Key Link Building Statistics at a Glance
| Metric | Value | Population / Scope | Data Period | Source |
|---|---|---|---|---|
| Median cost per paid link | $100–$500 | SEO and link building practitioners, global survey | 2023 | Ahrefs Link Building Survey |
| Pages with no inbound external links | 66.3% | Sample of ~1 billion pages from Ahrefs index | 2023 | Ahrefs Blog / Research |
| Cold outreach human reply rate | 5–10% (typical range) | Link building and PR outreach datasets, various campaigns | 2022–2024 | BuzzStream / Ahrefs research |
| Positive reply rate (cold outreach) | 1–5% (typical range) | Cold outreach campaigns across multiple studies | 2022–2024 | Compiled from disclosed campaign data |
| Link building listed among top 3 SEO activities | 51% of respondents | Global SEO professionals (Semrush survey) | 2023 | Semrush State of SEO 2023 |
| Annual link decay rate (approximate) | ~6% per year | Tracked link cohorts in Ahrefs index | 2023 | Ahrefs link decay analysis |
Link building statistics are quantitative measures of how campaigns acquire external links, what they cost, and how outreach performs. They come from practitioner surveys, platform-level telemetry, and published campaign datasets. Like all benchmarks, each figure reflects a specific population and measurement method — a campaign in a competitive finance niche will produce very different numbers from one targeting local hobby blogs.
Definitions Readers Need
Link building research uses several terms interchangeably that describe different things. Before comparing any two benchmarks, confirm they share the same counting rules.
A link from an external page to the target website. One domain can produce many backlinks.
A distinct domain linking to the target website. Ten links from one domain still equals one referring domain.
A link that meets explicit criteria defined before the campaign — such as minimum traffic threshold, topical relevance, or editorial context.
A published occurrence counted under the campaign's stated rules. One placement may or may not equal one referring domain.
Any response from a prospect, including rejections, automated out-of-office messages, or payment requests. Definition varies across studies.
A response expressing editorial interest or agreement. A smaller subset of all replies. What counts as "positive" must be defined per study.
Defined campaign cost divided by qualifying links acquired. Does not always include labor, overhead, or content production unless explicitly stated.
A link selected by a publisher on their own editorial judgment, without direct compensation for the placement itself.
A link placement involving direct payment to the publisher. Google's guidance requires these to carry the rel="sponsored" attribute.
A campaign that reports "we built 50 links" may have added 50 links from 3 domains or 50 links from 48 domains. Those are radically different outcomes for referring-domain diversification. Always track both metrics separately and specify which one a benchmark is measuring.
Link Building Cost Statistics
Cost data is among the most variable in link building research because it depends heavily on what work is included, what geography the agency or practitioner operates in, and what link quality criteria are in use. The figures below are drawn from surveys where sample and method are disclosed.
| Cost Model | Typical Range | Population | Period | Source |
|---|---|---|---|---|
| Monthly agency retainer | $1,500–$10,000+ | US and global link building agencies; varies by scope | 2023–2024 | Ahrefs / Backlinko pricing research |
| Cost per link (paid directly) | $100–$500 median | SEO practitioners, global survey | 2023 | Ahrefs Link Building Survey |
| Cost per link (high-authority, niche) | $1,000–$5,000+ | Agency reports; finance, health, legal niches | 2023–2024 | Semrush / agency pricing data |
| Digital PR campaign (per campaign) | $5,000–$30,000+ | UK and US digital PR agencies; project basis | 2023–2024 | PRCA / agency pricing surveys |
| In-house content + outreach (per link, fully loaded) | $200–$600+ | US in-house SEO teams with dedicated outreach resource | 2023 | Ahrefs / Moz community survey data |
| Respondents paying over $1,000 per link | ~17% of respondents | SEO practitioners, Ahrefs global survey | 2023 | Ahrefs Link Building Survey |
Source: Ahrefs Link Building Survey 2023; Semrush State of SEO 2023; Backlinko agency pricing analysis. All figures are approximate medians or reported ranges from the cited populations.
Link building costs are right-skewed: a small number of high-authority or competitive-niche placements cost far more than the majority. The mean is pulled upward by these outliers. When a report gives only an "average" without specifying mean or median, the figure is likely a mean and will overstate what most campaigns pay. For a clear explanation of how this distinction matters across distributions, see the site's guide to mean, median, and mode.
Cost Models Compared
Agency Retainer
Covers a defined scope and volume of prospecting, outreach, and placement verification. Costs include labor; link acquisition volume varies by niche competitiveness.
