Keyword Set Governance (Onboarding)

Keyword Set Governance (Onboarding)

1. 01

   Pull the client's last 90 days of search console queries before proposing any tracked keyword list

   Sort by impressions and clicks, then cut anything with zero impressions in the window. This grounds the initial set in observed demand rather than stakeholder guesses.

2. 02

   Cap the launch list at 50 to 150 keywords and tag each one by intent and funnel stage

   Brand, non-brand, competitor, and local terms should be separable in reporting. A list that mixes all four into one average position number hides the story the client actually pays for.

3. 03

   Assign a device, location, and check-frequency setting per keyword group

   Local service clients need city-level geo grids and daily checks; national B2B clients can run weekly. Mismatched frequency inflates cost without improving the trend line.

4. 04

   Add a parallel AI prompt set of 10 to 20 buyer questions and record baseline citation presence

   Run each prompt through ChatGPT, Perplexity, and Google AI Overviews and log whether the client is mentioned, cited, or absent. This baseline is the only honest starting point for AI visibility claims.

5. 05

   Name 3 to 5 competitor domains and map them to the same keyword groups

   Share-of-voice comparisons only hold when the competitor set is fixed at onboarding. Adding rivals mid-quarter makes every historical chart misleading.

6. 06

   Write the reporting cadence into the retainer scope before the first report ships

   State the day of week, the metric set, and who reviews the draft. Agencies that skip this step end up rebuilding the report format every month at delivery cost.

7. 07

   Schedule a 30-day calibration review to prune dead keywords and promote winners

   Expect 10 to 20 percent of the launch list to be wrong. Budget the review hour now so pruning is routine rather than a scope negotiation.