A coverage diff is a short LLM-generated assessment of how thoroughly your current site covers a topical cluster — and what's missing.
It's the section of a Content Brief that answers "should I actually write this, or do I already have a page that does the job?"
Example
For a hypothetical project, the coverage diff for a "data glossary tools" cluster might read:
Your site has one page on data glossaries (
/blog/build-a-data-glossary) but it focuses on the why and not the comparison shopping. The cluster's top queries (best data glossary tools,data glossary software,comparison) all want a comparison-shaped page, which doesn't exist. Recommend a new buyer's-guide page rather than expanding the existing post.
What inputs it sees
The LLM gets:
- The cluster's top queries by impressions
- URLs already ranking for cluster queries + their average positions
- The cluster's ownership status (gap / partial / cannibal / owned)
- A summary of the queries the dominant URL doesn't currently rank for
From that it can usually tell whether the right move is "expand the existing page" vs "write a new page" vs "this cluster is fine, skip."
When the coverage diff says "skip"
We don't filter out clusters from brief generation — if you generate a brief for an already-owned cluster, the coverage diff will tell you it's not worth the work. Trust it. The free tier's 3-brief lifetime cap exists so you spend that budget on real gaps.