinary B2B content to earn repeated independent references.

Content implication: B2B organisations seeking authority should identify a smaller number of high-value reference assets rather than assuming every article needs to become a link magnet.

Source:
Backlinko / BuzzSumo — B2B Content Analysis

48. The Median Number of Backlinks Across 100 Million Articles Was Zero

BuzzSumo’s analysis of approximately 100 million articles published during 2017 found that the median number of backlinks received by an article was zero.

In other words, the typical article in that very large dataset did not receive a backlink.

The research identified stronger backlink performance among authoritative:

• research content;
• reference content;
• opinion content;
• and long-form resources.

Content implication: Link acquisition is an exceptional outcome rather than the default result of content publication.

Source:
BuzzSumo — Content Trends 2018, Analysis of 100 Million Articles

49. More Than 70% of the 100 Million Articles Were Never Linked to From Another Domain

The same BuzzSumo analysis found that more than 70% of all content in the 100-million-article sample was never linked to from another domain.

BuzzSumo’s later, larger collaboration with Backlinko produced an even higher zero-link figure for blog content, demonstrating how results can vary according to:

• sample construction;
• content type;
• measurement period;
• and the underlying index.

Content implication: Zero-link percentages should not be treated as universal constants, but multiple large studies agree that most published content attracts little or no external linking activity.

Source:
BuzzSumo — Content Trends 2018

50. One Annual Research Report Earned 10× the Domain Links of Other Content With Similar Social Sharing

BuzzSumo’s research into content that attracts both links and shares examined the annual industry research published by Social Media Examiner.

Its annual survey report received approximately ten times the number of domain links generated by other content on the site receiving comparable levels of social sharing.

BuzzSumo identified original research and valuable insights as one of the content categories repeatedly represented among highly referenceable material.

This is a case study from an individual publisher rather than a universal content benchmark, but it illustrates why proprietary data can become a persistent citation asset.

Content implication: An organisation that produces the underlying evidence becomes the primary source other writers can cite when discussing the finding.

Source:
BuzzSumo / Majestic — Research Into Content That Gets Links and Shares

Statistics 51–60 — Link Building Costs & Investment Statistics

Modern link acquisition is not cost-free simply because the final backlink is editorially earned.

Investment can include:

• research;
• data collection;
• surveys;
• content production;
• creative assets;
• media databases;
• journalist outreach;
• PR expertise;
• analysis;
• and campaign management.

The statistics below primarily draw on two BuzzStream studies: its 2026 State of Digital PR survey of more than 150 professionals and a separate Digital PR pricing survey of approximately 70 agencies, freelancers and consultants.

The pricing survey is particularly relevant to this UK research because 54% of respondents were UK-based.

All monetary figures are reported in US dollars, matching the original research.

51. Around 60% of Digital PR Teams Report Monthly Budgets Below $10,000

BuzzStream’s 2026 State of Digital PR study found that approximately 60% of respondents operated with monthly Digital PR budgets below $10,000.

This demonstrates that Digital PR investment spans a wide range rather than being limited to large enterprise-level campaigns.

However, the research also shows increasing investment at the upper end of the market.

Investment implication: Digital PR should be budgeted as an ongoing authority-building activity rather than viewed only as the cost of individual links.

Source:
BuzzStream — State of Digital PR Report 2026

52. 25.7% of Digital PR Teams Report Monthly Budgets Below $5,000

In 2026, 25.7% of surveyed Digital PR professionals reported monthly budgets below $5,000.

The comparable figure in the previous year’s BuzzStream research was 34.1%.

The share of respondents operating at the lowest budget level therefore decreased by 8.4 percentage points.

Investment implication: The distribution is moving away from the lowest budget tier, suggesting greater investment among at least part of the Digital PR market.

Source:
BuzzStream — State of Digital PR Report 2026

53. The Share of Teams Spending $20,000+ Per Month More Than Doubled to 8.8%

BuzzStream found that the proportion of respondents reporting Digital PR budgets of $20,000 or more per month increased from 4% to 8.8%.

That represents more than a doubling of the share operating in the highest budget bracket measured in the report.

Higher campaign budgets can reflect investment in:

• larger proprietary datasets;
• consumer surveys;
• creative production;
• specialist PR staff;
• media monitoring;
• and sustained campaign activity.

Investment implication: At the upper end of the market, authority building is increasingly being funded as a substantial ongoing marketing function.

Source:
BuzzStream — State of Digital PR Report 2026

54. The Most Common Reported Cost per Link Range Is $300–$500

Among respondents able to report a cost per link, the most common range in BuzzStream’s 2026 survey was $300–$500.

This range was selected by 19.6% of respondents.

Cost per link remains widely used because it converts PR activity into an easily understood acquisition metric.

However, two backlinks costing the same amount can differ substantially in:

• publisher authority;
• editorial relevance;
• audience;
• brand exposure;
• traffic;
• and long-term value.

Investment implication: Cost per link can provide operational context, but it should not be used as a substitute for measuring link quality or business impact.

Source:
BuzzStream — State of Digital PR Report 2026

55. 10.2% Now Report a Cost per Link of $750 or More

BuzzStream found that 10.2% of respondents reported an average cost per link of at least $750 in 2026.

The equivalent figure in its previous research was only 3%.

The share of respondents operating in this higher cost bracket therefore increased by more than three times.

This should not be interpreted as meaning that every backlink has a market value of $750 or more.

Earned editorial links do not have a universal price.

Investment implication: Higher-quality campaigns can require significant research, production and specialist labour even when the resulting editorial link itself is not purchased.

Source:
BuzzStream — State of Digital PR Report 2026

56. 39.2% of Digital PR Professionals Do Not Know Their Cost per Link

Despite the continued use of cost per link as a reporting metric, 39.2% of respondents in BuzzStream’s 2026 State of Digital PR study said they did not know their average cost per link.

That was down from 51.4% in the previous year.

The figure illustrates the difficulty of reducing Digital PR to a single link-acquisition cost.

Campaigns may simultaneously generate:

• links;
• brand mentions;
• media coverage;
• referral traffic;
• Search visibility;
• and AI citations.

Investment implication: A complete measurement model needs to examine the wider return produced by earned media rather than calculating only link volume divided by campaign expenditure.

Source:
BuzzStream — State of Digital PR Report 2026

57. The Average Digital PR Contract in BuzzStream’s Pricing Study Was $5,458 Per Month

BuzzStream’s separate Digital PR pricing study surveyed approximately 70 agency leaders, freelancers and consultants.

The average monthly Digital PR contract across all respondents was approximately $5,458.

The dataset was particularly relevant to the UK market because 54% of survey participants were UK-based, compared with 23% from the United States.

Investment implication: Professional Digital PR commonly represents a recurring monthly investment rather than a low-cost transactional link acquisition service.

Source:
BuzzStream — Digital PR Cost Survey

58. The Average Reported Digital PR Cost per Link Was $597

Across BuzzStream’s Digital PR pricing survey, the calculated average cost per link was $597.

Approximately one quarter of respondents reported a cost below $300, while other responses were distributed across substantially higher ranges.

The overall figure should be interpreted cautiously because campaign types varied considerably.

The research included activities ranging from:

• guest posting;
• journalist-request responses;
• reactive PR;
• proactive PR;
• and larger hero-content campaigns.

Investment implication: Comparing Digital PR providers purely on cost per link can be misleading unless the underlying service model and quality expectations are also comparable.

Source:
BuzzStream — Digital PR Cost Survey

59. UK Respondents Reported an Average Cost per Link of $818

When BuzzStream separated its Digital PR pricing survey geographically, UK respondents recorded an average cost per link of $818.

The comparable figures were:

• UK: $818;
• Other markets: $461;
• United States: $350.

The reported UK cost per link was therefore more than twice the US figure within this particular survey.

BuzzStream observed that UK respondents were more heavily represented in comparatively resource-intensive services such as hero content and reactive/proactive PR, while US respondents were more likely to report journalist-request and guest-posting services.

UK implication: Cost comparisons between markets need to account for differences in the type of Digital PR being delivered rather than assuming identical service models.

Source:
BuzzStream — UK vs US Digital PR Cost Analysis

60. Around 75% of Digital PR Contracts Run for More Than Six Months

BuzzStream’s pricing research found that approximately three quarters of Digital PR contracts lasted longer than six months.

Only around one quarter fell into the three-month-or-shorter category.

The study also found substantial differences between agencies and freelancers:

• 53% of agency contracts were in the 6–11 month range;
• 35% of agency contracts lasted 12 months or longer;
• compared with 43% and 20% respectively among freelancers and consultants.

This supports the view that authority building is generally treated as a continuing programme rather than a short isolated intervention.

Investment implication: Organisations should evaluate Digital PR across a sufficiently long period to measure accumulated links, media coverage, authority and Search impact.

Source:
BuzzStream — Digital PR Cost Survey

Statistics 61–70 — Link Loss, Content Decay & the Disappearing Web

Backlinks are not permanent assets.

Pages disappear, domains close, publishers update articles, URLs change and links are removed.

This process creates link rot and wider digital decay across the web.

The statistics below draw primarily on Pew Research Center’s large-scale analysis of almost one million historical webpages and Ahrefs’ study of links pointing to more than two million websites.

Together, they demonstrate why authority building requires not only earning new links but also monitoring and preserving existing ones.

61. 25% of Webpages From 2013–2023 Were No Longer Accessible by October 2023

Pew Research Center analysed a random sample of just under one million webpages collected through Common Crawl between 2013 and 2023.

By October 2023, 25% of those webpages were no longer accessible.

The finding demonstrates that a substantial proportion of published web content disappears over time.

Authority implication: A backlink portfolio is not static. Links earned today may disappear as the pages and websites containing them change.

Source:
Pew Research Center — When Online Content Disappears

62. 38% of Webpages That Existed in 2013 Were Gone a Decade Later

The effect becomes more pronounced as content ages.

Pew found that 38% of webpages captured in its 2013 sample were no longer accessible in 2023.

In other words, almost four in ten sampled pages disappeared within approximately a decade.

Authority implication: Long-term authority strategies need to account for the natural erosion of external references as older content disappears from the web.

Source:
Pew Research Center — When Online Content Disappears

63. Around 22% of Webpages From 2021 Were Already Inaccessible by 2023

Digital decay is not confined to very old parts of the web.

Pew found that approximately 22% of webpages captured in its 2021 sample were no longer accessible by 2023.

This means that around one in five sampled pages had disappeared within roughly two years.

Authority implication: Link loss can occur much faster than organisations may assume, making backlink monitoring relevant even for recently acquired links.

Source:
Pew Research Center — When Online Content Disappears

64. 16% of Historical Pages Became Inaccessible While Their Parent Domain Remained Active

Of the webpages Pew examined from 2013–2023, 16% were individually inaccessible even though the root-level website remained functional.

This type of decay can occur when publishers:

• delete old articles;
• remove obsolete resources;
• change URL structures;
• consolidate content;
• or fail to implement redirects correctly.

Authority implication: A backlink can disappear even when the referring website itself remains healthy and active.

Source:
Pew Research Center — When Online Content Disappears

65. 9% of Historical Pages Were Inaccessible Because the Entire Root Domain No Longer Functioned

A further 9% of webpages in Pew’s 2013–2023 sample were inaccessible because their entire root-level domain was no longer functioning.

This is an important distinction from an individual article simply being deleted.

When an entire domain disappears, every backlink from that website may effectively disappear with it.

Authority implication: A diversified backlink profile can reduce dependence on any single external website or publisher remaining permanently active.

Source:
Pew Research Center — When Online Content Disappears

66. 23% of News Webpages Contained at Least One Broken Link

Pew sampled 500,000 pages from 2,063 news and information websites.

Across those pages, 23% contained at least one broken external link.

The research analysed more than 14 million external links from the sampled news pages.

Overall, approximately 5% of those external links pointed to pages that were no longer accessible.

Authority implication: Editorial coverage from news publishers can decay over time as the external resources journalists originally referenced disappear.

Source:
Pew Research Center — News Website Link Decay

67. 21% of Government Webpages Contained at Least One Broken Link

Pew also analysed approximately 500,000 government webpages.

It found that 21% contained at least one broken link.

Across the government websites studied, researchers identified approximately 42 million links.

Around 6% of sampled government links led to pages that were no longer accessible.

Authority implication: Link rot affects even highly established institutional websites. Publisher authority does not guarantee that every reference will remain available indefinitely.

