There is no universal number of backlinks. Here is a 5-step method to estimate how many referring domains you need for a keyword, and which links count most.

In short
no fixed number of backlinks guarantees a ranking. The number you need depends on the pages already ranking for your keyword: count their referring domains (unique linking sites, not raw links), take the median of the top 10 and aim to close the gap with relevant, editorially placed links. A low-competition query can be won with a handful of good links or none; a competitive commercial query can require dozens of referring domains plus a strong page.
Want to go further? Ask
"How many backlinks do I need" is one of the most common questions in SEO, and most answers either dodge it or give an arbitrary figure. Links still matter: Google's link best practices state that it uses links as a signal of relevance and to find new pages to crawl, and Google's spam policies exist precisely because people try to manipulate them. This guide gives a reproducible method to estimate your own target, explains why referring domains matter more than totals and shows which links are worth the effort. For the basics, see our definition of a backlink.
Four variables change the answer from one keyword to the next.
Large studies confirm both the importance and the scarcity of links. An analysis of 11.8 million Google results by Backlinko, run with Ahrefs data, found that the first result had on average 3.8 times more backlinks than results in positions 2 to 10 (a correlation, not proof of cause). A 2020 Ahrefs study reported that 90.63% of pages get no organic traffic from Google, with a lack of backlinks among the main reasons it identified. Most pages on the web have few or no links, which is why a small number of good ones can make a difference.
Use this five-step method for each priority keyword. It takes about 15 minutes with any backlink tool, including our free backlink analysis tool.
Search the keyword in the country and language you target. Note the 10 organic URLs, ignoring ads and features such as videos. Check the keyword difficulty score your tool gives as a first signal, but do not stop there: the score is an estimate built from the same link data.
For each ranking page, record the number of referring domains pointing to that exact URL, then the referring domains of the whole site. Use referring domains rather than total backlinks: one site that links 500 times from its footer counts as 500 backlinks but only one referring domain.
A Wikipedia page, a major media site or a giant marketplace in the top 10 distorts the average. Remove the highest and lowest values, then take the median of what remains.
Subtract your page's current referring domains from the median. The difference is your page-level gap. Then compare site-level figures: if your whole domain has far fewer referring domains than the ranking sites, plan links to the site in general (homepage, studies, tools) as well as to the target page.
Divide the gap by a sustainable monthly pace and check the cost with our link building cost calculator. Re-measure every month: competitors also gain links.
A worked example makes the method concrete. Suppose the top 10 pages for a keyword have 3, 6, 8, 11, 14, 17, 21, 28, 40 and 900 referring domains. Removing 3 and 900 leaves eight values with a median of 15.5. If your page has 4 referring domains, your page-level gap is about 11 to 12 relevant referring domains.
| What you observe in the top 10 | What it means | Priority |
|---|---|---|
| Median under 5 referring domains | Content and intent decide; links are a bonus | Publish the best page first, then a few links |
| Median between 5 and 30 | Links make the difference between similar pages | Steady link building to the page and the site |
| Median above 30, strong domains | Competitive query held by established sites | Build site authority, target long-tail keywords first |
| Weak pages on strong domains | Domain strength carries them | Grow site-level referring domains |
| Strong pages on weak domains | Page quality and page links carry them | Better content plus targeted links can win |
Yes, as a planning metric. Ten links from ten different relevant sites signal broader recognition than ten links from the same site. Ranking studies consistently find a stronger correlation with the number of referring domains than with raw backlink counts, and tools that estimate site strength (Domain Authority at Moz, Domain Rating at Ahrefs) are built mainly from linking domains. Keep in mind that these are third-party scores: Google does not use them.
| Criterion | Stronger link | Weaker link |
|---|---|---|
| Relevance | Site and page on your topic | Unrelated niche |
| Placement | In the body of an article, editorial | Footer, sidebar, sitewide template |
| Linking site | Real audience and organic traffic | Site built only to sell links |
| Attribute | Followed editorial link | rel="nofollow", "sponsored" or "ugc" (a hint, often not counted) |
| Anchor | Natural, descriptive text | Repeated exact-match commercial anchor |
Google's documentation on qualifying outbound links explains how paid and user-generated links should carry nofollow, sponsored or ugc attributes. Links bought to pass ranking credit, excessive reciprocal linking and low-quality directory links are listed as link spam. If you inherited such links, run a backlink audit before adding new ones.
There is no published safe speed. What matters is that the profile looks like the result of real interest: links arriving from varied, relevant sites, to varied pages, with varied anchor text. A sudden spike of hundreds of links from unrelated sites is the pattern spam systems look for; a steady flow of a few relevant links every month is the pattern of a growing business. For tactics that earn links, see how to get backlinks.
Indirectly, yes. ChatGPT search, Perplexity and Google's AI features retrieve pages from the web, and well-linked pages rank and get retrieved more often. AI answers also rely on what many sources say about a brand, so brand mentions on third-party sites count even when they carry no link. A link profile built from relevant sites in your field serves both goals: rankings on Google and recognition by AI.
Measure the gap for your five most important keywords, fix weak or risky links, then build a steady flow of relevant referring domains to the pages that are closest to the top 10. Track progress with a domain authority tracker rather than counting links by hand. The right number of backlinks is the gap between your page and the median of the pages that already rank. Sorank closes that gap on autopilot: it publishes optimized content and obtains natural backlinks automatically from its network of 5,000+ users through backlinks acquisition. See plans and pricing, with a 3-day free trial.
Yes, for low-competition and long-tail queries where the ranking pages have few or no links themselves. Internal links from the rest of your site and a page that answers the query precisely are often enough. On competitive queries, a page without external links rarely reaches the top 10.
Not entirely. Google treats nofollow, sponsored and ugc as hints, so it may or may not use them for ranking. They still bring referral traffic, brand visibility and mentions that AI assistants can read, and a natural link profile always contains some of them.
Google publishes no safe number. What raises risk is the pattern: sudden spikes, links from unrelated or link-selling sites and repeated exact-match anchors. A steady pace of relevant, editorial links that matches your publishing activity and audience growth is the safest approach.
THEY TRIED SORANK BEFORE YOU