Per-Link Pricing
Payment per published placement. Survey respondents report wide variation by domain authority, niche, and whether content creation is included.
Digital PR Campaign
Includes research, content creation, journalist outreach, and follow-up. Outcomes are not guaranteed; cost per acquired referring domain varies substantially by campaign quality and story hook.
In-House Fully Loaded
Includes outreach specialist time, tool costs, and content production. Excludes overhead, management time, and training. Many in-house estimates undercount total cost.
Note: These cost ranges are from survey data and agency pricing research, not from a single standardized dataset. Use them as directional reference points rather than planning targets.
Link Building Outreach Response Rate Statistics
Outreach performance data requires careful reading of denominators before any comparison. The figures below specify whether a rate is measured per email sent, per unique prospect reached, or per sequence delivered. These are different things and produce different numbers from the same raw campaign data.
| Metric | Value | Denominator | Population | Source |
|---|---|---|---|---|
| Email delivery rate | ~85–92% | Emails sent | Link building outreach tools; varies by list hygiene | BuzzStream / Hunter.io platform data |
| Human reply rate (any response) | 5–10% typical | Successfully delivered emails or unique prospects | Cold link building outreach, multiple published datasets | BuzzStream outreach research; Ahrefs blog |
| Positive reply rate (interest expressed) | 1–5% typical | Successfully delivered emails or unique prospects | Cold outreach; varies by personalization and niche | Compiled from disclosed campaign studies |
| Reply rate improvement with personalized subject lines | +30–50% relative increase | A/B tests comparing generic vs. personalized subjects | Email outreach A/B test data; Backlinko research | Backlinko Email Outreach Study |
| Follow-up email contribution to replies | ~30% of all replies | Share of replies that came after first outreach touch | BuzzStream multi-touch outreach data | BuzzStream Link Building Report |
| Placement rate from prospects reached | 1–3% typical | Unique prospects producing a qualifying published link | Cold outreach campaigns; estimate from industry data | Derived from combined reply and conversion data |
Source: BuzzStream Link Building and Outreach Research; Backlinko Email Outreach Study; Ahrefs link building community data. Ranges represent observed variation across multiple campaign types; a single campaign will sit at a specific point within or outside these ranges.
Apple's Mail Privacy Protection (iOS 15, September 2021) pre-fetches email tracking pixels for many recipients, inflating reported open rates across all outreach tools. Open rate is no longer a reliable proxy for actual engagement in most outreach campaigns. Click rate on links within the email and direct replies are more dependable signals of real interest. Where older link building benchmarks cite open rates, treat them with extra caution if predating 2022.
From Prospects to Published Links: The Outreach Funnel
Most outreach reporting collapses several distinct conversion stages into a single number. Understanding the funnel makes it possible to identify where a campaign is losing momentum and whether a benchmark comparison is genuinely relevant.
Positive reply rate = unique positive-reply prospects ÷ successfully reached prospects × 100
Placement rate = prospects producing a qualifying link ÷ successfully reached prospects × 100
Reply rate includes rejections and automated responsesPositive reply rate counts only interest expressedPlacement rate measures actual published linksThe funnel below uses the hypothetical worked example that is spelled out in full in the worked example section. The numbers are constructed for illustration.
This funnel shows why a 10% reply rate does not mean a 10% link acquisition rate. The reply-to-placement conversion depends on follow-through from both sides: prospects who agree to link do not always publish, and publication delays mean some links counted at one month will appear at three. Measurement windows matter significantly when comparing campaigns or benchmarks.
A prospect who agrees in month one may not publish until month three. Campaigns measured at 30 days will show lower placement rates than the same campaign measured at 90 days. When comparing placement rates across two campaigns or against a benchmark, confirm that both use the same observation window. A "2% placement rate" measured at 30 days is not comparable to a "5% placement rate" measured at 90 days.
Link Building Tactic Statistics and Evidence
Tactic effectiveness data in link building has a persistent quality problem: most of it comes from adoption surveys or self-reported effectiveness ratings, neither of which measures what actually happened in a controlled way. The section below distinguishes evidence types before presenting any findings.
Adoption Survey
Measures how many respondents report using a tactic. A popular tactic is not necessarily an effective one. Adoption reflects perceived value, familiarity, and resource constraints.
Perceived Effectiveness
Respondents rate whether they believe a tactic works. This is opinion data, not measured outcome data. Confirmation bias and sunk-cost effects influence ratings.
Campaign Outcome Data
Measured results from disclosed campaigns: links acquired, referring domains gained, cost per placement. Stronger than surveys but may suffer from publication bias (failed campaigns rarely appear).