Source:
Pew Research Center — Government Website Link Decay

68. At Least 66.5% of Links in an Ahrefs Nine-Year Study Had Rotted

Ahrefs examined links pointing to 2,062,173 websites using backlink data extending from January 2013.

It found that at least 66.5% of links identified during the study period had rotted.

Ahrefs defines link rot as links that no longer point to the original intended resource because pages have been:

• removed;
• redirected;
• changed;
• or become unavailable.

Authority implication: Link attrition is not an exceptional event. Over long periods it is a normal characteristic of the web.

Source:
Ahrefs — Link Rot Study

69. 74.5% of Links in the Ahrefs Study Were Classified as Lost for SEO Purposes

Ahrefs’ broader definition of links considered lost for SEO purposes produced an even higher figure.

A total of 74.5% of links in the study were classified as lost.

This includes the 66.5% identified as link rot along with additional technical conditions that prevented links from continuing to be counted normally within the Ahrefs index.

Authority implication: Historical backlink totals can substantially overstate the amount of external authority that remains active today.

Source:
Ahrefs — Link Rot Study

70. 47.7% of Lost Links Came From Dropped Pages and 34.2% From Links Being Removed

Ahrefs identified two dominant causes behind lost links.

• 47.7% came from pages that had been dropped from the Ahrefs index;
• 34.2% occurred because the page still existed but the backlink itself had been removed.

Other causes included:

• 6.45% from crawl errors;
• 5.99% from redirected pages;
• 4.11% because the linking page was no longer found;
• 0.82% because the page was no longer canonical;
• 0.73% because the linking page had been marked noindex.

This matters because different forms of link loss require different responses.

A removed backlink may justify relationship-based outreach, while a backlink pointing to a deleted page on the organisation’s own site may be recoverable through an appropriate redirect.

Authority implication: Lost-link analysis should identify why the backlink disappeared before deciding whether reclamation is possible.

Source:
Ahrefs — Link Rot Study

Statistics 71–80 — Brand Mentions, Entities & Authority Statistics

Authority is broader than backlinks alone.

Search engines, AI systems and consumers can encounter an organisation through a distributed network of external evidence including:

• reviews;
• business profiles;
• directories;
• media coverage;
• social platforms;
• industry websites;
• citations;
• and other third-party references.

These signals help establish whether information about an organisation is consistent, current and corroborated beyond its own website.

The statistics in this section primarily use BrightLocal’s 2026 Local Consumer Review Survey, conducted with a representative panel of 1,002 US adults. These are consumer-authority benchmarks rather than UK population estimates and should not be interpreted as direct Google ranking factors.

71. 97% of Consumers Read Online Reviews for Local Businesses

BrightLocal’s 2026 Local Consumer Review Survey found that 97% of respondents read online reviews when researching local businesses.

Reviews create a large body of third-party information about an organisation outside its own website.

They can contain references to:

• products;
• services;
• staff;
• locations;
• customer experiences;
• pricing;
• and operational characteristics.

Authority implication: External reputation has become part of the information environment through which consumers evaluate whether an organisation appears credible.

Source:
BrightLocal — Local Consumer Review Survey 2026

72. 41% of Consumers Always Read Reviews When Researching a Business

BrightLocal found that 41% of consumers always read reviews when browsing for businesses.

That was substantially higher than the 29% recorded in the previous year’s survey.

The finding illustrates how external validation can become a routine part of the decision journey rather than an occasional secondary check.

Authority implication: An organisation’s reputation increasingly exists across third-party platforms as well as on its own website.

Source:
BrightLocal — Local Consumer Review Survey 2026

73. Consumers Use an Average of Six Review Platforms When Choosing Businesses

BrightLocal reports that the average consumer uses six different review sites when researching businesses.

That means organisational reputation is increasingly distributed across multiple platforms rather than concentrated within one profile.

Potential sources can include:

• Google;
• Facebook;
• Trustpilot;
• Tripadvisor;
• Apple Maps;
• sector-specific directories;
• and other third-party platforms.

Entity implication: Consistent business information and reputation across independent sources becomes increasingly important when consumers move between multiple platforms.

Source:
BrightLocal — Local Consumer Review Survey 2026

74. AI Tools for Local Business Recommendations Increased From 6% to 45%

BrightLocal found that reported use of ChatGPT and other generative AI tools for local business recommendations increased from 6% to 45%.

That made AI tools the third most-used source for local recommendations within the survey.

This development is especially significant for authority building because AI-generated recommendations may draw on information distributed across:

• business websites;
• reviews;
• directories;
• media coverage;
• third-party articles;
• and other online sources.

Entity implication: Consistent third-party information is becoming relevant not only to human research but also to AI-mediated business discovery.

Source:
BrightLocal — Local Consumer Review Survey 2026

75. 47% of Consumers Will Not Use a Business With Fewer Than 20 Reviews

BrightLocal found that 47% of consumers would not use a business with fewer than 20 reviews.

Only 9% said they were willing to use a business with five or fewer reviews.

The number of independent customer references therefore contributes to perceived credibility for many consumers.

This is conceptually similar to referring-domain diversity in link building: repeated independent validation is different from one isolated endorsement.

Authority implication: The breadth of third-party evidence can affect how credible an organisation appears, even where each individual mention is relatively small.

Source:
BrightLocal — Local Consumer Review Survey 2026

76. 74% of Consumers Prioritise Reviews From the Previous Three Months

BrightLocal found that 74% of consumers look for reviews written within the previous three months.

More specifically:

• 18% were influenced only by reviews written within the previous week;
• 32% looked for reviews written within the previous two weeks.

This demonstrates that reputation evidence can decay in perceived relevance even when the underlying review remains online.

Authority implication: Authority signals require freshness as well as historical accumulation. An organisation with strong but outdated external evidence may appear less current than one receiving consistent recent validation.

Source:
BrightLocal — Local Consumer Review Survey 2026

77. 31% of Consumers Will Only Use Businesses Rated 4.5 Stars or Higher

BrightLocal found that 31% of consumers will only use a business with a rating of 4.5 stars or higher.

That increased from 17% in the previous year’s study.

The research also found that 68% will only use businesses rated four stars or higher, up from 55%.

Authority implication: Third-party reputation signals can directly affect whether an organisation enters the consumer’s consideration set.

Source:
BrightLocal — Local Consumer Review Survey 2026

78. 56% Say Consistent Sentiment Across Multiple Reviews Is the Most Important Review Signal

When BrightLocal asked which review characteristics made consumers feel positively about using a business, the highest-ranked factor was whether a review was supported by other reviews expressing similar sentiment.

This factor was selected by 56% of respondents.

Other factors included:

• 46% — the review describes a positive experience;
• 44% — the review was posted within the last month;
• 42% — the review has a high star rating;
• 37% — the business owner responded.

Entity implication: Repeated, consistent evidence across independent sources can be more persuasive than one isolated positive reference.

Source:
BrightLocal — Local Consumer Review Survey 2026

79. Positive Reviews Increase Likelihood of Using a Business for 85% of Consumers

BrightLocal found that 85% of consumers are more likely to use a business after reading positive reviews.

Conversely, 77% said negative reviews make them less likely to choose the organisation.

The effect demonstrates that third-party references can influence not only general reputation but actual inclusion or exclusion from consideration.

Authority implication: Brand Authority is shaped by what independent sources say about an organisation, not only by what the organisation says about itself.

Source:
BrightLocal — Local Consumer Review Survey 2026

80. 54% Visit the Business Website After Reading Positive Reviews

BrightLocal found that after reading positive reviews, 54% of consumers visit the business’s website.

The survey also found that 66% conduct additional research after reading positive reviews, while 34% are prepared to move directly toward a purchase or booking.

This demonstrates how off-site authority can influence subsequent owned-site discovery.

The journey can therefore look like:

Third-Party Evidence
↓
Trust
↓
Brand Research
↓
Website Visit
↓
Decision

Authority implication: External reputation should not be measured in isolation. It can contribute to downstream branded research, website visits and conversion behaviour.

Source:
BrightLocal — Local Consumer Review Survey 2026

Statistics 81–90 — AI Search, Citations & Source Authority Statistics

The growth of AI Search introduces another form of external authority: source citation.

A backlink and an AI citation are not the same thing.

A backlink is a persistent hyperlink published on an external webpage.

An AI citation is a source selected dynamically by a generative system while constructing an answer.

However, both involve an external system deciding that a source is sufficiently relevant or useful to reference.

The emerging evidence shows that AI citation visibility overlaps with traditional Search Authority, but the relationship is far from one-to-one.

The statistics below primarily use large 2025–2026 Ahrefs and Semrush datasets. Most are based on US or international AI Search activity and should therefore be treated as AI Search benchmarks rather than UK population statistics.

81. Only 37.9% of AI Overview Citations Also Appeared Within Google’s First 10 Result Blocks

Ahrefs analysed 863,000 keyword SERPs and approximately four million AI Overview URLs in its updated 2026 research.

Only 37.9% of URLs cited by AI Overviews also appeared within the first ten Google result blocks for the original query.

A further:

• 31.2% appeared between positions 11 and 100;
• 31.0% appeared beyond the top 100 result blocks.

This represents a substantial change from Ahrefs’ earlier 2025 study, when much greater citation overlap with direct Search results was observed.

AI authority implication: Ranking for the exact query remains relevant, but ranking alone does not determine which sources an AI Overview will select.

Source:
Ahrefs — 38% of AI Overview Citations Pull From the Top 10

82. 36.7% of AI Overview Citations Did Not Rank in Google’s Top 100 Organic Results

When Ahrefs repeated its analysis using only conventional organic blue links and excluded other SERP features, it found that 36.7% of AI Overview-cited URLs did not rank within the first 100 organic results for the same query.

The remaining citations were distributed as follows:

• 37.1% ranked in the organic top 10;
• 26.2% ranked between positions 11 and 100;
• 36.7% ranked outside the top 100.

Google’s use of query expansion and related Search journeys can allow AI Overviews to identify sources beyond the direct SERP for the original query.

AI authority implication: Organisations competing for AI citation visibility may need authority across the wider topic rather than only one exact keyword.

Source:
Ahrefs — AI Overview Citation and Organic Ranking Study 2026

83. 18.2% of AI Overview Citations Outside Google’s Top 100 Came From YouTube

Among AI Overview citations that did not rank within Google’s top 100 results for the same keyword, Ahrefs found that 18.2% were YouTube URLs.

YouTube URLs represented approximately 5.6% of all AI Overview citations in the complete dataset.

This demonstrates how AI systems can draw authority from content formats and sources that may not occupy a conventional blue-link position for the original query.

AI authority implication: Source Authority increasingly extends beyond conventional webpages into video, transcripts and other machine-readable media.

Source:
Ahrefs — AI Overview Citation Study 2026

84. YouTube’s AI Overview Citation Visibility Grew 34% in Six Months

Ahrefs reported that YouTube had become the most-cited domain within AI Overviews in its Brand Radar dataset.

Its citation presence had grown by 34% during the previous six months.

Separate Ahrefs research involving 75,000 brands also found a strong relationship between YouTube mentions and AI Overview Brand Visibility.

This does not mean that publishing a YouTube video automatically creates AI visibility.

It demonstrates that authoritative source ecosystems increasingly extend beyond text-only content.

AI authority implication: Authority strategies may need to include multiple content formats capable of being discovered, interpreted and cited by generative systems.

Source:
Ahrefs — AI Overview Citation Study 2026

85. ChatGPT Ultimately Cited Only Around Half of the URLs Retrieved During Search

Ahrefs analysed approximately 1.4 million ChatGPT prompts to investigate why some retrieved pages were ultimately cited while others were not.

The study found that ChatGPT retrieved approximately:

• 16.57 URLs per prompt that were subsequently cited;
• 16.58 URLs per prompt that were retrieved but not cited.

In practical terms, only around half of the candidate URLs ultimately received visible citation credit.

This introduces an important distinction:

Being Retrieved
≠
Being Selected as a Citation

AI authority implication: Machine discoverability is only the first stage. The source must still be selected from the wider candidate pool.

Source:
Ahrefs — Why ChatGPT Cites One Page Over Another

86. ChatGPT’s General Search Results Had an 88.46% Citation Rate in Ahrefs’ Retrieval Study

Ahrefs identified several different retrieval channels inside the ChatGPT dataset.

