{"id":"dat-mkt-email-send-readout","audience":"marketing","category":"data","technique":"context","title":"Write a 3-sentence readout on which email send won","free":false,"task":"Your director wants a 3-sentence readout on last month's three email sends. The first sentence gives each send's click rate, computed on delivered emails (not sent); the second names the winner. Open rates are inflated by mail privacy settings, so don't pick a winner on opens. At most 3 sentences, plain text, under 75 words, and use whole percents.","materialLabel":"Email platform export (attached to your prompt automatically)","material":"OCTOBER EMAIL SENDS - export from Mailcrest\n\nSend | Sent | Bounced | Opens | Clicks\nFall Sale (Oct 3) | 10,000 | 2,000 | 2,800 | 560\nNew Arrivals (Oct 12) | 8,000 | 0 | 4,000 | 320\nLoyalty Perks (Oct 24) | 5,200 | 200 | 1,750 | 300\n\nTeam notes:\n- Fall Sale went to the old list import, which is why bounces were so high. That list is now cleaned.\n- Opens include Apple Mail privacy auto-opens, so open rate runs high and is not reliable. We judge sends on click rate.\n- Our click rate is clicks ÷ delivered (sent minus bounced).\n- Loyalty Perks had the best subject line according to Jess (she just likes it).\n- The November calendar is in the shared drive.\n- New Arrivals got a shout-out in the all-hands for \"great engagement.\"","rules":[{"type":"maxSentences","max":3},{"type":"plainText"},{"type":"maxWords","max":75}],"mistakes":[{"title":"Not saying how to measure","why":"The AI either picks by opens or divides by emails sent, and crowns the wrong send.","fix":"Add: \"Winner by click rate = clicks ÷ delivered (sent minus bounced).\"","checks":["r0","r1"]},{"title":"Getting fooled by opens","why":"New Arrivals has a 50% open rate, so the AI calls it the star.","fix":"Add: \"Opens are inflated, don't pick a winner on them.\"","checks":["r2"]},{"title":"Leaving the length open","why":"The AI writes a full report with a table and next steps.","fix":"Say: \"At most 3 sentences, winner first.\"","checks":["maxSentences"]},{"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: \"each sentence under 18 words.\"","checks":["maxWords"]},{"title":"Forgetting where it goes","why":"AI loves markdown tables and bold, which show up as clutter in a director's inbox.","fix":"Add: \"Plain text, no bold, headings or tables.\"","checks":["plainText"]}],"promptHints":[{"label":"Defines click rate","patterns":["delivered","bounce","clicks? ?(÷|/|divided)"]},{"label":"Warns about opens","patterns":["opens?.{0,30}(inflat|unreliab|ignore|not)","privacy"]},{"label":"Says what each sentence holds","patterns":["sentence 1|first sentence","sentence 2|second sentence","then"]},{"label":"Sets a word limit","patterns":["\\d+\\s*words","word (limit|count)","under \\d+"]},{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no (headings|tables)"]}],"hint":"The AI can't know your team measures on clicks over delivered. Tell it, and tell it what each sentence holds.","rubricLabels":["Names Fall Sale as the winner with a 7% click rate (560 ÷ 8,000 delivered), right after the rates","Any other click rates given are on delivered emails: New Arrivals 4%, Loyalty Perks 6% (not 5.6% or 5.8% on sent)","Does not call New Arrivals the winner for its high open rate; if opens come up, they are described as unreliable"]}