Reading the gap
How to check inbound links when GSC gives you a sample
The Links report is the first place to look and a poor place to stop. Here is what it covers, what a separate index adds, and how to reconcile the two.
Reading the gap
Treat the Links report as a sample, not a census
Use Search Console for your own verified sites and as a sanity check. Use an index when you need rivals, scale, repeatability or a script. Neither one is complete, and the useful number is the difference between them.
- Overlap tells you the links both sources can see
- Only-in-GSC rows are usually fresh or niche links
- Only-in-index rows are older links Google may already ignore
Six differences that change how you work
| Aspect | Search Console Links report | Our link index |
|---|---|---|
| Scope | Properties you verified | Any domain, including competitors |
| Completeness | Not exhaustive; tables are trimmed | Up to 1,000 rows per call, paged to 50,000 |
| Export | Manual CSV from the interface | JSON by REST, scriptable from Sheets |
| Refresh | Follows Google's crawl and processing | About once a month, on our schedule |
| Fields | Linking sites, linked pages, anchor text | Referring domains, backlinks, anchors, top pages, DR |
| Cost | Free | 1 credit is $0.0001; free actions need no key |
Where each source misleads you
A GSC export looks tidy and feels official. That is the trap. Because the list is trimmed, a sheet that counts its rows is counting a ceiling set by the report, and not the number of sites that link to you.
An independent index has the opposite flaw, since it sees plenty of links that Google may well discount, such as footers, directories and scraped copies, and it can run a month behind. Counting every row as value inflates the story. Be sceptical.
Match on the referring domain, not the URL. Two sources rarely list the same linking page, but they often agree on the domain, so a domain-level overlap table is the fairest comparison you can build in a single afternoon. Twenty minutes is realistic.
One more caution. Since the Penguin 4.0 release in September 2016, Google has mostly devalued spammy links rather than punishing the sites that receive them, and SpamBrain, announced in 2022, extended that approach. A long list of poor links is clutter, not a verdict.
A reconciliation recipe for one afternoon
Export the Links report's linking sites into a tab called gsc-links, and pull refdomains for the same site into a second tab, so that you hold two columns of domain names. That is all you need. Keep the monthly snapshot date in each tab's name.
Now add a helper column to each tab with =COUNTIF(other_tab!A:A, A2), where a result of 1 means both sources saw the domain and 0 means only one did. Filter, count and write the three totals at the top. Done.
The totals are the finding. If 70% of your GSC domains also appear in the index, the index is a decent proxy for your site. If only 20% do, lean on Search Console and use the index mainly for looking at rivals.
Large lists? For anything above a few hundred domains, run bulk_dr on the lot: up to 2,000 domains for 1 credit. Then sort the only-in-index rows by DR and skim the top 50 before you decide whether any of them matter. Our join walkthrough covers the Apps Script half.
Save both exports with the month in the filename. In three months you can diff them, and a diff is the only way to see whether links were lost or simply stopped being reported.