Controlled Test
Conditions varied deliberately to isolate the effect of one factor. Rare in link building research. Without this, causal claims about "what works" cannot be made from observational data alone.
| Tactic | Adoption Rate | Rated Effective | Evidence Type | Source |
|---|---|---|---|---|
| Guest posting / editorial contributions | ~64% | ~58% | Adoption + perceived effectiveness survey | Semrush State of SEO 2023 |
| Original research / data studies | ~49% | ~72% (of those who used it) | Adoption + perceived effectiveness survey | Content Marketing Institute / Ahrefs |
| Resource page outreach | ~44% | ~52% | Adoption + perceived effectiveness survey | Ahrefs Link Building Survey 2023 |
| Digital PR (media coverage campaigns) | ~38% | ~67% (of those who used it) | Adoption + perceived effectiveness survey | PRCA / UK digital PR practitioner survey |
| Broken link building | ~37% | ~44% | Adoption + perceived effectiveness survey | Ahrefs Link Building Survey 2023 |
| Unlinked brand mention outreach | ~32% | ~61% (of those who used it) | Adoption + perceived effectiveness survey | Ahrefs / Semrush practitioner data |
| Link exchange (reciprocal linking) | ~27% | ~33% | Adoption + perceived effectiveness survey | Semrush State of SEO 2023 |
Source: Semrush State of SEO Report, 2023; Ahrefs Link Building Survey, 2023; Content Marketing Institute B2B Research; PRCA Digital PR Practitioner Survey. All figures are from adoption and effectiveness surveys — not controlled experiments or measured campaign outcomes.
The fact that 64% of practitioners use guest posting tells you it is popular, not that it works. Effectiveness survey ratings are self-reported opinions from people who have already invested in the tactic — a population with strong confirmation bias. The only evidence type that actually measures outcomes is campaign data with disclosed methods. Where that data exists, it is labeled clearly in the tables on this page.
Digital PR and Original Research: Outcome Data
Unlike most link building tactics, digital PR and original research campaigns occasionally produce disclosed outcome data, because the referring-domain results are publishable as case studies. This data carries its own limitation: successful campaigns are far more likely to be written up than failed ones (publication bias), so treating any single case study as a typical result is not warranted.
| Campaign Type | Referring Domains Acquired | Campaign Duration | Notes |
|---|---|---|---|
| Original data study with PR distribution | 50–300+ (wide range) | 1–6 months from publish | Publication bias; successful examples heavily over-represented |
| Industry survey report (gated asset) | 20–150 | 3–12 months | Referral traffic value often exceeds raw link count |
| Free tool or calculator | 30–500+ over lifetime | Long-term compounding | Tool development cost not included in most reported CPL figures |
| Reactive PR / data newsjacking | 5–80 per story | Days to weeks | Speed-dependent; most stories produce fewer than 20 unique RDs |
These are ranges from published case studies and are not representative of typical campaigns. No random sample of campaigns exists. Source: Compiled from Ahrefs blog case studies, BuzzStream published research, and UK digital PR agency case study data (2022–2024).
Link Quality and Persistence Statistics
A link acquired today is not necessarily a link that exists in six months. Persistence data helps answer whether an investment in link building maintains value over time or requires ongoing spend to sustain it.
A link appearing "lost" in a third-party index like Ahrefs or Semrush may have been removed from the page, or it may simply not have been recrawled recently. Third-party indexes do not crawl all pages on the same schedule. Before concluding a link has gone, verify directly at the source URL. Decay rates derived from third-party index data will overstate actual removal rates to some degree.
| Metric | Value | Measurement Method | Source |
|---|---|---|---|
| Approximate annual link loss rate | ~6% per year | Cohort tracking within Ahrefs index; includes uncrawled and removed | Ahrefs link decay analysis |
| Pages with zero referring domains | 66.3% | Sample of ~1 billion pages from Ahrefs crawl database | Ahrefs Web Analytics Study |
| Pages getting organic search traffic despite no/few links | ~10% of link-poor pages | Ahrefs database cross-reference (index sample) | Ahrefs blog research |
| Share of link building campaigns tracking persistence | ~28% | Self-reported survey; practitioner sample | Ahrefs Link Building Survey 2023 |
| Links with referral traffic from acquired placements | Minority; varies by publisher audience size | Campaign-level tracking required; no universal benchmark | Contextual inference from traffic data |
Source: Ahrefs content research and database studies, 2022–2024. The 66.3% figure refers specifically to a sample of pages in the Ahrefs crawl database and should not be generalized to "all pages on the internet."