URLs classified within the general search retrieval channel had a citation rate of 88.46%.

The comparable citation rates for specialist retrieval channels were substantially lower:

• News: 12.01%;
• Reddit: 1.93%;
• YouTube: 0.51%;
• Academia: 0.40%.

Ahrefs notes that Reddit and YouTube pages can also appear through the ordinary Search channel, so these specialist-category rates should not be interpreted as the overall citation rate for those domains.

AI authority implication: Conventional Search discoverability continues to play an important role inside at least some generative retrieval systems.

Source:
Ahrefs — ChatGPT Citation Selection Study 2026

87. 67.8% of ChatGPT’s Non-Cited Retrieval Pool Came From Reddit

One of the more unusual findings in Ahrefs’ ChatGPT study was that 67.8% of all retrieved URLs that were not ultimately cited came from Reddit.

The dedicated Reddit retrieval channel contained more than 16 million data points, yet only 1.93% were directly cited through that particular channel.

This suggests that a source can potentially contribute to:

• context;
• consensus;
• idea formation;
• or retrieval

without necessarily receiving visible attribution in the final answer.

AI authority implication: Visible citations represent only one part of the information AI systems may use while constructing an answer.

Source:
Ahrefs — ChatGPT Citation Selection Study 2026

88. 61.7% of AI Source Appearances Were “Ghost Citations” With No Brand Mention

Semrush and Growth Memo analysed 3,981 domain appearances across 115 prompts, 14 countries and four AI Search systems.

They found that 61.7% of appearances were “ghost citations”.

The AI system used the domain as a source link but did not explicitly name the brand within the generated answer.

Across the full dataset:

• 61.7% were citation-only appearances;
• 13.2% were both cited and mentioned;
• 25.1% were mentioned without a citation.

Overall, 74.9% of appearances included a citation, while only 38.3% included a Brand Mention.

AI authority implication: Citation visibility and Brand Visibility need to be measured separately. A website can provide the evidence while the brand remains largely invisible to the user.

Source:
Semrush / Growth Memo — Ghost Citations Study 2026

89. 45.5% of AI Overview Citations Changed Between Consecutive Observations

Ahrefs tracked more than 43,000 keywords, each with at least 16 recorded AI Overviews, across approximately one month.

It found that only 54.5% of citation URLs overlapped on average between consecutive observations.

That means approximately 45.5% of cited sources changed.

Ahrefs estimated this as roughly one citation changing each time the same AI Overview query was observed again.

The generated content itself changed from one observation to the next approximately 70% of the time.

AI authority implication: AI citation visibility is considerably more volatile than a conventional static backlink and should be measured repeatedly over time.

Source:
Ahrefs — AI Overviews Change Every Two Days

90. AI Mode and AI Overviews Shared Only 13.7% of Their Citation URLs

Ahrefs compared Google AI Mode and AI Overviews using approximately 540,000 query pairs for citation analysis and 730,000 pairs for broader response analysis.

The same URLs appeared in both systems only 13.7% of the time.

For the top three citations, overlap increased only slightly to 16.3%.

This means approximately 87% of cited URLs differed between AI Mode and AI Overviews for the same queries.

Yet the systems reached strongly similar conclusions: average semantic similarity was approximately 86%, with 89.7% of response pairs scoring above 0.8 for semantic alignment.

In other words:

Similar Answer
≠
Same Sources

AI authority implication: Citation visibility should be monitored independently across AI platforms and interfaces. Authority in one generative environment does not guarantee selection in another.

Source:
Ahrefs — AI Mode vs AI Overviews Citation Study

Statistics 91–100 — Link Building Performance, Measurement & Future Authority Trends

The final ten statistics examine how the measurement of link building and Digital PR is expanding beyond conventional backlink counts.

Links remain important, but organisations are increasingly monitoring a wider authority environment involving:

• quality links;
• Brand Mentions;
• organic visibility;
• AI citations;
• AI Share of Voice;
• and wider brand presence across the web.

The evidence below primarily combines BuzzStream’s 2026 State of Digital PR research with Ahrefs’ analysis of 75,000 brands across AI Search environments.

Practitioner survey findings should be interpreted as reported industry behaviour, while the Ahrefs figures represent statistical correlations rather than evidence that any one factor directly causes AI visibility.

91. 66.2% of Digital PR Professionals Say AI Citations Are an Effective Outcome of Digital PR

BuzzStream’s 2026 research found that 66.2% of Digital PR professionals identified getting mentioned in AI citations as an effective outcome of Digital PR.

Backlink building remained the most frequently selected outcome at 85.8%, but AI citation visibility appeared as a major new measurement category.

This demonstrates how the perceived role of Digital PR is widening from:

Media Coverage + Backlinks

toward:

Media Coverage + Backlinks + Brand Mentions + AI Citations

Future authority implication: Digital PR is increasingly being evaluated for its contribution to both Search Authority and generative Search visibility.

Source:
BuzzStream — State of Digital PR Report 2026

92. 55.4% Track AI Citation Mentions as a Digital PR Success Metric

More than half of the Digital PR professionals surveyed by BuzzStream said they now monitor AI citation mentions as part of campaign measurement.

The figure reached 55.4% in 2026.

For comparison, traditional metrics remained widely used:

• quality links: 85.1%;
• total mentions: 72.3%;
• total links: 70.9%;
• organic traffic and rankings: 63.5%.

Measurement implication: AI citation tracking is moving from an experimental measurement area toward a mainstream Digital PR KPI.

Source:
BuzzStream — State of Digital PR Report 2026

93. 78.4% of Digital PR Professionals Track AI Visibility for Clients

BuzzStream found that 78.4% of surveyed Digital PR professionals track AI visibility for their clients.

This is significantly broader than measuring direct AI citations alone.

AI visibility can include:

• Brand Mentions;
• citations;
• recommendations;
• Share of Voice;
• source appearances;
• and competitor visibility.

Measurement implication: Link-building and Digital PR reporting is beginning to incorporate visibility inside answer engines alongside traditional Search metrics.

Source:
BuzzStream — State of Digital PR Report 2026

94. Around 40% Say They Have a Repeatable Process for Earning AI Citations

BuzzStream reported that approximately four in ten Digital PR professionals said they had developed a repeatable way of helping brands earn citations from AI systems.

BuzzStream contrasted this with its separate Link Building Trends research, where only 11% of link builders reported having a repeatable AI-citation process.

The difference suggests that Digital PR teams may currently be adapting more quickly to citation-focused authority building than traditional link-building teams.

Future authority implication: AI citation acquisition is becoming a distinct operational discipline, although standardised methodologies are still developing.

Source:
BuzzStream — State of Digital PR Report 2026

95. 83.1% Say Digital PR Has Become More Important Because of AI

BuzzStream found that 83.1% of Digital PR professionals believe Digital PR has become more important during the previous 12 months in relation to AI.

This is a practitioner perception rather than a measured causal effect.

However, it demonstrates how strongly the industry is connecting earned media with the emerging AI visibility environment.

The underlying logic is that Digital PR can create:

• editorial links;
• Brand Mentions;
• expert references;
• research citations;
• and third-party corroboration.

Future authority implication: Activities historically associated with off-page SEO are increasingly being evaluated for their potential influence on AI discovery as well.

Source:
BuzzStream — State of Digital PR Report 2026

96. 75% Have Been Asked to Use Digital PR to Improve AI Citation Visibility

Three quarters of the Digital PR professionals surveyed by BuzzStream said they had been approached during the previous 12 months about using Digital PR to help brands appear in AI citations.

The statistic provides evidence of changing client demand.

Organisations are increasingly asking PR teams not only:

“Can this campaign earn coverage and backlinks?”

but also:

“Can this improve our visibility inside AI-generated answers?”

Future authority implication: AI visibility is becoming part of the commercial brief for Digital PR and off-page authority programmes.

Source:
BuzzStream — State of Digital PR Report 2026

97. Brand Web Mentions Had a 0.664 Correlation With AI Overview Brand Visibility

Ahrefs analysed 75,000 brands to examine which measurable factors were most strongly associated with Brand Visibility in Google AI Overviews.

The strongest relationship identified was the number of branded web mentions.

These produced a Spearman correlation of 0.664 with AI Overview Brand Visibility.

Importantly, the metric included both linked and unlinked Brand Mentions.

This was substantially stronger than the relationship observed for conventional backlink counts.

Authority implication: The wider frequency with which credible web content discusses a brand may be increasingly relevant to understanding AI visibility.

Source:
Ahrefs — AI Overview Brand Visibility Factors: 75,000 Brands Studied

98. Branded Anchor Mentions Had a 0.527 Correlation With AI Overview Visibility

Ahrefs found that branded anchors — hyperlinks where the visible anchor text contains the brand name — had a correlation of 0.527 with Brand Visibility in AI Overviews.

This was the second-strongest relationship identified in the study.

It sat below general Brand Mentions at 0.664 but above measures including:

• branded Search volume;
• Domain Rating;
• referring domains;
• and total backlinks.

Authority implication: Links can potentially provide both conventional link authority and explicit semantic information connecting a brand with the surrounding topic.

Source:
Ahrefs — AI Overview Brand Visibility Factors

99. Referring Domains Correlated More Strongly With AI Visibility Than Total Backlinks

Within the same 75,000-brand study, Ahrefs recorded a correlation of:

• 0.295 for referring domains;
• 0.218 for total backlinks.

Domain Rating recorded a correlation of 0.326.

All three link-related relationships were substantially weaker than the 0.664 correlation for Brand Mentions.

The result does not mean backlinks are unimportant to AI visibility.

It suggests that link volume alone provides only a partial view of the broader external authority environment.

Future authority implication: Referring-domain diversity continues to be useful, but organisations increasingly need to measure how extensively they are discussed and referenced across the web as well.

Source:
Ahrefs — AI Overview Brand Visibility Factors

100. Brands With the Most Web Mentions Earned Up to 10× More AI Overview Mentions

When Ahrefs grouped the 75,000 brands according to the number of web mentions they received, the brands in the highest-mention group generated up to ten times more Brand Mentions in AI Overviews than the next closest quartile.

The study also found that 26% of the analysed brands had zero Brand Mentions in AI Overviews.

Again, the study measures correlation rather than proving that creating additional web mentions will directly produce a proportional increase in AI visibility.

Nevertheless, it provides some of the strongest quantitative evidence currently available connecting broad third-party Brand Presence with generative Search visibility.

Future authority implication: The emerging authority objective is increasingly broader than acquiring hyperlinks. Organisations need to become recognised, discussed and corroborated within the wider information ecosystem.

Source:
Ahrefs — Analysis of AI Overview Brand Visibility Factors

Overall Analysis — What the 100 Link Building, Digital PR & Authority Statistics Tell Us

Taken together, the 100 statistics show that link building is becoming broader, more evidence-led and more closely connected to Brand Authority, Digital PR and AI Search visibility.

The evidence does not support the idea that backlinks have stopped mattering.

It supports a more nuanced conclusion:

Links Still Matter — But the Context Around the Link Matters More Than Ever.

The strongest authority strategies increasingly combine:

• relevant backlinks;
• diverse referring domains;
• editorial coverage;
• original research;
• Brand Mentions;
• entity signals;
• credible source citations;
• and visibility across both Search and AI systems.

1. Backlinks Remain Associated With Search Visibility

The ranking studies examined in this research continue to show measurable relationships between backlink strength and higher Google rankings.

Top-ranking pages tend to have:

• more backlinks;
• more referring domains;
• greater referring-domain diversity;
• and stronger external authority profiles.

However, the measured correlations are not strong enough to justify treating links as the only ranking factor.

That is important.

The evidence supports:

Links as Part of Search Authority

rather than:

Links as the Entire Ranking System

2. Referring-Domain Diversity Matters More Than Raw Link Volume Alone

Several datasets point toward the importance of earning links from multiple independent domains.

The number of referring domains often showed a slightly stronger relationship with Search performance than total backlink volume.

This makes intuitive sense.

One website linking 100 times represents a different external evidence pattern from 100 independent websites each choosing to reference the same organisation.

The stronger authority question is therefore:

“How Many Independent Sources Consider This Organisation Worth Referencing?”

3. Backlink Quality Cannot Be Reduced to One Metric

The evidence around anchor text, page traffic, Domain Rating and raw link counts demonstrates why no single backlink metric provides a complete assessment of quality.

