Use this 90-day plan to establish measurement, correct important problems, and begin a repeatable content process. The schedule is a planning example. Adjust task sizes to the team and site.
Commit to deliverables you control: documented tests, access checks, corrected facts, reviewed pages, and a reporting process. Changes in recommendations are outcomes to measure, not results to guarantee by day 90.
Weeks 1-2: baseline and audit
Choose representative customer questions and record the initial responses. Repeat prompts under consistent conditions and score competitors in the same sample. Use the audit method to identify issues that need further checks.
Completion checks:
- A documented prompt inventory.
- Initial repeated measurements with test conditions recorded.
- A list of observed problems and the checks supporting them.
Weeks 3-4: technical and entity fixes
Resolve confirmed access errors on important pages. Establish maintained business facts and correct conflicting copies you control. Add supported markup where useful and verify that it matches the page.
Completion checks:
- Access checks completed with their limits stated.
- Maintained business facts and corrected copies you control.
- Outdated claims corrected on the priority pages.
Month 2: content and the tracking system
Revise pages that omit or obscure information customers need. Create additional pages only where the subject warrants them. Establish scheduled collection and review of the relevant answers, with complete response records.
Completion checks:
- The first reviewed pages published.
- A repeatable collection and review schedule.
- A report format reviewed by the responsible team.
Month 3: independent coverage and first experiments
Evaluate relevant independent publications and communities. Share useful, supported information through appropriate channels. Run a defined experiment on a specific content or access change, recording other changes during the period.
Completion checks:
- Relevant publications and communities assessed.
- At least one experiment documented, including inconclusive results.
- The next period’s priorities based on the review.
The two variants
Small team. Reduce the number of simultaneous tasks. Complete the minimum useful measurement and correction work before expanding coverage.
Larger organization. Assign owners and arrange engineering and subject-matter reviews early. Use common fact records and scoring definitions across teams.
The scorecard
| By day | What should be true |
|---|---|
| 30 | Initial measurements recorded; observed problems documented; access checked; business facts maintained |
| 60 | Reviewed pages published; repeatable tracking established; first report delivered |
| 90 | Relevant publication opportunities reviewed; experiment documented; next-period priorities recorded |
Report completed work and measured outcomes separately. Include answer counts and known configuration changes. An increase or decrease during the period does not by itself establish campaign impact.
Reviewing the first 90 days
At the end of 90 days, review which changes are verified and which results remain uncertain. Extend the plan according to observed problems and customer value. Future model updates are outside the publisher’s schedule.Research sources and limitations. Identifies the available report data and the analysis still needed to support earlier draft figures.
Use Running the Program for ongoing responsibilities and Running a Campaign for an engagement structure.
Use the completed example and worksheets
Start with the Example Cooling audit to see how a prompt becomes evidence, a correction, and a follow-up report. The example numbers are invented and labeled throughout.
Use the measurement workbook for the first project. Finish one complete audit and verification cycle before expanding the number of prompts or platforms. The first cycle should reveal which records or approvals your team is missing.