Links, Rankings, and Business Outcomes
The relationship between backlinks and search rankings is one of the most cited correlations in SEO. It is also one of the most commonly overstated. The evidence is clear that pages with more referring domains tend to rank for more queries. The evidence does not support the conclusion that acquiring links will cause your rankings to improve by a predictable amount.
High-ranking pages also tend to have stronger content, established domain histories, better user engagement signals, and more thorough technical optimization. A campaign adding links to a page that lacks those other factors may produce no ranking movement at all. Attributing a ranking change to a link building campaign requires ruling out other simultaneous changes (content updates, technical fixes, seasonal patterns, competitor activity), which is rarely done in published case studies.
| Finding | Evidence Type | Limitation | Source |
|---|---|---|---|
| Number of referring domains correlates with keyword rankings | Correlation study across large SERP sample | Correlation; cannot establish causation; many confounders | Ahrefs / Backlinko SERP studies |
| Top-ranking pages have more referring domains than pages on pages 2–3 | Observed association in index data | Survivorship: well-ranked pages attract links organically | Ahrefs SERP analysis data |
| Link velocity (rate of acquisition) associated with ranking trajectory | Observational time-series study | Many other variables change simultaneously; causation unproven | Semrush SEO research |
| Controlled experiment isolating link acquisition as ranking cause | No publicly available controlled experiment exists | All evidence is observational; causal attribution is not established | N/A — evidence gap |
Google's Search documentation confirms that links are one of several signals used to assess relevance and authority. The same documentation notes that links from low-quality, irrelevant, or paid-for sources can be discounted or treated as spam signals. Google's link spam guidance explicitly requires that compensated links carry the rel="sponsored" or rel="nofollow" attribute. For current details, see Google's crawlable links documentation.
This limitation is relevant to how link building is measured and justified at a business level. Attribution of ranking changes to a link campaign requires controlling for other simultaneous variables, tracking before and after over a sufficient window, and comparing against periods with no link activity. This level of rigor is rarely present in published case studies or agency reports.
For an introduction to how statistical associations and causal claims differ, see the site's guide to correlation vs. causation. For A/B testing frameworks that can isolate variables in SEO experiments, see the guide to how statistics powers A/B testing.
How to Read Link Building Statistics Correctly
This section addresses the interpretation errors that appear most often when link building benchmarks are applied in planning or reporting.
- Who is in the sample? Agency survey data from high-budget clients produces different numbers from a small-business practitioner survey. A benchmark drawn from US finance-sector campaigns does not apply to a local home improvement business.
- What is the exact denominator? "10% reply rate" means different things if it is per email sent, per email delivered, per unique prospect contacted, or per outreach sequence. Two reports can give the same number for completely different quantities.
- Is it mean or median? Cost-per-link data is right-skewed. High-value, competitive-niche placements pull the mean well above the median. The median is more representative of what most practitioners pay.
- Is it survey opinion or measured outcome? Practitioner surveys about what "works" are opinion data. Campaign datasets with verified placements are outcome data. They are not interchangeable.
- What time window was used? A placement rate measured at 30 days will be lower than the same campaign at 90 days. A link persistence rate measured at 6 months will be higher than at 3 years.
Percentage Points vs. Relative Change
A commonly cited example from personalization research: if a personalized subject line raises reply rate from 5% to 8%, two things are true at once. It increased by 3 percentage points. It also increased by 60% relative to the starting rate (because 3 ÷ 5 = 0.60). Both statements are accurate — they describe different things. When a link building vendor reports a "60% improvement in reply rates," checking whether that is a relative or absolute change matters considerably.
+3 pp absolute (percentage-point) increase+60% relative change (3 ÷ 5)Both correct — describe different thingsFor a detailed explanation of how these two measures differ in interpretation and when each applies, see the site's reference on descriptive statistics, particularly the section on relative vs. absolute change.
Sample Size and Representativeness
A large number of emails in an outreach dataset does not automatically mean a large number of independent observations. Recipients at the same company, campaigns run by the same agency, and prospects in the same niche are correlated — they are not fully independent data points. Benchmark reports that cite millions of outreach emails without discussing clustering are presenting misleadingly precise-looking figures from what may be a relatively small number of truly independent campaigns.
Worked Example: Outreach Funnel and Acquisition Cost
The following is a hypothetical educational example designed to show how reply rate, placement rate, and cost per acquired link answer different questions from the same campaign data. The figures are constructed for illustration and do not represent any real campaign.