Useful evaluation increasingly requires considering:

• topical relevance;
• editorial context;
• referring-domain diversity;
• source credibility;
• placement;
• audience value;
• and the relationship between the linking page and the linked resource.

This makes simplistic strategies based on:

• Domain Rating thresholds;
• exact-match anchor percentages;
• or raw backlink totals

increasingly inadequate.

4. Most Outreach Fails

The outreach studies are a reminder that link acquisition is difficult.

Most cold outreach emails receive no response.

The data consistently shows that performance improves when campaigns become more relevant and more structured.

The strongest variables include:

• target selection;
• personalisation;
• subject-line relevance;
• contact selection;
• and appropriate follow-up.

The important strategic lesson is:

More Emails Do Not Automatically Produce Better Link Building.

The quality of the proposition matters.

5. Digital PR Works Best When the Story Is Worth Publishing

The Digital PR evidence shows that journalists continue to use PR pitches as a source of story ideas.

At the same time, they overwhelmingly ignore material outside their area of coverage.

This creates a simple rule:

Journalistic Relevance Comes Before Link Acquisition.

The link is usually the consequence of editorial value.

It is not the reason a journalist publishes the story.

6. Most Content Never Earns a Backlink

The content studies are among the clearest findings in the research.

Across datasets containing hundreds of millions of articles, the overwhelming majority of published content received no external links.

This means publishing more ordinary content is not automatically a link-building strategy.

The formats with stronger link potential tend to provide something another publisher can reference.

Examples include:

• new research;
• statistics;
• original data;
• benchmark reports;
• visualisations;
• tools;
• and authoritative reference resources.

The distinction is between:

Content Designed to Be Read

and:

Content Designed to Become a Source

7. Link Building Is an Investment, Not a Free Outcome

Earned links may not be purchased directly, but credible link acquisition still requires investment.

That investment can include:

• research;
• surveys;
• data licensing;
• analysis;
• copywriting;
• design;
• journalist databases;
• PR expertise;
• and campaign management.

This is why cost per link can be misleading.

The actual investment may generate far more than the backlink itself.

One campaign can create:

• links;
• coverage;
• Brand Mentions;
• referral traffic;
• journalist relationships;
• Search visibility;
• and AI citation visibility.

8. Authority Decays Unless It Is Maintained

The link-rot data demonstrates that a backlink profile does not remain intact indefinitely.

Pages disappear.

Domains close.

Links are removed.

URLs are redirected.

This means authority building has two components:

Authority Acquisition
+
Authority Preservation

Organisations therefore need to monitor both:

• new links gained;
• and existing links lost.

This is particularly important for research resources that may accumulate citations over several years.

9. Authority Is Broader Than Backlinks

The review, Brand Mention and entity evidence demonstrates that organisations are evaluated through multiple external information sources.

Consumers already move across:

• Search;
• reviews;
• social platforms;
• business directories;
• media coverage;
• and AI systems.

They look for consistency and corroboration.

The same principle increasingly matters to machine understanding.

An organisation repeatedly and consistently referenced across credible sources creates a stronger external evidence network than one visible only through its own website.

10. AI Search Is Expanding the Definition of Authority

The AI citation studies add a new layer to the authority model.

AI systems can:

• retrieve a source;
• use it without citing it;
• cite it without mentioning the brand;
• mention a brand without citing its website;
• or recommend the organisation directly.

This means several separate measurements are now required:

• machine discoverability;
• retrieval visibility;
• citation visibility;
• Brand Mentions;
• recommendation visibility;
• and AI Share of Voice.

The evidence also shows that AI citations are dynamic.

Sources can change between repeated generations of the same query.

Different AI systems can reach similar conclusions using substantially different sources.

The resulting authority environment is therefore more fluid than traditional backlink measurement.

100 Link Building, Digital PR & Authority Statistics at a Glance

The table below summarises the 100 statistics examined throughout this research.

The evidence combines ranking studies, backlink datasets, outreach research, Digital PR surveys, journalist research, web-decay analysis, consumer authority data and emerging AI citation studies.

UK-specific evidence has been used where available. International and US datasets are clearly identified and should be interpreted as industry or behavioural benchmarks rather than direct measurements of the UK market.

CGO Link & Authority Framework

The 100 statistics demonstrate that modern link building should no longer be understood as a process whose sole objective is acquiring hyperlinks.

The stronger model is based on creating information, evidence and expertise that independent sources consider worth referencing.

The CGO Link & Authority Framework maps that process across seven stages:

1. Research Asset Creation
2. Media & Publisher Discovery
3. Editorial Mention
4. Link & Citation Acquisition
5. Entity Reinforcement
6. Authority Accumulation
7. Search & AI Visibility

1. Research Asset Creation

Authority building begins with something worth referencing.

This may include:

• original research;
• statistics;
• consumer surveys;
• benchmark reports;
• proprietary datasets;
• expert analysis;
• interactive tools;
• calculators;
• industry indexes;
• visualisations;
• or a genuinely useful reference resource.

The central principle is that the organisation contributes information that does not simply repeat what is already available elsewhere.

A strong authority asset should ideally answer one or more of the following questions:

• What new evidence can we provide?
• What can we measure that others have not measured?
• What question can we answer better than existing sources?
• What information would a journalist, researcher or industry writer need to cite?
• What could become a recurring reference within our sector?

The asset is therefore not created only to rank.

It is created to become a source.

Framework question: What does the organisation possess or know that credible third parties would have a legitimate reason to reference?

2. Media & Publisher Discovery

Creating a strong asset does not automatically generate external authority.

The relevant publishers need to discover it.

Potential audiences can include:

• journalists;
• trade publications;
• researchers;
• industry analysts;
• bloggers;
• professional associations;
• universities;
• government organisations;
• business publications;
• and specialist websites.

Discovery can occur organically through Search, but it can also be accelerated through Digital PR and targeted outreach.

The statistical evidence in this research shows why relevance matters.

Most outreach receives no response, and journalists overwhelmingly ignore pitches outside their area of coverage.

The objective is therefore not maximum distribution.

It is relevant distribution.

Framework question: Which publications and people already write about the subject covered by the asset?

3. Editorial Mention

The first form of external validation may not be a backlink.

A journalist or publisher may:

• name the organisation;
• quote its research;
• reference a statistic;
• mention an expert;
• include the brand in analysis;
• or discuss the findings without providing a hyperlink.

This still creates external evidence that the organisation is connected with a particular topic.

An editorial mention can therefore contribute to:

• Brand Awareness;
• topic association;
• entity recognition;
• reputation;
• and future discoverability.

The research examined earlier also suggests that Brand Mentions can be associated with AI visibility even when a conventional backlink is absent.

Framework question: Is the organisation being discussed by credible external sources in connection with the topics it wants to own?

4. Link & Citation Acquisition

The next stage is the explicit reference.

This can take several forms:

• editorial backlink;
• research citation;
• data-source attribution;
• reference link;
• journalistic source link;
• or AI citation.

These should not all be treated as identical.

A traditional backlink can:

• send referral traffic;
• create a persistent connection between two webpages;
• contribute to Search Authority;
• and reinforce topical relationships.

An AI citation is different.

It may be:

• dynamic;
• query-specific;
• temporary;
• and subject to change between answer generations.

Nevertheless, both forms of citation indicate that an external information system has selected the organisation’s content as relevant supporting evidence.

Framework question: Which independent sources are explicitly linking to or citing the organisation’s evidence?

5. Entity Reinforcement

Repeated third-party references can strengthen the digital evidence surrounding an organisation as an entity.

For example, external sources may repeatedly associate an organisation with:

• a particular industry;
• a specialist topic;
• research expertise;
• a location;
• a product category;
• a methodology;
• or an individual expert.

When these relationships are consistent across multiple independent sources, the organisation becomes easier to understand within the wider information ecosystem.

The pattern can be represented as:

Organisation
+
Topic
+
Independent References
+
Consistent Context
=
Stronger Entity Association

This does not mean that every external mention has equal value.

Authority depends on the quality, credibility and relevance of the source.

Framework question: Are independent sources consistently reinforcing the same understanding of who the organisation is and what it is authoritative about?

6. Authority Accumulation

Authority develops cumulatively.

A single backlink can be useful.

A single article can provide visibility.

A single Brand Mention can create awareness.

But a sustained pattern of independent references creates a much stronger evidence base.

This wider authority can include:

• Link Authority — backlinks and referring domains;
• Editorial Authority — coverage in credible publications;
• Brand Authority — repeated external Brand Mentions;
• Entity Authority — consistent associations around the organisation;
• Content Authority — evidence that its resources are repeatedly referenced;
• Citation Authority — selection as a source by Search and AI systems.

The organisation is therefore no longer measuring one isolated signal.

It is measuring an authority network.

Framework question: Is external evidence becoming broader, stronger and more diverse over time?

7. Search & AI Visibility

The final stage is visibility.

Authority can contribute to several discovery environments:

• traditional Google Search;
• Google AI Overviews;
• Google AI Mode;
• ChatGPT;
• Gemini;
• Copilot;
• Perplexity;
• and other generative Search systems.

The mechanisms are different.

A backlink may contribute to conventional Search visibility.

A source citation may contribute to AI visibility.

A Brand Mention may reinforce broader entity recognition.

Media coverage may contribute to all three.

The strongest outcome therefore occurs when several forms of authority reinforce one another.

Framework question: Does growing external authority correspond with stronger presence across both Search results and AI-generated answers?

Authority Must Also Be Maintained

The framework is not complete without preservation.

The link-decay research reviewed earlier shows that external authority can deteriorate over time.

Organisations therefore need an ongoing maintenance cycle:

Earn
↓
Monitor
↓
Preserve
↓
Reclaim
↓
Refresh
↓
Earn Again

This is particularly important for research pages, statistics resources and reference assets that can accumulate links over many years.

Where possible, important source URLs should be preserved and updated rather than unnecessarily replaced.

From Campaign Thinking to System Thinking

Traditional link building is frequently organised around individual campaigns.

For example:

Campaign → Outreach → Links → Finish

The CGO Link & Authority Framework uses a longer-term model:

Research
↓
Publication
↓
Distribution
↓
Coverage
↓
Links & Mentions
↓
Authority
↓
Visibility
↓
New Research

This turns authority building into a continuous system rather than a sequence of disconnected link campaigns.

How to Measure Link Building & Authority

Traditional link-building reporting often focuses on the number of backlinks acquired.

The 100 statistics examined in this research show why that is no longer sufficient.

Modern authority measurement needs to consider the complete external evidence environment around an organisation, including:

• referring domains;
• link quality;
• editorial relevance;
• Brand Mentions;
• media coverage;
• citation visibility;
• link retention;
• Search performance;
• AI visibility;
• and commercial outcomes.

The objective is to move from:

“How Many Links Did We Build?”

toward:

“How Much Independent Authority Did We Create — and What Did It Change?”

1. Referring Domain Growth

Referring-domain growth measures how many new independent websites begin linking to the organisation.

This is generally more useful than monitoring total backlink count alone because multiple links from one domain do not represent the same diversity of external validation as links from multiple independent sources.

This metric should be segmented by:

• authority;
• topical relevance;
• geography;
• publisher type;
• and campaign source.

2. Link Quality Distribution

Not every backlink should be counted as equivalent.

A useful reporting model groups earned links according to quality characteristics rather than presenting one undifferentiated total.

Possible dimensions include:

• editorial relevance;
• publisher credibility;
• page relevance;
• referring-domain strength;
• placement within the article;
• audience relevance;
• and whether the link was genuinely editorially earned.

Rather than relying on one third-party authority metric, organisations can create internal quality bands.

For example:

• Tier 1: highly authoritative and directly relevant editorial sources;
• Tier 2: credible industry or specialist sources;
• Tier 3: relevant supporting sources;
• Monitor: low-value, questionable or irrelevant links.

3. Editorial Placement Rate

A response to outreach is not the same as editorial coverage.

Editorial Placement Rate measures how frequently outreach results in actual publication.

This can be further separated into:

• linked coverage;
• unlinked Brand Mentions;
• quotes;
• data citations;
• and feature coverage.

4. Link Conversion Rate

Link Conversion Rate measures how frequently earned coverage includes a backlink.

This helps distinguish:

• campaigns that generate awareness;
• campaigns that generate links;
• and campaigns that generate both.

5. Brand Mention Growth

Unlinked Brand Mentions should be measured separately from backlinks.

