Backlink profile

What a profile is made of

Individual links get all the attention. Profiles decide outcomes, because search engines evaluate the aggregate — the shape of who points at you, not the merits of link number 47.

Six things make up that shape. Everything else in a tool's dashboard is a derivative of these.

• Referring domains — how many separate sites link to you. The headline number, not total backlinks. One site linking you 3,000 times through a footer is one domain and roughly one vote.
• Topical relevance — how many of those domains cover your subject. Four hundred unrelated domains are weaker than forty relevant ones.
• Anchor text distribution — the words people used when linking. The health check nobody runs.
• Link type — followed, nofollow, rel="sponsored", rel="ugc". A profile that's 95% followed on paid-looking placements is a flag by itself.
• Velocity — the growth curve of referring domains. Shape matters more than slope.
• Link quality — is the linking page indexed, does it get traffic, does it link out to forty unrelated businesses.

The six columns to check in any export

Export referring domains from Ahrefs, Semrush or the Links report in Google Search Console — the free one is less complete, but it's Google's own view. Then work through this in order. Forty minutes on a mid-size site, and it beats any automated audit score you'll be sold.

1. Sort by first-seen date. Look for clusters. Twenty new domains in a fortnight, on a site that normally earns two a month, is a campaign — and if nobody on your team ran one, somebody bought it.
2. Sort by anchor text and group it. Add up brand, URL, generic ("here", "this guide") and exact-match commercial anchors. That last bucket is the one to watch.
3. Sort by referring domain traffic. A decent authority score with near-zero organic traffic means a link shop. If a third of your profile sits there, you know what happened.
4. Scan the topics. Open ten linking sites at random. If you sell dental software and casino, essay-writing and crypto blogs link to you, that profile was purchased.
5. Check the link type. Paid placements should carry rel="sponsored". If every commercial-looking guest post is followed, somebody ignored Google's policy on your behalf.
6. Open the linking pages. Indexed? Any other links? Is yours in the body, or in a bio box under a name with no surname?

Reading the anchor text distribution

This is the check that separates people who read profiles from people who read dashboards. When strangers link to you naturally, they use your brand name, your URL, the page title, or a shrug of a phrase like "this breakdown". Almost nobody spontaneously types "best seo agency in bangalore" as anchor text — so when 30% of a profile does, it wasn't spontaneous.

There's no official safe ratio, and anyone quoting one made it up. The useful test is comparative: pull the same distribution for two competitors ranking above you. If theirs is 60% brand and URL while yours is 40% exact-match commercial, you've found the anomaly without needing a threshold. The mechanics are in anchor text — anchors are a strong relevance hint, which makes them the first thing manipulators reach for and the first thing spam systems watch.

Comparing yours against two competitors

A profile in isolation means nothing. Two hundred referring domains is dominant in one market and negligible in another. Only the comparative reading produces a decision.

Pick two sites in the top five for the query that actually sells for you — not your biggest-name competitor, the one currently taking the traffic. Export their referring domains. Subtract yours. What's left is a target list of sites that demonstrably link to businesses like yours.