{"id":"res-pm-prd-metrics-vs-dashboard","audience":"product-managers","category":"fact-checking","technique":"verification","title":"Check the metrics in your PRD against the analytics export","free":false,"task":"Your PRD's background section quotes several metrics, and you exported the real numbers from your analytics dashboard. You want only the numbers the export contradicts, as numbered lines in the form \"PRD says X; export says Y\", under 60 words total, in plain text. Skip matching numbers and claims the export doesn't cover, and don't rewrite the section.","materialLabel":"The analytics export and your PRD background section (attached to your prompt automatically)","material":"ANALYTICS EXPORT - Habitly app, Sept 2026 (pulled Oct 2)\nMonthly active users: 84,200\nNew signups: 11,300\nDay-7 retention: 31%\nUsers who set up at least one reminder: 46%\nAverage sessions per active user per week: 4.1\nPremium conversion (trial to paid): 6.2%\n\n-----\nPRD BACKGROUND SECTION (draft)\nHabitly has 84,200 monthly active users and added 13,100 new signups in September. Only 31% of new users are still active after a week. Fewer than half of users (46%) ever set up a reminder, which we believe is the main reason people churn. Active users open the app about 4 times a week, and 9.2% of trial users convert to paid.","rules":[{"type":"maxWords","max":60},{"type":"plainText"}],"mistakes":[{"title":"Missing swapped digits","why":"13,100 vs 11,300 and 9.2 vs 6.2 look alike at a glance, and the AI skims.","fix":"Say: \"Compare every number digit by digit against the export.\"","checks":["r0"]},{"title":"Arguing with your hypothesis","why":"The AI critiques the churn claim, which the data doesn't address.","fix":"Add: \"Skip claims the export doesn't cover.\"","checks":["r1"]},{"title":"Getting a rewrite","why":"The AI rewrites the section, so reviewers can't see what changed.","fix":"Say: \"Only list contradictions, no rewrite.\"","checks":["r2"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for a total word limit and it often lands 10-20% over.","fix":"Cap each line: \"each line under 12 words.\"","checks":["maxWords"]},{"title":"Getting markdown back","why":"AI defaults to bold labels and headings, which show up as stray symbols in a PRD comment.","fix":"Ask for \"plain text, no bold, no headings.\"","checks":["plainText"]}],"promptHints":[{"label":"Names the source of truth","patterns":["export.{0,30}(only|source|truth)","only.{0,30}export","against the export"]},{"label":"Asks for contradictions only","patterns":["contradict","only.{0,20}(wrong|errors|mismatch)","skip.{0,30}match"]},{"label":"Gives the line format","patterns":["prd says","export says","\\bX\\b.{0,20}\\bY\\b"]},{"label":"Sets a length limit","patterns":["\\d+\\s*words","one (short )?sentence","word (limit|count)"]},{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no headings"]}],"hint":"Transposed digits survive many reviews. Make the export the judge, ask only for mismatches, and keep the AI away from your hypothesis.","rubricLabels":["Flags both contradictions: new signups (PRD: 13,100; export: 11,300) and premium conversion (PRD: 9.2%; export: 6.2%)","Does not flag numbers that match (84,200 MAU, 31% day-7 retention, 46% reminders, about 4 sessions a week) or the claim about why people churn, which the export doesn't cover","Outputs only numbered lines in the form 'PRD says X; export says Y': no intro, no rewritten section, no analysis"]}