{"id":"data2-kpi-better-worse-flat-labels","audience":"managers","category":"data","track":"data","technique":"verification","level":"beginner","title":"Label five KPIs as better, worse or flat","free":false,"task":"Your ops director, Sasha, wants to see at a glance how the week went. Write exactly 5 plain-text lines, one per KPI, in the form 'KPI: Better, Worse or Flat - last week to this week'. A change under 2% of last week's value counts as Flat. Remember that for some KPIs lower is better. Under 80 words, nothing else.","materialLabel":"KPI table (attached to your prompt automatically)","material":"Weekly operations KPIs - Granite Fulfillment\n\nKPI | Last week | This week\nOn-time delivery | 91% | 94%\nReturn rate | 4.1% | 5.0%\nAverage handle time (minutes per ticket) | 6.0 | 5.5\nOrders shipped | 1,000 | 1,015\nCost per order | $8.00 | $8.60\n\nReference notes:\n- Return rate and handle time and cost per order are things we want to go down.\n- On-time delivery and orders shipped are things we want to go up.\n- Sasha reads these in 20 seconds on her phone, between meetings.\n- Last week was a normal week; there was no holiday.","rules":[{"type":"plainText"},{"type":"maxWords","max":80}],"mistakes":[{"title":"Treating up as good","why":"The AI calls Return rate and Cost per order improvements because they rose.","fix":"Add: \"Return rate, handle time and cost per order: lower is better.\"","checks":["r0"]},{"title":"Fuzzy flat","why":"The AI calls small changes 'slightly better' or marks a 1.5% move as Better.","fix":"Add: \"Flat means the change is under 2% of last week's value.\"","checks":["r1"]},{"title":"Wrong layout","why":"The AI writes a table with commentary and a summary.","fix":"Add: \"Exactly 5 lines like 'On-time delivery: Better - 91% to 94%'. No notes.\"","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":"Gives the labels","patterns":["better","worse","flat"]},{"label":"Handles direction","patterns":["lower is better","go down","want to go down"]},{"label":"Defines flat","patterns":["under 2%","2%"]},{"label":"Shows the line shape","patterns":["91% to 94%","like ['\"]?on.time","exactly 5"]},{"label":"Rules out extras","patterns":["no (intro|notes)","nothing else","start with"]}],"hint":"Direction matters. Say which KPIs are better when they fall, and define Flat with a number.","mission":"Which KPIs got better or worse?","musts":["Five lines","Lower can be better","Flat under 2%"],"rubricLabels":["On-time delivery is Better, Average handle time is Better, and Orders shipped is Flat (up 1.5%, under the 2% threshold)","Return rate is Worse (4.1% to 5.0%) and Cost per order is Worse ($8.00 to $8.60), not Better just because the number went up","There are exactly 5 lines, each showing the KPI name, one of the labels Better, Worse or Flat and both values (last week to this week), with nothing before or after"]}