What a natural backlink profile looks like

A natural backlink profile is mixed. Most anchors are brand names, URLs and titles, most links point to the homepage and to informational pages, followed and nofollow links sit side by side, and growth is uneven. No fixed ratio defines it, so the test is comparison with similar sites.

In short

  • Google publishes no ratio for anchors, follow status or link sources, so every percentage you see quoted is an industry rule of thumb.
  • Ahrefs' anchor text study of 384,614 pages, last updated in April 2020, found that most top-ranking pages have a low share of exact-match anchors.
  • Natural links go mostly to homepages and informational pages, while product, category and service pages attract few on their own.
  • The 2024 Google documentation leak contains fields that measure spikes of spam-phrase anchors, which suggests the share and timing of commercial anchors are evaluated, with weights unknown.
  • The working test is to benchmark your profile against the top five sites ranking for the same kind of page in the same market.

A natural backlink profile looks untidy. The anchors are mostly brand names, bare URLs and page titles. The links point mainly at the homepage and at a few useful pages. Some are followed and some are not. They come from many kinds of site, and they arrive unevenly, with quiet months and busy ones. There is no formula for it. Google publishes no safe ratio for anything in a link profile, so “natural” can only be judged against other sites of the same type in the same market. What follows describes the traits, the evidence behind each, and a working method for comparison.

Why “natural” matters at all

Google’s link spam policy targets links made mainly to manipulate rankings. Its systems look for patterns that people who link freely would not produce. A site owner who links to you because they liked something chooses their own words, their own page and their own timing. Hundreds of such choices produce variety, because no two of those people were working from the same instructions. Links that one person planned tend to look alike, since the same brief, the same list of target pages and the same preferred anchors sit behind every one of them.

So a natural profile is evidence, not a goal in itself. You cannot make a planned campaign natural by matching percentages, because the percentages are a by-product of many independent decisions and copying the by-product does not recreate the decisions. You can avoid the sameness that gives planned links away. That is a more modest aim, and a more useful one: it turns the question from “what ratio is safe” into “what would this campaign look like to someone reading the whole list of links at once”, which is a question you can answer by looking.

The traits, one by one

Trait Natural pattern Pattern that stands out
Anchor text Mostly brand, URL, title and generic anchors Many identical commercial keyword anchors
Target pages Homepage and informational pages get most links Nearly all links point at product or service pages
Attributes Followed and nofollow mixed Almost every link followed
Sources Many site types, mostly relevant, with some junk One site type, such as only guest posts on general blogs
Links per domain Mostly one or a few links per referring site Sitewide footer links from unrelated sites
Timing Uneven growth, with spikes that have a cause A burst of similar links with no visible reason

Anchor text is the trait with the most evidence behind it. People who link without being asked mostly use your name, your URL or the title of the page. Homepages attract brand anchors most of all. Deep informational pages attract more topical and title anchors, because the title describes the subject. Ahrefs’ anchor text study of 384,614 pages, last updated in April 2020, found only a weak positive link between exact-match anchor share and rankings, and most top pages had a low share. The same study cites an industry convention of keeping exact match to roughly 1% to 5%. That is a convention, not a rule from Google, and the study is old enough that it should be read as a description of what ranking pages looked like and not as a current threshold.

The 2024 Google documentation leak adds some support. It contains fields that record the fraction of spam-phrase anchors and whether they arrived in a spike. The fields exist. How much weight they carry is not known. What can be said is that the two things the fields record, the share of a certain kind of anchor and the suddenness of its arrival, are the same two things that separate the left and right columns of the table. That fits the idea that sameness and bursts are what get noticed, without proving how they are used.

Where the links point is the second trait. Most natural links go to homepages and to informational assets such as research, tools and guides. Commercial pages attract few, for the plain reason that a writer has little cause to send readers to a page whose job is to sell. A profile where most links point at category and product pages, with keyword anchors, does not match how people link. This is why the standard approach is to earn links to linkable assets and pass the value on with internal links. Which pages need links covers that decision, and it is the part of a campaign where a natural shape and a commercial aim can be made to agree.

Natural profiles contain both followed and nofollow links. Social platforms, forums, Wikipedia and many news sites add nofollow or similar attributes to outbound links as a matter of policy, so any site that gets talked about collects them. No reliable public figure gives a “correct” share, and we have not found one we would cite. The useful reading is therefore about the extreme and not the ratio: a site that is discussed in public picks up nofollow links whether it wants them or not, so a profile in which almost every link is followed raises the question of where the discussion is.

Sources and relevance come next. Relevance works at three levels: the linking page, the section it sits in, and the site as a whole. Page-level relevance counts for most. A natural profile is mostly relevant with a tail of oddities: scrapers, small directories, foreign-language aggregators. Google’s John Mueller has said the idea of “toxic” links was made up by SEO tools and that Google ignores random spam links. Their presence is normal, so finding them in an export is no reason for alarm. Backlink audits explains when action is needed, which is rarely.