An organisation may be referenced by:

• journalists;
• industry publications;
• researchers;
• review platforms;
• directories;
• and other third parties

without receiving a hyperlink.

Mentions should ideally be classified according to:

• linked vs unlinked;
• positive, neutral or negative context;
• topic;
• publisher quality;
• and geography.

6. Citation Visibility

AI Search introduces another measurable authority layer.

Citation Visibility measures how often an organisation’s website or research appears as a source in relevant AI-generated answers.

This should be measured separately from:

• Brand Mentions;
• recommendations;
• and general AI visibility.

A source can be cited without the brand being named.

7. Authority Retention Rate

Because links disappear over time, organisations should measure how much of their earned authority remains active.

This can be monitored over:

• three months;
• six months;
• 12 months;
• and longer periods.

Lost links should be categorised according to cause:

• link removed;
• page deleted;
• domain disappeared;
• URL changed;
• redirected page;
• or technical issue.

8. Link Reclamation Rate

Some lost links can be recovered.

Link Reclamation Rate measures how successful the organisation is at restoring links that have disappeared but remain potentially recoverable.

Reclamation can involve:

• correcting broken destination URLs;
• implementing redirects;
• contacting publishers;
• restoring removed source content;
• or updating obsolete references.

9. Search Visibility Lift

Authority campaigns should ultimately be connected with changes in organic Search performance.

Potential measurements include:

• ranking improvements;
• organic impressions;
• organic clicks;
• non-brand visibility;
• topic-level visibility;
• and organic traffic.

Changes should be interpreted carefully because Search performance is affected by many factors beyond links.

These include:

• content changes;
• technical SEO;
• Google updates;
• competitor activity;
• Search demand;
• and seasonality.

10. AI Visibility Lift

AI Visibility measures whether the organisation appears more frequently across relevant generative Search prompts after authority-building activity.

This can include changes in:

• Brand Mentions;
• source citations;
• recommendation visibility;
• Share of Voice;
• and cross-platform visibility.

11. Referral Traffic From Earned Media

Links can generate direct visits in addition to potential Search Authority.

Referral traffic should be measured by:

• publisher;
• campaign;
• landing page;
• engagement;
• conversion;
• and revenue where available.

12. Branded Search Lift

Digital PR can also increase demand for the brand itself.

Users exposed to editorial coverage may later search directly for:

• the company name;
• the founder;
• a product;
• a research study;
• or a framework.

Branded Search Lift should not automatically be attributed entirely to Digital PR because advertising, social media and other activities can influence the same metric.

13. Commercial Contribution

Authority activity should ultimately be connected to business outcomes where attribution is possible.

Possible measurements include:

• leads;
• enquiries;
• sales;
• bookings;
• pipeline;
• revenue;
• and customer acquisition.

This may involve:

• referral attribution;
• assisted-conversion analysis;
• CRM data;
• customer surveys;
• and campaign-level attribution.

Recommended Link & Authority Dashboard

Measure by Campaign Type

Not every authority campaign should be judged using identical KPIs.

A research-led Digital PR campaign may prioritise:

• editorial citations;
• high-authority referring domains;
• media coverage;
• and AI citation visibility.

A broken-link reclamation programme may prioritise:

• recovered links;
• restored referring domains;
• and preserved authority.

A journalist-commentary programme may prioritise:

• expert mentions;
• linked quotes;
• Brand Mentions;
• and repeated publisher relationships.

Measurement should therefore reflect the purpose of the campaign.

Measure by Authority Type

A mature authority dashboard should separate several forms of external evidence.

These include:

• Link Authority — backlinks and referring domains;
• Editorial Authority — independent media coverage;
• Brand Authority — external Brand Mentions;
• Entity Authority — consistent topic and organisational associations;
• Content Authority — repeated citation of specific resources;
• AI Citation Authority — selection as a source in generative answers.

These areas overlap but should not be treated as identical.

Measure Quality and Quantity Separately

A campaign can generate many links while producing little meaningful authority.

Another campaign may generate only a handful of links from highly authoritative and relevant sources.

Reporting should therefore separate:

Volume
from
Value

At minimum, dashboards should show:

• total links earned;
• unique referring domains;
• quality distribution;
• topical relevance;
• and high-value placements.

Measure New Authority Against Lost Authority

Reporting only new links can create an inaccurate impression of progress.

If a campaign earns 30 new referring domains but loses 25 existing referring domains, the net authority growth is very different from a campaign earning 30 while losing only two.

The relevant metric is therefore:

New Authority − Lost Authority = Net Authority Growth

Measure Over Time

Authority should be tracked longitudinally.

Useful reporting intervals can include:

• monthly campaign activity;
• quarterly authority growth;
• six-month Search impact;
• annual link-retention analysis;
• and long-term AI citation visibility.

The objective is to distinguish short-term campaign spikes from sustained authority growth.

Research Methodology

This research was developed to provide a structured evidence base for understanding link building, Digital PR, Brand Authority and emerging AI citation behaviour in 2026.

The final dataset contains 100 quantitative findings drawn from published research, large-scale Search datasets, outreach studies, practitioner surveys, journalist surveys, consumer research, web-decay analysis and AI citation studies.

The objective was not to prove that any single metric determines Search or AI visibility.

Instead, the research examines how multiple forms of external evidence contribute to the wider authority environment around organisations and their content.

Research Scope

The study covers ten principal evidence areas:

1. Backlinks and Google rankings;
2. Link Authority and relevance;
3. Link acquisition and outreach;
4. Digital PR;
5. Content that earns links;
6. Link-building costs and investment;
7. Link loss and web decay;
8. Brand Mentions and Entity Authority;
9. AI citations and Source Authority;
10. Link-building performance and future authority measurement.

These categories were selected because modern authority building increasingly sits across the intersection of SEO, Digital PR, research publishing, brand development and AI Search.

Source Selection

Sources were selected according to a hierarchy designed to favour substantial datasets and directly published research.

Priority was given to:

• large-scale empirical studies;
• original research published by the organisation that conducted the analysis;
• recognised Search and Digital PR research platforms;
• journalist and practitioner surveys with disclosed sample sizes;
• consumer research with documented panels;
• and studies providing enough methodological detail to interpret the results responsibly.

Where possible, the research cites the original study rather than secondary commentary describing it.

Primary Research Organisations Used

The evidence base includes research published by organisations including:

• Ahrefs;
• Backlinko;
• BuzzStream;
• BuzzSumo;
• Semrush;
• Muck Rack;
• BrightLocal;
• Pew Research Center;
• Pitchbox;
• Majestic;
• and Growth Memo.

The inclusion of a source does not imply that CGO Media endorses every methodology, metric or conclusion published by that organisation.

Each statistic is used for the specific evidence it contributes to this research.

UK Evidence and International Evidence

This research is published as Link Building Statistics UK 2026, but not every large-scale study used in the evidence base was conducted exclusively in the United Kingdom.

UK-specific evidence was prioritised where credible data was available.

For example, the Digital PR pricing study used in this research had a participant base in which 54% of respondents were UK-based.

Other datasets were based on:

• US Search results;
• international Digital PR professionals;
• US consumer panels;
• multi-country AI Search datasets;
• or global web indexes.

Where a dataset is not UK-specific, it should be interpreted as an industry benchmark, platform benchmark or behavioural indicator rather than a direct estimate of the UK population.

Use of Historical Research

Several important studies included in the dataset were published before 2026.

They were retained where they met one or more of the following conditions:

• the dataset was exceptionally large;
• the study examined a fundamental characteristic of the web;
• no newer equivalent dataset offered the same scale;
• or the finding remained useful as a historical benchmark.

Examples include large-scale analyses of hundreds of millions of articles and outreach datasets containing millions of emails.

Older evidence is not presented as though it describes the precise state of the market in 2026.

Its publication period should be considered when interpreting the finding.

Ranking Correlation Studies

Several statistics in this research are based on correlations between backlink metrics and Google rankings.

Correlation must not be interpreted as proof of causation.

For example:

More Referring Domains
↕
Higher Rankings

does not automatically prove:

More Referring Domains
→
Guaranteed Higher Rankings

Higher-quality pages may attract more links because they already perform well.

Strong brands may simultaneously have:

• more backlinks;
• better content;
• greater Search demand;
• stronger user signals;
• better technical infrastructure;
• and greater commercial authority.

Correlation studies are therefore used to identify relationships rather than prove that one variable independently causes another.

Practitioner Survey Data

Some figures in this research come from surveys of Digital PR and link-building professionals.

These datasets are useful for understanding:

• budgets;
• costs;
• workflow;
• measurement practices;
• campaign timescales;
• and changing industry priorities.

However, practitioner surveys measure reported experience and opinion.

They should not be interpreted as independent measurements of:

• Google ranking algorithms;
• AI model architecture;
• or guaranteed campaign performance.

Journalist Survey Data

Journalist research was included to examine the receiving side of Digital PR.

This is important because link-building research often evaluates the outreach process only from the perspective of the sender.

Journalist survey data helps provide evidence about:

• pitch relevance;
• preferred pitch length;
• the role of PR pitches in story generation;
• and the volume of material journalists receive.

These findings provide context for why targeted media outreach can perform differently from mass distribution.

Consumer Authority Data

Consumer review research was included because authority is increasingly distributed across sources beyond conventional backlinks.

Review platforms can provide external evidence about:

• business identity;
• services;
• locations;
• reputation;
• customer experience;
• and recency.

The BrightLocal consumer research used within the 100-statistic dataset was conducted with a representative panel of US adults.

It is therefore treated as a consumer-behaviour benchmark rather than a direct estimate of UK consumer behaviour.

AI Citation Research

AI Search evidence requires particular caution because generative systems are changing rapidly.

The research distinguishes between several different concepts:

• retrieval — a system accesses or considers a source;
• citation — the source receives visible attribution;
• Brand Mention — the organisation is named in the generated answer;
• recommendation — the organisation is explicitly suggested to the user;
• AI visibility — the wider frequency with which the organisation appears across relevant prompts.

These events should not be treated as equivalent.

A page may be retrieved without receiving a visible citation.

A domain may be cited without the brand being named.

A brand may be mentioned without its own website being cited.

AI Citation Volatility

Unlike a conventional backlink, an AI citation is not necessarily persistent.

The same prompt can generate different:

• wording;
• sources;
• citations;
• Brand Mentions;
• and recommendations

across different generations or platforms.

AI visibility therefore requires repeated observation rather than a single one-time check.

Any AI citation statistic should be interpreted in the context of:

• the model or interface studied;
• the date of data collection;
• the query sample;
• the geography;
• and the methodology used by the original researcher.

Selection of the 100 Statistics

Potential statistics were assessed against five criteria.

Avoiding Duplicate Evidence

Where one study produced several distinct findings, individual statistics were retained only when they measured meaningfully different variables.

For example, the research may separately report:

• total backlinks;
• referring domains;
• followed backlinks;
• and Brand Mentions

because each represents a different measurement.

Repeated presentation of the same percentage using slightly different wording was avoided where possible.

Source Dates

The title of this research contains the year 2026.

That year identifies the publication and analytical context of the CGO Media research.

It does not mean that all 100 underlying studies were published in 2026.

The dataset combines:

• recent 2025–2026 evidence;
• earlier large-scale benchmark studies;
• and historical evidence used specifically to examine long-term patterns.

Readers should consider the original publication date of each source when applying an individual statistic.

Rounding

Where an original study reports decimal values, those values have generally been retained.

Where wording such as:

• approximately;
• around;
• more than;
• or fewer than

appears in the original evidence or is necessary to avoid false precision, the same cautious wording is used here.

Currency

Digital PR cost statistics are presented in the currency used by the original source.

Where BuzzStream reported pricing in US dollars, CGO Media has retained those figures in dollars rather than converting them into pounds sterling using a temporary exchange rate.

This preserves comparability with the original dataset.

Interpretation of Commercial Impact

The presence of links, Brand Mentions or AI citations does not automatically prove a direct commercial outcome.

Revenue can be influenced by:

• brand strength;
• pricing;
• product quality;
• sales processes;
• website conversion;
• market demand;
• advertising;
• and many other variables.

Where possible, organisations should therefore connect authority metrics with:

• Search visibility;
• referral traffic;
• branded demand;
• leads;
• pipeline;
• and revenue

without automatically attributing every downstream change to link-building activity.

Research Limitations

The 100 statistics provide a broad evidence base for understanding link building, Digital PR, Brand Authority and AI citation visibility, but the dataset has important limitations.

