{"id":"data2-email-test-request-reply-runtime","audience":"product-managers","category":"data","track":"data","technique":"audience","level":"intermediate","title":"Reply to a test request with how long it would take","free":false,"task":"A marketer asked you to A/B test a new headline and you need to reply with what it takes. Write a plain-text email to Hannah Boyd under 100 words with a Subject: line. The first sentence states about how many days the test would take, with the division shown, then ask her exactly 3 questions on lines starting Q1:, Q2: and Q3:, one for each missing decision. Don't promise a result.","materialLabel":"Request and traffic notes (attached to your prompt automatically)","material":"Hannah's message (Slack): \"Can we A/B test the new headline on the pricing page? I think it'll lift sign-ups. Let's start Monday.\"\n\nTraffic and baseline:\n- The pricing page gets about 800 visitors a day.\n- Current sign-up rate on that page: 3.0%.\n- Hannah hopes for a lift to 3.5% (0.5 points).\n- Sample size calculator (standard settings, two versions, 50/50 split): 19,500 visitors per version, so 39,000 in total.\n- Days needed = 39,000 / 800 = 48.75, so about 49 days (7 weeks).\n\nNot yet decided:\n- Which one metric decides the test (sign-ups, paid conversions, or revenue per visitor).\n- Which visitors are included (everyone, or only new visitors).\n- The smallest lift that would be worth shipping.\n\nPolicy: don't start a test without these three answers.\nHannah is a marketer; plain words please.","rules":[{"type":"plainText"},{"type":"maxWords","max":100}],"mistakes":[{"title":"Agreeing to start Monday","why":"The AI says 'sounds great, let's start Monday'.","fix":"Add: \"Don't start until she answers the three questions.\"","checks":["r0"]},{"title":"Missing the runtime","why":"The AI doesn't say it takes about 49 days.","fix":"Add: \"First sentence: 39,000 / 800 = about 49 days.\"","checks":["r1"]},{"title":"Vague questions","why":"The AI asks 'what are your goals?'","fix":"Add: \"Q1 the deciding metric, Q2 who is included, Q3 smallest lift worth shipping.\"","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":["hannah"]},{"label":"Asks for the runtime","patterns":["49 days","39,?000 ?/ ?800"]},{"label":"Asks for three Q lines","patterns":["q1:","q2:","q3:"]},{"label":"Rules out promises","patterns":["don.t promise","no promise"]},{"label":"Asks for the subject","patterns":["subject"]}],"hint":"A good test reply gives the time cost with its arithmetic, then asks only the missing decisions.","mission":"Reply to Hannah about the headline test.","musts":["Runtime with division","Three Q lines","No promises"],"rubricLabels":["The email starts with a Subject: line and the first body sentence says the test would take about 49 days, showing 39,000 visitors / 800 per day","There are exactly three question lines starting Q1:, Q2: and Q3: asking which single metric decides, which visitors are included, and the smallest lift worth shipping","The email does not promise the headline will lift sign-ups, and does not start Monday without the answers"]}