The last two traits are links per domain and timing. The first link from a site is worth far more than the fiftieth. Natural profiles reflect that: most referring domains link once or a few times. Sitewide links from a partner or a parent company are normal. Sitewide links with keyword anchors from unrelated sites are not. As for timing, growth is uneven. A product launch, a piece of research or a news story produces a spike, and that is fine. Mueller has said the nature of the links matters, not the speed. What stands out is a spike of similar keyword-rich anchors from weak sources, so the thing to examine in a busy month is what the new links have in common. Link velocity goes through the evidence.

Natural differs by market

A local plumber, a software company and a betting affiliate do not share one natural shape. A local business has citations and directory listings in its profile. A software company has integration pages and review sites. In gambling, the research for our iGaming hub describes regulated brands as holding thousands of referring domains with a majority of brand and URL anchors and heavy homepage concentration, though we found no dated benchmark to confirm numbers. Each of those shapes follows from how the business is talked about, which is why a benchmark borrowed from another sector misleads. A directory-heavy profile is ordinary for the plumber and would look odd on the software company.

Agencies in that market are reported to work to about 60% to 70% brand and URL anchors, 15% to 20% generic or topical, and under 10% exact commercial. That is a low-confidence rule of thumb from practitioners, not data. It does show the direction: even in markets where almost every link is negotiated, careful teams copy the brand-heavy shape. The point to take from it is the ordering, with brand and URL anchors as the largest group and exact commercial anchors as the smallest, and not the particular numbers, which are a rule of thumb and belong to one market in any case.

A working method: benchmark, do not guess

This is a method, not a guarantee. It replaces a borrowed ratio with a range taken from the sites you are competing against, which is the only kind of benchmark the absence of a published ratio leaves open.

  1. Pick the five sites that rank best for the same kind of page in your country and language.
  2. In one of the backlink analysis tools, export for each: referring domains, anchor text, top linked pages and follow status. Use the same tool for all five, because tools disagree with each other.
  3. Record four numbers per site: share of brand and URL anchors, share of links to the homepage, share of followed links, and new referring domains per month.
  4. Note the range across the five. That range is “normal” for your market.
  5. Compare your own site and mark where you sit outside the range.
  6. Fix the outliers through what you do next, not by deleting what exists. If your anchors are too commercial, make the next links branded.

The sixth step deserves a word, because the instinct on finding an outlier is to remove things. Adding branded links dilutes a commercial anchor share over time without touching links that may be doing useful work, whereas deleting or disavowing changes the profile in ways that are hard to predict. Control anchors only where control is normal, such as an author byline. Use exact-match anchors mainly in internal links, where the risk is far lower. Both rules come to the same thing: put your chosen wording in the places where chosen wording is expected, and leave the wording of other people’s links to them.

What a natural profile is not

It is not a clean profile. Scrubbing every weak link through the disavow tool does not make a profile more natural, and Mueller said in 2026 that disavowing does have effects, so removing good links by mistake can cost rankings. That remark cuts against the habit of treating a disavow file as harmless tidying. If the tool has effects, then a file built from a tool’s “toxic” list can remove links that were helping, and the junk tail it was aimed at is, by Mueller’s account, something Google already ignores.

It is not a set of percentages copied from a blog post. And it is not proof of safety: a profile can match every benchmark and still be made of paid placements on sites Google ignores. The benchmark describes the shape of a profile and says nothing about whether the links inside it count, so a profile that passes the comparison has cleared one test and still has to pass the test of whether its sources are real sites with readers. To see the raw data behind your own profile, start with how to check backlinks. Terms such as referring domain, exact match and co-citation are defined in the glossary, and settling those definitions first makes the four numbers in the method above much easier to record consistently across five sites.

Common questions

What is a backlink profile?

It is the full set of links pointing to a site, described by where they come from, which pages they point to, what anchor text they use, which attributes they carry and how they arrived over time.

What is a good ratio of dofollow to nofollow links?

There is no official ratio. Natural profiles contain both kinds, because social sites, forums and many news sites add nofollow by default. A profile with almost no nofollow links is a sign the links were chosen.

What percentage of anchors should be exact match?

Google gives no figure. Ahrefs' anchor text study cites an industry convention of roughly 1% to 5%, and found most top-ranking pages have a low share. Treat any number as a rule of thumb and compare with competitors.

Does a natural profile contain bad links?

Yes. Every site that has been online for a while collects links from scrapers and low-quality directories. Google has said it ignores random spam links, so their presence is normal.

How do I know if my profile looks unnatural?

Look for sameness. Many links with the same commercial anchor, from the same type of site, all pointing at money pages and arriving in a burst, is the pattern that stands out.

Vendors to look at

  • BazoomEditor's pick

    Sponsored content and link marketplace, managed service

  • MotherlinkEditor's pick

    Backlink services, guest posts, niche edits, full SEO