These limitations should be considered when applying individual findings to a specific organisation, industry or Search environment.

1. The Research Is Not Based Exclusively on UK Data

This research is published for a UK audience, but there is not enough high-quality UK-specific evidence to support 100 substantial statistics covering every area examined.

The dataset therefore combines:

• UK evidence;
• US Search datasets;
• US consumer surveys;
• international practitioner surveys;
• global backlink indexes;
• and multi-country AI Search studies.

International findings can provide useful benchmarks, but they should not automatically be assumed to represent UK behaviour.

Differences may exist in:

• media markets;
• consumer behaviour;
• publisher economics;
• Digital PR practices;
• industry competition;
• and Search behaviour.

Where a dataset is not UK-specific, it is used as a comparative industry benchmark rather than presented as a direct UK population estimate.

2. Several Important Sources Account for Multiple Statistics

Some organisations appear repeatedly throughout the dataset because they have published unusually large or detailed studies.

These include:

• Ahrefs;
• BuzzStream;
• Backlinko;
• BuzzSumo;
• Semrush;
• BrightLocal;
• and Pew Research Center.

This creates a degree of source concentration.

For example, multiple AI citation statistics may originate from different analyses conducted using the same underlying Search intelligence platform.

The statistics remain separate because they measure different variables, but they should not be mistaken for 100 completely independent research programmes.

3. Third-Party SEO Databases Do Not Contain the Entire Web

Backlink platforms build their own indexes of the web.

These indexes differ according to:

• crawler coverage;
• crawl frequency;
• URL discovery;
• data processing;
• link classification;
• and index-retention policies.

The number of backlinks or referring domains reported by one platform can therefore differ from another platform examining the same website.

Metrics such as:

• Domain Rating;
• Domain Authority;
• Authority Score;
• Trust Flow;
• and similar third-party measures

are proprietary metrics rather than Google ranking scores.

4. Correlation Does Not Establish Causation

Several studies within the dataset identify statistical relationships between backlink characteristics and Search rankings.

Those relationships do not prove that backlinks alone caused the ranking position.

A page with many links may also have:

• better content;
• stronger brand recognition;
• greater Search demand;
• better technical SEO;
• higher user engagement;
• and greater historical authority.

The same limitation applies to emerging AI visibility studies.

A strong correlation between Brand Mentions and AI Overview visibility does not prove that increasing Brand Mentions alone will directly produce more AI mentions.

5. Google Does Not Publish a Complete Link-Weighting Formula

Google provides guidance about links and link spam, but it does not publish a complete formula explaining exactly how every backlink is valued within ranking systems.

Third-party studies therefore observe Search outcomes and attempt to identify relationships within the available data.

They cannot independently determine the precise internal weighting Google assigns to:

• individual backlinks;
• anchor text;
• link placement;
• source relevance;
• referring-domain authority;
• or combinations of these signals.

Link studies should therefore be interpreted as observational evidence rather than a reverse-engineered version of Google’s ranking algorithm.

6. Digital PR Survey Data Is Self-Reported

Several statistics relating to:

• Digital PR effectiveness;
• cost per link;
• monthly budgets;
• campaign timescales;
• AI citation tracking;
• and link output

come from practitioner surveys.

These respondents report their own experiences.

Self-reported data can be affected by:

• different interpretations of questions;
• incomplete internal measurement;
• recall bias;
• commercial incentives;
• different definitions of a successful link;
• and varying client portfolios.

Practitioner survey results are therefore most useful for identifying industry patterns rather than establishing universal performance benchmarks.

7. Cost per Link Is Particularly Difficult to Standardise

The reported cost of earning a link can vary dramatically according to what expenditure is included.

For example, one organisation may include:

• research;
• survey costs;
• data acquisition;
• creative production;
• PR salaries;
• software;
• and outreach time.

Another may divide only agency fees by the number of links secured.

The resulting cost-per-link figures are therefore not perfectly comparable.

A link from a major national publication also cannot reasonably be treated as economically identical to a link from a small specialist website simply because both are counted as one backlink.

8. Historical Content Studies May Not Reflect the Exact 2026 Web

Some of the largest content studies used in this research were conducted several years before 2026.

They remain valuable because of their scale, including datasets containing:

• 100 million articles;
• 912 million blog posts;
• and millions of outreach emails.

However, the web has changed since those studies were published.

Changes include:

• growth in AI-generated content;
• changes to social platforms;
• new Search features;
• publisher consolidation;
• changes in Digital PR practices;
• and the emergence of generative Search.

Historical findings are therefore used as large-scale benchmarks rather than assumed to describe every characteristic of content publishing in 2026.

9. Journalist Surveys Do Not Represent Every Journalist

Journalist research provides useful insight into pitching preferences, but individual journalists differ substantially.

Preferences can vary according to:

• publication;
• sector;
• country;
• seniority;
• news cycle;
• working pattern;
• and the type of story being pitched.

A journalist covering breaking financial news may behave very differently from a journalist preparing a long-form travel feature.

Survey averages should therefore inform outreach strategy without replacing research into individual journalists.

10. Consumer Review Evidence Is Not the Same as Backlink Evidence

Review statistics were included to illustrate the broader external evidence environment surrounding organisations.

They should not be interpreted as evidence that reviews function in the same way as backlinks.

Reviews primarily provide evidence relating to:

• reputation;
• consumer trust;
• business identity;
• freshness;
• and third-party corroboration.

Backlinks, reviews, media mentions and AI citations are different signals and should be measured separately.

11. AI Search Systems Are Changing Rapidly

AI Search is the fastest-changing area covered by this research.

The behaviour of:

• Google AI Overviews;
• Google AI Mode;
• ChatGPT;
• Gemini;
• Copilot;
• Perplexity;
• and other AI systems

can change as models, retrieval systems, indexes and product interfaces are updated.

A citation pattern measured in one month may not remain identical several months later.

For that reason, AI citation statistics should be considered time-specific observations rather than permanent rules.

12. AI Outputs Are Non-Deterministic

A traditional backlink is relatively straightforward to verify.

A link either appears on a page or it does not.

Generative AI outputs are more variable.

The same prompt may produce different:

• answers;
• sources;
• citations;
• Brand Mentions;
• and recommendations.

Results can also differ according to:

• location;
• model version;
• logged-in state;
• personalisation;
• query wording;
• and time of observation.

A single prompt check should therefore never be treated as a reliable measurement of overall AI visibility.

13. Retrieval Cannot Always Be Observed Directly

Generative systems may retrieve, process or use information without exposing every source involved in constructing the final answer.

This creates an important measurement gap.

The visible citation set may not represent the complete information environment considered by the system.

This is particularly relevant to the distinction between:

Retrieved
↓
Used
↓
Cited
↓
Mentioned
↓
Recommended

These are separate stages and cannot always be fully observed externally.

14. AI Citation Studies Depend on Prompt Selection

AI visibility research is heavily influenced by the prompts chosen for analysis.

A dataset focused on:

• commercial recommendations;
• informational questions;
• product research;
• local Search;
• health queries;
• or B2B topics

may produce very different citation patterns.

Results from one query set should therefore not automatically be generalised to every industry or Search intent.

15. Source Authority Can Vary by Topic

A domain that is highly authoritative for one subject may be less relevant for another.

For example:

• a financial regulator may be authoritative for regulatory information;
• a university may be authoritative for academic research;
• a specialist trade publication may be authoritative within a narrow industry;
• a national newspaper may provide broad editorial authority.

Authority should therefore be interpreted in context rather than reduced to a universal domain score.

16. Link Quantity Does Not Equal Link Quality

Several statistics in the dataset report link volumes.

Those numbers should not imply that every backlink contributes equal value.

Links can differ according to:

• editorial quality;
• relevance;
• source credibility;
• placement;
• traffic;
• link attributes;
• and the reason the link exists.

A campaign producing fewer but highly relevant editorial references may create more useful authority than one generating a much larger number of weak links.

17. Brand Mentions Are Not Automatically Positive Signals

A Brand Mention can occur within:

• positive coverage;
• neutral reporting;
• criticism;
• complaints;
• controversy;
• or irrelevant contexts.

Counting Brand Mentions without considering context can therefore create a misleading picture of authority.

Mention analysis should ideally include:

• source quality;
• sentiment;
• topic;
• prominence;
• and relevance.

18. Link Acquisition and Commercial Performance Are Separated by Multiple Variables

A backlink can contribute to visibility without directly producing a sale.

Between authority acquisition and commercial conversion sit many additional variables:

Authority
↓
Visibility
↓
Discovery
↓
Website Experience
↓
Trust
↓
Conversion
↓
Revenue

A poor product, weak landing page or ineffective sales process can prevent strong authority from producing commercial results.

Link-building ROI should therefore be interpreted within the wider customer journey.

19. Search Algorithms and AI Systems Can Change After Publication

This research reflects evidence available during the 2026 publication period.

Future changes to:

• Google ranking systems;
• spam policies;
• AI retrieval architecture;
• publisher behaviour;
• Search interfaces;
• and generative models

may change the relationships described in this study.

The page should therefore be treated as a living research resource and updated when sufficiently strong new evidence becomes available.

20. The Framework Is an Analytical Model, Not a Search Engine Specification

The CGO Link & Authority Framework developed from this research is intended to organise the evidence into a useful strategic model.

It should not be interpreted as a description of Google’s internal algorithm or the architecture of any AI model.

The framework describes an observable authority journey:

Evidence Creation
↓
External Discovery
↓
Independent Reference
↓
Links & Citations
↓
Authority Accumulation
↓
Search & AI Visibility

It is an analytical framework derived from the research, not a claim about undisclosed ranking mechanisms.

Related CGO Media Research

Link building sits within a much wider system of Digital Authority.

The 100 statistics in this study connect directly with CGO Media research examining Digital PR, Brand Authority, Entity Authority, Content Authority, source selection, citation selection and AI recommendation systems.

The following research papers and frameworks provide deeper analysis of the individual authority mechanisms discussed throughout this study.

Supporting Authority Frameworks

The wider CGO Media Framework Library contains additional models that connect directly with link and authority development.

Frequently Asked Questions — Link Building, Digital PR, Backlinks & AI Authority

The following questions summarise the principal findings from the 100 statistics and address how link building is changing as traditional Search, Digital PR, Brand Authority and AI discovery increasingly overlap.

Do backlinks still matter for SEO in 2026?

Yes. The large-scale ranking studies examined in this research continue to show measurable relationships between backlinks, referring domains and higher Search rankings.

However, backlinks should not be interpreted as the only factor determining rankings.

Modern Search visibility also depends on factors including:

• content quality;
• technical accessibility;
• Search intent alignment;
• Brand Authority;
• entity understanding;
• user demand;
• and wider site quality.

The stronger interpretation is that backlinks remain an important component of Search Authority rather than a standalone ranking system.

Are referring domains more important than total backlink numbers?

In several studies examined here, referring-domain diversity showed a slightly stronger relationship with rankings than total backlink volume.

This reflects an important difference between:

100 Links From One Website

and
100 Independent Websites Choosing to Link

Both can have value, but the second pattern represents much broader independent external validation.

For this reason, organisations should normally measure both total backlinks and unique referring domains.

What makes a backlink high quality?

No single metric determines backlink quality.

A useful assessment can consider:

• topical relevance;
• editorial context;
• publisher credibility;
• page quality;
• placement;
• audience relevance;
• referring-domain diversity;
• and whether the link exists for a legitimate editorial reason.

Third-party authority metrics can help with comparison, but they should not replace contextual evaluation.

Does Domain Rating determine whether a backlink is valuable?

No.

Domain Rating is a proprietary Ahrefs metric describing characteristics of a site’s backlink profile.

It is not a Google ranking score.

A relevant link from a specialist publication can sometimes be strategically more useful than a less relevant link from a domain with a higher third-party authority score.

Domain-level metrics are best used as one diagnostic input rather than as the sole definition of link quality.

Is Digital PR the same as link building?

No, although the two disciplines increasingly overlap.

Traditional link building focuses primarily on acquiring backlinks.

Digital PR can generate a much wider range of outcomes, including:

• editorial coverage;
• backlinks;
• Brand Mentions;
• expert quotations;
• research citations;
• referral traffic;
• branded Search demand;
• and potentially AI citation visibility.

A backlink is therefore one possible output of Digital PR rather than its complete purpose.

