Find page-two keywords worth working on

Use this guide to find non-branded queries with average positions between 11 and 20. The assistant maps the queries to existing pages and returns a shortlist for review.

When to use this use case

Use this use case when you want to improve pages that already appear in search. Average position is an aggregate measure, so the page-two label describes the selected position range rather than a fixed result for every search.

What to ask

Replace the domain, dates, and business context before using this prompt.

Find non-branded queries for example.com with at least 100 impressions and average positions between 11 and 20 during the 28 days ending September 30, 2026. Map each query to an existing page and check for competing pages. Our priority is [products, services, or audiences]. Return a shortlist based on business relevance, evidence, and estimated click opportunity. Label estimates as projections and flag missing page inventory.

How the assistant runs it

The assistant verifies the query filters before choosing pages:

  1. Call get_server_status, list_sites, and get_connection_status to confirm the property and attached engines.
  2. Call get_workspace_brand_terms with the workspace ID resolved for the site. Check that the effective terms represent the brand rather than an internal workspace label.
  3. Call find_low_hanging_keywords with the position range set to 11 through 20 and branded queries excluded.
  4. Call get_query_pages for selected queries. Use the returned full URLs for page lookups.
  5. Call detect_keyword_cannibalization for the queries you plan to act on. Review its evidence and inventory status.
  6. Choose pages to inspect before recommending content or internal-link changes.

Example tool input

This input selects queries in the page-two position range.

json
{  "tool": "find_low_hanging_keywords",  "arguments": {    "site": "example.com",    "to": "2026-09-30",    "windowDays": 28,    "positionMin": 11,    "positionMax": 20,    "minImpressions": 100,    "excludeBranded": true,    "limit": 20,    "searchVertical": "web"  }}

What the result includes

The shortlist contains the following information:

  • Query, engine, impressions, clicks, CTR, and average position.
  • The existing page that appears for the query.
  • Estimated click opportunity and the assumptions behind it.
  • Cannibalization verdict and missing inventory information.
  • Business relevance and the next page review.

How to read the result

Set the position range explicitly because the tool's default range starts at position 4. Adjust the impression threshold to suit the site's size, and disclose the threshold used.

The uplift estimate assumes promotion to the site's measured CTR for positions 1 through 3. It does not estimate the work required or guarantee the ranking gain. Check the benchmark before relying on a large estimate. A null estimate means the tool cannot price the opportunity, not that it has no value.

Multiple ranking pages do not prove that a merge is needed. Review search intent and page content first. Missing rows and incomplete inventory prevent the assistant from claiming that the shortlist covers every opportunity.