Backlink checker numbers look wrong: causes and quick tests
You expected links after launch week and the report says none. Before you blame the tool, or the launch, run through the likely causes below.
LinksApp research team ·
Symptom, likely cause, test
| Symptom | Likely cause | Quick test |
|---|---|---|
| Zero links a week after launch | The index refreshes roughly monthly | Wait for the next refresh and look again |
| Fewer than another tool shows | Different crawlers see different pages | Compare referring domains, not raw backlinks |
| DR differs from Ahrefs or Moz | DR here is our own metric on a log scale | Compare trends, never the raw figure |
| Link you can see in a browser is missing | Page is new, blocked or behind a script | Check again after a month |
| HTTP 401 | Missing or bad key | Check the header spelling and the sk_ prefix |
| HTTP 402 | Credit has run out | Check _credits on the last good call |
| Huge backlink count, few domains | Site-wide link in a footer or sidebar | Call refdomains and compare |
| Lookup for the wrong host | www, non-www or a subdomain was used | Try each form of the domain |
The two causes behind most surprises
Timing. Our link graph is rebuilt about once a month, so any link younger than that is a coin toss. If you shipped on Tuesday and checked on Wednesday, you checked the world as it stood weeks earlier.
Scope. Every link database sees a different slice of the web, and none sees all of it. The honest comparison is trend against trend: is your referring-domain count rising across several months?
Keep a log with the date, the DR and the count of referring domains. Three rows tell you more than any single lookup.
When to stop debugging
Stop when two independent checks agree. If a second tool also shows few links, your domain really is quiet, and the useful work is writing something worth linking.
If they disagree on a single page you know exists, note the URL and the date, then look again after the next monthly refresh. Not every gap is a fault.
A bug report with a host, a date and the exact action is easy to act on. 'It shows nothing' is not.
Before you file a bug report with anyone
- Confirm the exact host you looked up, including www.
- Note the date of the link you expected, and whether a month has passed.
- Check the link is on a public page without a login.
- Check that the response had ok set to true.
- Compare referring domains, not backlinks.
A debugging order that saves time
Start with the response. Every call returns ok, and authenticated ones add _credits. If ok is false, read the HTTP status before anything else, since 401 and 402 have different fixes.
Next, widen the lens. Look up the www form, the bare domain and your docs subdomain. Apps often scatter links across three hosts, and a lookup for one host never sees the others.
Then compare with a second tool, such as Ahrefs or Semrush, but only on referring domains and only on trend. Two crawlers rarely agree on raw counts, and nobody is wrong when they differ.
Last, wait. A monthly refresh means some answers only arrive with the next cycle. Google Search Console's links report is a useful cross-check for your own property, since it comes from Google's own view of your site.
If the table does not explain it, rerun the steps on the landing page walkthrough with a rival's domain to see whether the tool behaves the same way. For changelog-specific oddities read the changelog mistakes. Developers can compare with the docs site setup.