What type of content earns the most backlinks?

The evidence suggests that content is more likely to earn links when it gives another publisher something genuinely useful to reference.

Examples can include:

• original research;
• statistics;
• proprietary datasets;
• benchmark studies;
• industry reports;
• visualisations;
• calculators;
• interactive tools;
• and authoritative reference resources.

The key distinction is not simply content length or format.

It is whether the resource contains information another writer has a legitimate reason to cite.

Does longer content automatically earn more links?

No.

Large-scale content studies have found that longer resources can earn more referring-domain links on average, but length itself does not guarantee backlinks.

A 5,000-word article containing no original value may be less linkable than a concise dataset containing genuinely new evidence.

Depth is useful when it adds:

• new information;
• analysis;
• evidence;
• context;
• or reference value.

How long does Digital PR take to produce results?

The practitioner research examined in this study suggests that initial media coverage can often occur relatively quickly, while broader measurable campaign outcomes may take several months.

That distinction is important.

A campaign may secure its first article within days but require considerably longer to build:

• multiple referring domains;
• Search visibility;
• Brand Authority;
• AI citation visibility;
• and commercial impact.

Digital PR should therefore normally be assessed over a longer period than the launch week alone.

How much does link building or Digital PR cost?

There is no universal price.

Costs depend on the service model and can include:

• research;
• data collection;
• consumer surveys;
• content creation;
• design;
• media databases;
• journalist outreach;
• specialist staff;
• and measurement.

The pricing research reviewed in this study showed significant variation in monthly retainers and reported cost per link.

For that reason, comparing providers only on cost per backlink can be misleading.

Should businesses buy backlinks?

Paid links intended to manipulate Search rankings create risks and should not be confused with editorially earned Digital PR coverage.

The authority model developed in this research is based on creating something worth referencing and earning independent recognition.

The stronger sequence is:

Create Value
↓
Earn Coverage
↓
Receive a Link or Citation

rather than manufacturing the appearance of editorial authority.

Do backlinks disappear over time?

Yes.

The link-decay evidence reviewed in this study shows that a substantial proportion of webpages and links disappear over longer periods.

Backlinks can be lost because:

• the linking page is deleted;
• the entire website disappears;
• the publisher removes the link;
• the URL structure changes;
• the destination page disappears;
• or technical changes alter how the page is indexed.

This means authority building requires both acquisition and preservation.

What is link reclamation?

Link reclamation is the process of identifying valuable links that have been lost or broken and determining whether they can reasonably be restored.

Examples include:

• redirecting an old URL to the correct replacement resource;
• restoring an accidentally removed page;
• asking a publisher to update a broken reference;
• or reclaiming an unlinked Brand Mention where a source citation would be genuinely useful.

Not every lost link is recoverable, so the first step is identifying why it disappeared.

Do unlinked Brand Mentions matter?

They can matter as part of the wider external evidence surrounding an organisation.

A Brand Mention can reinforce:

• recognition;
• topic association;
• reputation;
• entity understanding;
• and awareness.

Emerging AI visibility research has also found meaningful relationships between broader web mentions and Brand Visibility in generative Search.

This does not mean an unlinked mention should be treated as identical to a backlink.

They are different forms of external evidence.

Are backlinks important for ChatGPT and AI Search?

The emerging evidence suggests a more complex relationship than conventional SEO.

Backlinks and referring domains remain part of the wider authority environment, but AI systems can also evaluate information using:

• Brand Mentions;
• source relevance;
• content quality;
• entity relationships;
• Search retrieval;
• third-party corroboration;
• and other source-selection signals.

A page therefore does not necessarily need to rank first for the exact query to appear as an AI citation.

Is an AI citation the same as a backlink?

No.

A backlink is a persistent hyperlink published on an external webpage.

An AI citation is typically generated dynamically in response to a particular query.

AI citations can change between:

• different prompts;
• different users;
• different models;
• different Search interfaces;
• and repeated generations of the same query.

AI citations should therefore be measured separately from conventional backlinks.

Can an AI system use a source without citing it?

Yes.

The retrieval studies examined in this research indicate that generative systems can retrieve or consider considerably more material than ultimately receives visible citation.

This creates several separate stages:

Discoverable
↓
Retrieved
↓
Evaluated
↓
Cited
↓
Mentioned
↓
Recommended

An organisation can therefore contribute information to an answer without receiving visible attribution.

Can a website be cited by AI without the brand being mentioned?

Yes.

The “ghost citation” research examined earlier found many cases where a domain appeared as an AI source but the associated brand was not named in the generated answer.

This is why organisations should measure:

• AI citations;
• Brand Mentions;
• and recommendations

as separate visibility outcomes.

What is the difference between Link Authority and Brand Authority?

Link Authority primarily concerns the external hyperlink network surrounding a website or webpage.

Brand Authority is broader.

It can include:

• editorial mentions;
• recognition;
• reputation;
• reviews;
• expert references;
• branded Search demand;
• citations;
• and consistent third-party corroboration.

Strong organisations can therefore possess significant Brand Authority that is not fully represented by backlink totals alone.

What is Entity Authority?

Entity Authority describes the strength and consistency of the information environment connecting an organisation with particular topics, people, products, locations or areas of expertise.

For example, repeated independent references connecting one organisation with AI Search research can reinforce the relationship:

Organisation ↔ AI Search Research

Entity Authority depends on more than links alone.

It also involves identity clarity, consistency and external corroboration.

How should businesses measure link-building performance?

A modern authority dashboard should normally monitor more than raw backlink totals.

Useful KPIs can include:

• new referring domains;
• lost referring domains;
• net authority growth;
• link quality distribution;
• editorial placements;
• Brand Mentions;
• link-retention rate;
• AI citation visibility;
• Search visibility;
• referral traffic;
• branded Search demand;
• leads;
• and commercial outcomes.

What is the most important lesson from these 100 statistics?

The most important lesson is that link building is becoming part of a much broader authority system.

The strongest strategy is not simply:

Build More Links

It is:

Create Original Value
↓
Earn Independent References
↓
Build Links, Mentions & Citations
↓
Strengthen Brand & Entity Authority
↓
Increase Visibility Across Search & AI

The backlink remains important.

But it is increasingly one component within a wider system of recognition, evidence and external authority.

Conclusion — Link Building in 2026: From Backlinks to Search, Brand & AI Authority

The evidence in this research shows that link building remains relevant in 2026, but its role is changing.

The older model focused heavily on acquiring as many backlinks as possible.

The stronger model is broader:

Create Value
↓
Earn Independent Recognition
↓
Generate Links, Mentions & Citations
↓
Build Brand, Entity & Source Authority
↓
Increase Search & AI Visibility

Across the 100 statistics, several patterns repeatedly emerge.

Backlinks Still Matter

Large-scale ranking studies continue to show measurable relationships between backlinks, referring domains and higher Search positions.

This means backlinks should not be dismissed simply because Search interfaces are changing.

However, backlink counts alone provide an incomplete picture.

The strongest authority profiles tend to include:

• diverse referring domains;
• editorially relevant links;
• credible publishers;
• and sustained external recognition.

Referring Domains Matter More Than Repetition

The research repeatedly highlights the importance of independent sources.

One website linking many times can be useful.

But multiple independent websites choosing to reference the same organisation create a different type of authority signal.

This is why referring-domain diversity should remain a core measurement within any serious link-building programme.

Most Content Is Not Naturally Linkable

One of the clearest findings in the evidence is that most published content earns few or no backlinks.

The practical implication is significant.

Publishing more articles does not automatically create more authority.

Linkable content usually contains something another publisher needs.

That can include:

• new data;
• research;
• statistics;
• benchmarks;
• tools;
• visualisations;
• or authoritative reference material.

The strongest question is therefore not:

“What Should We Publish This Week?”

It is:

“What Can We Publish That Other People Will Need to Reference?”

Digital PR Is Becoming a Core Authority Channel

Digital PR sits increasingly close to SEO because editorial coverage can produce multiple forms of value at the same time.

A successful campaign can generate:

• backlinks;
• Brand Mentions;
• expert recognition;
• referral traffic;
• research citations;
• and wider visibility.

The strongest Digital PR therefore starts with a story that is relevant enough to be published, not with the backlink as the only objective.

Authority Has to Be Maintained

The web-decay evidence shows that backlinks disappear.

Pages are deleted.

Domains close.

Links are removed.

URLs change.

This means authority is not only something an organisation builds.

It is something it must preserve.

Long-term authority management should therefore include:

• lost-link monitoring;
• link reclamation;
• redirect management;
• preservation of high-value URLs;
• and continuous acquisition of new independent references.

Brand Mentions Are Becoming More Important

The emerging AI evidence suggests that Brand Mentions across the wider web may be particularly important within the broader authority environment.

This does not mean Brand Mentions replace backlinks.

It means organisations should measure both.

A backlink can create:

• Search Authority;
• referral value;
• and an explicit relationship between two webpages.

A Brand Mention can create:

• recognition;
• entity reinforcement;
• topic association;
• and broader external corroboration.

The strongest authority profile increasingly combines the two.

AI Search Adds a New Citation Layer

AI Search introduces another stage in the authority journey.

A source can now be:

• retrieved;
• evaluated;
• cited;
• mentioned;
• or recommended.

These events are different.

The evidence also shows that AI citation behaviour is more volatile than conventional backlinks.

Different AI systems can reach similar answers using very different sources.

This means AI authority cannot be measured using one static ranking position.

The Future Is Multi-Signal Authority

The direction of travel across Search, Digital PR and AI discovery points toward a broader authority model.

That model includes:

Backlinks
+
Referring Domains
+
Editorial Coverage
+
Brand Mentions
+
Entity Clarity
+
Source Citations
+
Recommendation Visibility

No single element explains the complete system.

The stronger organisations are likely to be those that build credible external evidence across all of them.

References

The following sources support the statistics, benchmarks and analysis presented throughout this research.

Where several statistics originate from the same large-scale study, the source is listed once below rather than repeated for every individual finding.

Readers should refer to the original research for complete methodology, sample construction, publication dates, qualifications and subsequent updates.

