Do-it-yourself game plans · No. 21 / 22
What files should you create today?
Last updated
A spreadsheet and an organised evidence folder are enough to start. Use these nine workbook tabs:
| Tab | Minimum fields |
|---|---|
01_Charter | Buyer, decision, hypothesis, scope, owner, budget, success criteria, stop date |
02_Prompts | Prompt ID, exact text, family, cohort, provenance, priority, approver, version |
03_Observations | Observation ID, prompt ID, surface, collection route, timestamp, settings, repetition, status, raw-answer path |
04_Scores | Observation ID, brand, mention, owned citation, recommendation, negative recommendation, accuracy flag, evidence excerpt, reviewer |
05_Citations | Observation ID, original URL, resolved URL, domain, source type, supported claim, brands discussed, verification status |
06_Accuracy | Claim, reference evidence, discrepancy, severity, recurrence, owner, correction route, recheck |
07_Backlog | Issue, evidence, hypothesis, priority, owner, acceptance test, estimated effort |
08_Changes | Ticket, affected URLs, before/after description, approval, deployment timestamp, rollback, related prompts |
09_Outcomes | Native visibility, assistant referrals, qualified actions, customer reports, costs, limitations, decision |
Preserve observation-level answers separately from brand-level scores. One answer can mention several brands and cite several sources; forcing everything into one spreadsheet cell makes later auditing difficult.
What should your first five working days contain?
- Day 1: secure the budget owner, implementation owner and one-page charter.
- Day 2: gather anonymised buying questions and approved product facts.
- Day 3: approve the initial prompts and collection specification; check native reports.
- Day 4: collect a small batch, time the work and test the scoring rubric.
- Day 5: resolve scoring disagreements and decide whether to complete the diagnostic or commit to the larger pilot.
The most useful DIY outcome is not “we now have an AI visibility score”. It is “we know which gaps matter, what we can responsibly change, and what evidence would justify spending more”.