A backlink dashboard gave me a number I could count. Opening the source pages showed that I had not yet decided what I was counting.
In a September 13 audit, Bing reported five referring domains and roughly 17-18 backlinks for a site I maintain. I wanted to know something narrower: had independent writers actually chosen to link to the site from their articles?
Those are different questions. Here is what the saved audit found, rather than a claim about what those pages look like today.
Twelve entries pointed to a feed
The largest group was twelve entries attributed to DEV. Their source was my profile's RSS feed, not twelve independently written articles recommending the site.
The same audit also checked 26 of my DEV posts that linked to the site and recorded them as noindex. That is a separate observation: a page's indexing directive and the relationship expressed by a link are different properties. Neither observation tells me exactly how a search engine valued those links.
What I could say was simpler: this group did not establish twelve independent editorial recommendations.
The other domains needed different labels
The remaining reported sources did not all mean the same thing either:
| Source category | What the audit recorded |
|---|---|
| Another site I own | Two links to a related subdomain, rather than independent recommendations of the main site |
| Unavailable site | A Cloudflare 1016 origin DNS error; a live link could not be verified |
| Aggregator | Its “Read full article” link went to DEV; no direct link to my site was found on the inspected page |
| Independent article | A relevant anchor linked directly to a comparison page; no rel attribute was recorded, and the page had index,follow
|
The unavailable source belongs in an unverified category. A failed fetch does not prove that a link never existed. Likewise, the aggregator observation describes the inspected page; it does not reconstruct what Bing previously crawled.
The second editorial source was missing from Bing
There was another independently written article in the audit, outside Bing's list. It contained two direct links to a relevant page, with rel="noopener" and index,follow recorded for the source page.
That made two verified editorial source pages, not two individual anchor elements. One came from the reported list; the other was absent from it. Saying “I filtered 17 backlinks down to two” would hide that distinction.
I also should not turn those HTML observations into a promise of ranking credit. The absence of nofollow is observable; the weight a search engine assigns is not.
A ledger that answers a narrower question
For the next audit, I would keep one row per inspected source page, with:
- the reported source URL and the final page actually inspected;
- the date and whether retrieval succeeded;
- the exact destination and anchor text, when a link is present;
- link
relvalues and page indexing directives in separate fields; - whether the source is mine, a feed, an aggregator, or an independent article;
- an explicit unknown state when verification fails.
Only then would I aggregate, naming the unit: referring domains, source pages, or individual links. A feed entry and a third-party recommendation can both appear in a report without answering the same question.
At the time, Google's Links report showed zero external and internal links. That did not erase the live anchors the audit recorded. It meant that its report and my page inspection were different observations, with different coverage.
The useful result was not a smaller scoreboard. It was knowing which records supported the specific claim I wanted to make, and which still needed another check.
Top comments (0)