{"id":"data2-email-churn-definition-change-vp","audience":"analysts","category":"data","track":"data","technique":"audience","level":"intermediate","title":"Email a VP about a churn definition change","free":false,"task":"The churn metric's definition changed and last month's number looks better only because of that. Write an email to VP of Customer Success Lena Ostrow, plain text, under 110 words. Start with a Subject: line, put the restated like-for-like number in the first sentence of the body, show both old and new definitions' numbers, and end with one question for her. Don't apologize or thank her.","materialLabel":"Notes on the change (attached to your prompt automatically)","material":"Churn metric change - notes (Quillfeather Software)\n\nOld definition (used through September): a customer churns if they cancel OR downgrade to a cheaper plan. October churn under the old definition: 5.0% (50 of 1,000 customers).\n\nNew definition (effective for October reporting): a customer churns only if they cancel. Downgrades are tracked separately as \"contraction\".\n\nOctober under the new definition: 3.8% (38 of 1,000 customers cancelled). 12 customers downgraded.\n\nSeptember churn (old definition): 5.4%. September under the new definition, recomputed: 4.2% (42 cancelled of 1,000).\n\nThe October dashboard tile currently shows 3.8% next to the September tile showing 5.4%, which looks like a 1.6-point improvement.\n\nLike-for-like: Sept 4.2% to Oct 3.8% on the new definition is a 0.4-point improvement. On the old definition, Sept 5.4% to Oct 5.0% is also 0.4 points.\n\nPending decision: whether the dashboard should restate all months back to January under the new definition. Data engineering estimates one day of work.\n\nLena's last email asked: \"Why did churn drop so much?\"","rules":[{"type":"plainText"},{"type":"maxWords","max":110}],"mistakes":[{"title":"Leading with 1.6 points","why":"The AI repeats the dashboard's apparent improvement of 1.6 points.","fix":"Add: \"First sentence: like-for-like it improved 0.4 points, 4.2% to 3.8%.\"","checks":["r0"]},{"title":"Skipping the definitions","why":"The AI says 'methodology changed' without saying what changed.","fix":"Add: \"Say it now counts cancellations only; downgrades are tracked separately.\"","checks":["r1"]},{"title":"Ending with pleasantries","why":"The AI closes with thanks, an apology and an open invitation.","fix":"Add: \"End with one question about restating earlier months. No apology or thanks.\"","checks":["r2"]},{"title":"Leaving the format to chance","why":"The AI picks its own length and layout unless told the exact shape.","fix":"Add: \"Plain text, no bold or headings.\"","checks":["plainText"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for a total word limit and it often lands over it.","fix":"Cap each part instead, for example \"each line under 15 words.\"","checks":["maxWords"]}],"promptHints":[{"label":"Names the reader","patterns":["lena","vp"]},{"label":"Leads with the real number","patterns":["first (body )?sentence","0\\.4","like.for.like"]},{"label":"Gives both definitions","patterns":["old definition","5\\.4","cancellations only"]},{"label":"Asks for a Subject line","patterns":["subject"]},{"label":"Ends with one question","patterns":["one question","restate"]}],"hint":"A good analyst email leads with the answer and the like-for-like figure, then ends with one decision for the reader.","mission":"Explain the churn change to the VP in an email.","musts":["Lead with like-for-like","Show old and new","One question at the end"],"rubricLabels":["The first sentence of the body gives the like-for-like improvement as 0.4 points (September 4.2% to October 3.8% on the new definition), not 1.6 points","The email explains in plain words that the definition changed (cancellations only, downgrades now tracked separately) and gives the old-definition figures 5.4% to 5.0% too","It starts with a Subject: line, ends with one question about restating earlier months (for example back to January), and has no apology or thank-you"]}