500-prospect cold outreach campaign: calculating five distinct metrics
| Outcome | Count |
|---|---|
| Unique prospects contacted | 500 |
| Prospects successfully reached (delivered) | 450 |
| Prospects providing any human reply | 45 |
| Prospects providing a positive reply | 18 |
| Prospects producing a qualifying published placement | 12 |
| Qualifying links published | 12 |
| Newly acquired referring domains | 10 |
| Defined campaign cost (labor + tools + content) | $3,000 |
Calculated Metrics:
| Metric | Calculation | Result |
|---|---|---|
| Delivery rate | 450 ÷ 500 × 100 | 90% |
| Reply rate (among reached) | 45 ÷ 450 × 100 | 10% |
| Positive reply rate (among reached) | 18 ÷ 450 × 100 | 4% |
| Placement rate (among reached) | 12 ÷ 450 × 100 | ~2.7% |
| Positive-reply-to-placement conversion | 12 ÷ 18 × 100 | ~66.7% |
| Cost per qualifying link | $3,000 ÷ 12 | $250 |
| Cost per newly acquired referring domain | $3,000 ÷ 10 | $300 |
The $3,000 here includes all defined campaign costs (labor, tools, content) — not a direct payment to publishers. A "10% reply rate" sounds promising but translates to a 2.7% placement rate once rejections and non-follow-through are counted. Cost per qualifying link ($250) differs from cost per new referring domain ($300) because two of the 12 links came from domains already linking to the site. Neither figure tells you anything about ranking impact or business ROI on its own. That requires measuring organic visibility, referral traffic, and business outcomes separately.
How to Apply These Benchmarks to Your Campaign
A benchmark is a reference point drawn from a specific population. The most useful application is to compare it against your own campaign data measured the same way, over the same window, with the same definitions. Here is a practical sequence.
Define qualifying criteria before outreach begins
Decide what counts as a qualifying link before you contact anyone: minimum estimated traffic, topical relevance, editorial context, and whether the placement must be dofollow. Changing criteria mid-campaign invalidates comparisons.
Track unique prospects, not just emails sent
Record delivery, any reply, positive reply, and published placement per unique prospect. This gives you four distinct metrics rather than one. Message-level tracking conflates what should be measured at prospect level.
Record fully loaded costs
Include labor time, tool subscriptions, content production, and any fees. A cost per link that excludes labor will be far lower than the real cost. Decide before the campaign whether to include overhead and document it for future comparisons.
Separate total links from new referring domains
Report both metrics. Multiple links from the same domain count once toward referring-domain growth. Track incremental referring domain acquisition as the primary diversification measure.
Allow a full observation window before judging placement rate
Measure at 30, 60, and 90 days from initial outreach. Some publishers agree but delay; early measurement understates final placement rate. Consistent window lengths are essential for comparing campaigns against each other or against benchmarks.
Evaluate search and business results separately
Track organic visibility, referral traffic, and business outcomes independently of link acquisition metrics. If rankings improve, identify what else changed (content, technical, competition) before attributing the improvement to links alone.
Data Sources and Methodology
Every statistic on this page carries explicit source attribution. The table below identifies the primary research organizations cited and notes limitations relevant to applying their findings to your own campaigns.
| Organization | Coverage on This Page | Methodology | Key Limitation |
|---|---|---|---|
| Ahrefs | Cost per link, link decay, referring domain data, tactic surveys | Index-level data + practitioner surveys; global sample | Index coverage ≠ all web; practitioner surveys have self-selection and vendor bias |
| Semrush | Tactic adoption, SEO activity rankings | Annual State of SEO survey; global SEO professionals | Semrush customer base may not represent all practitioner segments |
| BuzzStream | Outreach reply rates, follow-up contribution | Platform-level outreach telemetry; largely US/UK | BuzzStream user base skews toward agencies and experienced practitioners |
| Backlinko | Personalization effect on reply rates, outreach studies | Outreach A/B studies and SERP correlation research | Study conditions may not generalize; SERP studies are observational |
| Content Marketing Institute | Perceived effectiveness of original research as a tactic | Annual B2B survey; North American marketers | Self-reported effectiveness; sample skews to content-focused teams |
| Google Search documentation | Link attributes, spam guidance, compensated link policy | Official technical documentation | Policy documentation may not reflect all ranking signal implementation details |
Link building statistics shift as platform data changes, search algorithms update, and practitioner sampling evolves. The figures on this page reflect research available as of October 2026, with data periods noted for each finding. Older research is included only where it remains directionally relevant and no recent equivalent exists. This content is for research and planning purposes only and does not constitute professional SEO advice. Statistics described as estimates or derived from ranges are approximate and should not be used as guarantee thresholds.