Finding Content Gaps from Search Impression Data
Contrast ranked queries against the absence of dedicated content to surface SEO candidates
Definition
A process for exporting page and query data from Google Search Console, cross-referencing them with an AI tool, and finding queries that already rank in search but have no content directly addressing them — treating those as content candidates.
Perspectives
Tim Bennetto (2026-08-10, X)
Set the Search Console Performance date range to 12 months, export Pages.csv and Queries.csv, then upload both files to an AI tool to find "queries you rank for but don't have content for." Writing content for those queries lets you rank quickly.
harsehaj (2026-08-24, X)
Use live Search Console near-misses — queries already receiving at least 150 impressions and ranking around positions 5–20 without a dedicated post — as generator candidates. In one Browserbase run this surfaced 120 candidates, including “what is a captcha solver?” at 19,700 impressions and position 10.6.
Before saving a draft, gate it on strategy and persona fit, required structure, a docs URL for every code snippet, and cannibalization against existing posts. If another post owns the query, stop generation and send that page to the edit audit; otherwise create an unlisted draft and wait for server-side approval.
How to apply
- Fits a site that already has months of Search Console history — enough pages and queries that something ranks by accident. A new site with no impressions has nothing to export; start from SaaS SEO Page Type Priority instead.
- Assumes a person reviews the AI's candidates against existing pages and search intent before writing anything; harsehaj's implementation makes this explicit with a cannibalization gate and server-side approval.
- Order the candidates by page type with SaaS SEO Page Type Priority, then confirm the lift by re-checking the same queries' impressions and rank in Google Search Console.
- Two neighbouring uses of the same export are different tactics with their own pages: pulling question-shaped queries for a video FAQ is Winning YouTube AI Citations for Buy-Intent Search, and sorting existing pages into delete or re-optimize buckets is Pruning SEO Pages by Search Performance.
Limits
- The original gives no criterion for judging that content "does not exist."
- "Rank quickly" is a single author's claim, with no timeframe, sample, or performance figures.
- It does not explain which AI tool or prompt to use, or how to split a new post from expanding an existing one.