If you use backlink data to select prospects, you need to know what an observed link or gap establishes about a site. Two reports use the backlink data. Backlinks DB shows the links observed for one site. Backlinks DB: Link Gap compares several sites and shows where they differ. This page explains what those results mean and the conditions that change how you should read them.
What a link gap result says
You enter your website and the competitor websites you want to compare against. The report lists sites with an observed link to those competitors and no observed link to you.
Read the phrase "observed link" literally. It means a link was seen in the collected data. It does not mean the link exists today, and its absence does not mean no link exists.
The result is a starting list, not a verdict. Open a referring page before acting on it and check two things: whether the link is still there, and why the page links to your competitors at all.
That second check is where most of the value sits, because the report gives you the gap and not the reason.
An example of why the reason matters
Suppose a conference site lists three of the competitor websites you entered on its exhibitors page, with no observed link to yours. The report treats that as a gap, correctly.
Open the page and the requirement appears. Those three took a stand at the event, which is a budget decision with a date attached rather than a link you can request. A past-speakers page has the same shape and a different condition: someone submitted a talk and it was accepted.
None of that is visible in the report, and all of it decides whether the row is worth anything. Several competitors appearing together on one page usually means there is a condition they all met.
Choosing which competitors to enter
The comparison set determines the result more than any other input, and it is the input people give least thought to.
Entering four sites that closely resemble yours produces a focused list of places your direct market already appears. Entering four much larger sites produces a long list dominated by placements earned through size, brand recognition or press coverage, most of which are not available to you by asking.
Run one comparison against direct competitors of similar size to find placements you could obtain. Run another against the largest sites to see the extent of their coverage. Read the two lists differently.
The detail is covered in how competitor selection affects a backlink gap report.
Hosts, www variants and subdomains
Sites are counted at the host level, and a host is not the same thing as a brand. example.com, www.example.com and blog.example.com are three different hosts, and links pointing at one are not links pointing at another.
This matters when you enter competitors. Entering the bare domain of a competitor who serves from the www hostname can return a different picture from entering the www hostname, and neither is wrong. The same distinction appears in the link graph measurements, where all three levels are always reported for that reason.
Freshness
The collection is updated from monthly source releases, each incorporating several collection periods. It is an observation history rather than a live crawl, so a link can change after the latest source measurement.
The report describes links observed during collection. Use the current page to confirm whether those links are still present or absent. Check it before you write to a publisher, particularly when your message depends on a link being absent. A publisher who added a link last week and receives a request to add one will reasonably conclude that nobody looked.
The data cannot separate a link that was removed from one that was never seen. Both read as "no observed link", which is why the phrase is written that way.
Each link record carries its page section, filtered as Article, Navigation or Comments, and its type, filtered as Followed, Nofollow, Sponsored or UGC.
What this data does not establish
- It does not value a link. The report says a link was observed and gives its section and type. Whether any engine counts it, and what it is worth to you, are separate questions the data does not answer.
- It does not prove exhaustiveness. No link collection sees everything, so a site absent from a gap report may still link to a competitor.
- It does not date the link. Observation is not publication. A link appearing in a report today may have existed for years.
- It does not supply the editorial requirements that decide whether you can obtain the same placement.
Before adding a gap result to an outreach list, check the host setup and the referring page. That keeps the prospect decision grounded in the actual link context.
See the Link Gap report, finding competitor backlink gaps, or Backlinks DB.