1. Backlinko. “We Analyzed 11.8 Million Google Search Results.” Large-scale analysis of Google rankings, backlinks, referring domains and other ranking characteristics. https://backlinko.com/search-engine-ranking
2. Semrush. “Google Ranking Factors.” Research and analysis of factors associated with Google Search performance, including backlink and referring-domain variables. https://www.semrush.com/blog/google-ranking-factors/
3. Ahrefs. “Links Matter Less, But They Still Matter.” Analysis of approximately one million US keywords examining relationships between backlinks, referring domains and Google rankings. https://ahrefs.com/blog/links-matter-less-but-still-matter/
4. Ahrefs. “How Fast Do Top-Ranking Pages Get New Links?” Backlink growth study examining the acquisition of new followed referring domains by high-ranking pages. https://ahrefs.com/blog/backlink-growth-study/
5. Ahrefs. “How Many Web Pages Get Organic Search Traffic?” Large-scale Search traffic analysis examining billions of pages and the relationship between external links and Google traffic. https://ahrefs.com/blog/search-traffic-study/
6. Semrush. “Referring Domains: What They Are and Why They Matter.” Analysis and guidance concerning referring-domain diversity and Search visibility. https://www.semrush.com/blog/referring-domain/
7. Semrush. “How to Find Backlinks.” Search ranking and backlink research including data on backlinks among top-ranking pages. https://www.semrush.com/blog/find-backlinks/
8. Ahrefs. “Anchor Text: A Data-Driven Guide.” Analysis of 19,840 keywords and 384,614 webpages examining relationships between different anchor-text types and ranking positions. https://ahrefs.com/blog/anchor-text/
9. Ahrefs. “Do Links From Pages With Traffic Help You Rank Higher?” Study of 44,589 SERPs examining backlinks from pages receiving organic Search traffic. https://ahrefs.com/blog/links-with-traffic-study/
10. Backlinko and Pitchbox. “We Analyzed 12 Million Outreach Emails.” Large-scale outreach study examining response rates, personalisation, subject lines, follow-ups and contact strategies. https://backlinko.com/email-outreach-study
11. BuzzStream. “How Long Does It Take to Get a Reply to a Link Building Email?” Analysis of more than 51 million outreach emails and response timing. https://www.buzzstream.com/blog/how-long-to-get-email-replies/
12. BuzzStream. “State of Digital PR Report 2026.” Survey of more than 150 Digital PR professionals examining effectiveness, campaign timescales, backlinks, budgets, cost per link, AI citations and AI visibility measurement. https://www.buzzstream.com/blog/state-of-digital-pr-2026/
13. Muck Rack. “State of Journalism 2025.” Research involving more than 1,500 journalists examining relationships with PR professionals, pitching and editorial workflows. https://muckrack.com/resources/webinars/takeaways-state-of-journalism-2025
14. Muck Rack. “State of Journalism 2025 Webinar FAQs.” Journalist pitching preferences including pitch length, volume and relevance. https://muckrack.com/blog/2025/07/09/state-of-journalism-2025-webinar-faqs
15. Backlinko and BuzzSumo. “We Analyzed 912 Million Blog Posts.” Large-scale content study examining external links, referring domains, content length, content formats and social sharing. https://backlinko.com/content-study
16. BuzzSumo. “Content Trends 2018.” Analysis of approximately 100 million articles examining sharing, backlinks and changes in content engagement. https://buzzsumo.com/wp-content/uploads/2018/02/BuzzSumo-ContentTrends-2018.pdf
17. BuzzSumo and Majestic. “The Magical Content That Gets Links and Shares.” Research examining content formats that attract both backlinks and social engagement, including original research and reference assets. https://buzzsumo.com/blog/magical-content-gets-links-shares-new-research-buzzsumo-majestic/
18. BuzzStream. “How Much Does Digital PR Cost?” Pricing research involving agencies, freelancers and consultants, including UK and US comparisons, retainers, contract lengths and cost-per-link data. https://www.buzzstream.com/blog/digital-pr-costs/
19. Pew Research Center. “When Online Content Disappears.” Analysis of approximately one million historical webpages alongside large samples of government and news pages examining digital decay and broken links. https://www.pewresearch.org/data-labs/2024/05/17/when-online-content-disappears/
20. Ahrefs. “Link Rot Study.” Nine-year analysis of links pointing to more than two million websites examining link loss, page disappearance, removed links and other causes of backlink decay. https://ahrefs.com/blog/link-rot-study/
21. BrightLocal. “Local Consumer Review Survey 2026.” Consumer research examining review usage, review recency, ratings, platform diversity, business research and use of generative AI for local recommendations. https://www.brightlocal.com/research/local-consumer-review-survey/
22. Ahrefs. “38% of AI Overview Citations Pull From the Top 10.” Analysis of 863,000 SERPs and approximately four million AI Overview URLs examining citation overlap with conventional Search rankings. https://ahrefs.com/blog/ai-overview-citations-top-10/
23. Ahrefs. “Why ChatGPT Cites One Page Over Another.” Analysis of approximately 1.4 million ChatGPT prompts examining retrieved URLs, citation selection and specialist retrieval channels. https://ahrefs.com/blog/why-chatgpt-cites-pages/
24. Semrush and Growth Memo. “The Ghost Citations Study.” Multi-country analysis of AI source appearances examining citation-only appearances, Brand Mentions and the distinction between source attribution and brand visibility. https://www.semrush.com/blog/the-ghost-citations-study/
25. Ahrefs. “AI Overviews Change Every Two Days.” Longitudinal analysis of more than 43,000 keywords examining changes in AI Overview content and citation URLs over repeated observations. https://ahrefs.com/blog/ai-overview-change/
26. Ahrefs. “AI Overviews vs AI Mode.” Analysis of hundreds of thousands of query pairs comparing source overlap and semantic similarity between Google’s AI Search interfaces. https://ahrefs.com/blog/ai-overviews-vs-ai-mode/
27. Ahrefs. “AI Overview Brand Visibility Factors: 75,000 Brands Studied.” Analysis examining relationships between web mentions, branded anchors, backlinks, referring domains and Brand Visibility in Google AI Overviews. https://ahrefs.com/blog/ai-overview-brand-correlation/

CGO Media Research Ecosystem

This Link Building Statistics UK 2026 study forms part of a wider CGO Media research programme examining how organisations become discovered, understood, referenced, cited and recommended across traditional Search and AI-powered discovery environments.

The research ecosystem separates different forms of evidence and analysis rather than treating every publication as the same type of research asset.

Research papers provide depth. Statistics provide quantitative evidence. Research Observations examine emerging patterns. Frameworks convert accumulated evidence into structured models. Research Architecture connects those assets into a wider knowledge system.

Explore the Research System

Research Identity & Authorship

CGO Media research is developed as part of an independent programme examining Search, artificial intelligence, Digital Authority and information discovery.

The programme is developed under the research direction of Roger Wilkinson, independent Search and AI researcher, SEO practitioner and founder of CGO Media.

Research areas include:

• AI Search;
• Generative Engine Optimisation;
• Search Engine Optimisation;
• AI citation systems;
• source selection;
• recommendation authority;
• Entity Authority;
• Brand Authority;
• Content Authority;
• Digital PR;
• Knowledge Architecture;
• and the future of digital discovery.

External Research Distribution

Where appropriate, CGO Media research is also distributed beyond the main website through external research repositories, researcher-profile systems and academic discovery environments.

These can include:

• Zenodo — repository publication and DOI-supported research records;
• Figshare — research outputs, datasets and publication records;
• ORCID — persistent researcher identity and publication connections;
• Academia.edu — research discovery and researcher visibility;
• Google Scholar — academic and research discovery where eligible material is indexed.

Repository availability can vary by publication. Where a paper has a DOI or external repository record, readers should use the publication-specific citation information supplied with that research.

Why External Distribution Matters

Publishing research only on an organisation’s own website creates a limited evidence environment.

External research distribution can help create:

• independent publication records;
• persistent identifiers;
• research provenance;
• additional discovery pathways;
• citation opportunities;
• and clearer connections between researchers and their published work.

This is particularly relevant to research-led authority building because the objective is not simply to create more pages.

It is to create research that can be found, traced, referenced and cited.

Research Team, Authorship & Research Governance

CGO Media’s research programme is designed to separate research, interpretation and commercial application as clearly as possible.

The purpose is not simply to publish large quantities of content.

It is to build a traceable body of research that can be examined, challenged, cited, updated and reused by:

• businesses;
• journalists;
• researchers;
• SEO and Digital PR professionals;
• marketing teams;
• and organisations studying the development of AI Search and digital discovery.

Research Direction

The CGO Media research programme is developed under the research direction of Roger Wilkinson, founder of CGO Media and an independent researcher and practitioner working across SEO, Search strategy, Digital Authority, GEO and AI Search.

The programme focuses particularly on the transition from traditional Search Engine Optimisation toward a wider discovery environment involving:

• Google Search;
• AI Overviews;
• AI Search interfaces;
• generative answer engines;
• source selection;
• citation systems;
• Brand Authority;
• Entity Authority;
• Digital PR;
• and recommendation systems.

CGO Media Research Team

Research development can involve a combination of:

• research question development;
• source identification;
• data review;
• statistical extraction;
• comparative analysis;
• framework development;
• technical Search analysis;
• editorial review;
• visualisation;
• and publication management.

The exact contributors may vary by research project depending on the subject, methodology and technical requirements.

Further information about the wider research programme is available on the CGO Media Research Team page.

Explore the CGO Media Research Team →

Authorship

Research pages should identify authorship clearly enough for readers to understand who is responsible for the work.

Depending on the publication, authorship may be attributed to:

• Roger Wilkinson;
• the CGO Media Research Team;
• or named contributors where additional specialist involvement is appropriate.

Authorship indicates responsibility for the development and presentation of the research.

It does not imply that every statistical finding originated from CGO Media.

Where external research is used, the original organisation or researcher should remain clearly attributed.

Primary Research vs Secondary Research

CGO Media publications can contain different types of evidence.

These should be distinguished clearly.

Source Attribution

Where CGO Media uses statistics or findings produced by another organisation, the original source should be identified wherever practical.

The objective is to allow readers to:

• inspect the underlying evidence;
• review the original methodology;
• check publication dates;
• understand sample limitations;
• and distinguish the original finding from CGO Media’s interpretation.

Source attribution is particularly important where one statistic is widely repeated across the internet without clear identification of the original study.

Interpretation and Analysis

CGO Media may analyse or interpret third-party evidence, but that interpretation should remain separate from the underlying finding.

The preferred structure is:

Original Finding
↓
Source Attribution
↓
Context
↓
CGO Media Analysis
↓
Practical Implication

This allows readers to distinguish between what the original research established and what CGO Media concludes from that evidence.

Research Independence

CGO Media’s commercial services and its research programme can overlap in subject matter, but commercial relevance should not be used as a reason to alter or selectively present research findings.

Where evidence contradicts a common industry assumption, the evidence should be reported rather than adjusted to support a preferred commercial narrative.

Where evidence is weak, incomplete or contradictory, that uncertainty should be stated.

Evidence Before Commentary

The research programme follows a simple principle:

Commentary can identify useful questions.

It should not replace evidence.

Research should therefore prioritise:

• original datasets;
• published studies;
• large-scale platform analysis;
• documented experiments;
• reproducible observations;
• and clearly sourced quantitative evidence.

Research Review

Before publication, substantial CGO Media research should be reviewed for:

• source accuracy;
• numerical consistency;
• citation quality;
• unsupported claims;
• duplicate statistics;
• correlation-versus-causation errors;
• geographical limitations;
• methodological limitations;
• and internal consistency.

This review process is particularly important for large statistics resources where a single incorrect figure can be replicated by other publishers once the page begins receiving citations.

Corrections

Research should be corrected when a material error is identified.

Examples include:

• an incorrect statistic;
• an incorrect source;
• a misinterpreted sample;
• an outdated URL;
• a calculation error;
• or wording that materially overstates what a study demonstrates.

Minor editorial changes that do not alter the meaning of the research may be made without changing its fundamental conclusions.

Research Updates

Search and AI research can become outdated quickly.

Research assets should therefore be reviewed when:

• important new datasets become available;
• AI products materially change;
• Search platforms alter their interfaces;
• existing sources publish updated studies;
• or new evidence contradicts an earlier conclusion.

Updating a research page is preferable to leaving an outdated statistic in place simply because it appeared in the original version.

Versioning

Where a substantial revision materially changes the evidence, methodology or conclusions, the publication can be treated as a new version.

Versioning is particularly useful for research distributed through repositories because it creates a clearer record of how the publication has evolved.

A version update may be appropriate where:

• the dataset is materially expanded;
• new research sections are introduced;
• the methodology changes;
• major statistics are replaced;
• or the analytical framework changes substantially.

Persistent Research Identity

Where appropriate, research publications can be connected with persistent external research identifiers and repositories.

These may include:

• ORCID researcher identifiers;
• DOIs;
• repository records;
• and external publication metadata.

Persistent identifiers can make it easier for other researchers and journalists to identify:

• the correct author;
• the correct publication;
• the correct version;
• and the correct citation record.

Use of Artificial Intelligence in Research Production

AI tools can assist parts of the research and publication workflow, including:

• information organisation;
• draft structuring;
• data classification;
• quality-control checks;
• formatting;
• and analysis support.

AI-generated output should not be treated as an independent primary source.

Where a factual claim depends on external evidence, that claim should ultimately be traceable to an identifiable source, dataset or documented observation.

Human research oversight remains necessary for:

• source selection;
• interpretation;
• methodological decisions;
• quality control;
• and publication responsibility.

Framework Development

CGO Media frameworks are developed to organise evidence into practical models.

They should not be presented as though they are independently verified laws of Search or AI systems.

A framework can be informed by:

• published evidence;
• observational research;
• professional experience;
• comparative analysis;
• and emerging platform behaviour.

The role of the framework is to make that evidence easier to apply.

Research Governance Structure

Usage, Citation & Republishing

Journalists, researchers, publishers and organisations may quote individual statistics and findings from this research with clear attribution to CGO Media and a link to the original research page.

Where a statistic originates from external research, the original source identified within this study should also be acknowledged where appropriate.

Charts, figures and CGO Media analytical frameworks may be referenced or reproduced for editorial, research or educational purposes provided that attribution is retained and the material is not presented as independent original research by another organisation.

Suggested citation:

CGO Media (2026). Link Building Statistics UK 2026: 100 Backlink, Digital PR & Authority Statistics. CGO Media Research.

Digital Object Identifier: 10.5281/zenodo.22960611

For research, media or citation enquiries, visit Press & Media Resources or contact CGO Media.