Worked example
Google incoming link checker post-mortem: the guide that vanished
This is a made-up site, invented to show the method. Every number below is an example, not a customer result and not a measurement.
The site, the dates and the counts are hypothetical. The steps, the API calls and the limits of the data are real. Use the structure on your own domain.
What happened, in the order it happened
The clean-up
An editor merges 60 thin posts into 12 longer ones. A 2021 how-to guide on invoice templates is folded into a general finance post and its URL is redirected to the blog index.
The first dip
Sessions to the finance section fall by roughly a fifth in the example. Nobody connects it to the clean-up, because the guide itself only had modest traffic.
The question
A colleague asks where the invoice guide went. The answer is nobody remembers, and the audit sheet has no reason written beside the row.
The link pass
The editor runs a lookup on the domain, pulls the top pages and sees the old guide URL listed with 34 referring domains in the example. The redirect points to a page that does not mention invoices.
What the link pass showed
The guide had never been a traffic star. It was a citation magnet, linked from a couple of accounting blogs and a tools roundup. Sessions told the editor nothing. Referring domains told the whole story.
The anchors column made it worse. Most linkers had written something like 'free invoice template for freelancers', and the merged post no longer offered a template. A reader arriving from those links met a general essay and left, which is a poor use of a hard-won citation.
Notice the caveat too. Our index refreshes roughly once a month, so the example editor could not tell from link data alone whether the linkers had already dropped the page. A check in Search Console's links report filled that gap.
Run the same pass through Ahrefs or Semrush. Compare. The picture would likely look similar, with different counts, since each tool crawls the web on its own schedule. The point is to catch the pattern, not to settle a number to the digit.
Three things the example teaches
Traffic is a poor proxy
The guide earned few clicks and many citations. Rank pages by both.
The redirect target matters
A redirect to a page on another topic wastes the citation. Choose the closest match.
A written reason helps
One dated line in the sheet would have saved a month of confusion.
The repair, as a runbook
- 1
Restore a stand-alone URL
Publish the template page again on its original address, with the quoted claim near the top.
- 2
Redirect on topic
Send the old address to the restored page, and keep the finance post as a related link, not the destination.
- 3
Rewrite internal links
Point newer posts at the restored guide with descriptive anchors.
- 4
Log the reason
Write the verdict and the date beside the row, so the next editor does not repeat the move.
The audit question that would have prevented it
- Did I look at referring domains before the redirect was chosen?
- Did I read the anchors for the three strongest linkers?
- Is there a dated reason in the sheet?
- Did I set a calendar note to re-check in a month, once the index has refreshed?
Sites differ. An example proves a method, not an outcome. Recovery of lost traffic is never guaranteed, and a restored page may take weeks to settle. The honest claim is smaller: you stop making the same mistake twice.
For the order of questions to run before any redirect, see the refresh-or-prune tree. For the common slips behind this story, see pruning mistakes and fixes.