Calm Serum - 30ml: runs out in 4.1 days (≈ 39.1/day, supplier lead 21 days + safety 7). Order 2112 units (176 packs of 12). Night Oil: runs out in 20.4 days (≈ 14.7/day, supplier lead 21 days + safety 7). Order 564 units (47 packs of 12).
Saves ~17 days over 21 days
CS30: reorder point = 39.1/day × (21 + 7) = 1095, on hand 160 + on order 0; quantity covers lead + safety + 30 days, packs of 12, at least 2 packs; demand ×0.93 (last year's SKU pace ×0.91, planned ad budget ×1.02). NO1: reorder point = 14.7/day × (21 + 7) = 412, on hand 300 + on order 0; quantity covers lead + safety + 30 days, packs of 12, at least 2 packs; demand ×0.87 (last year's SKU pace ×0.85, planned ad budget ×1.02).
ugc_night_h3_v1: frequency 3.6 and rising, CTR 1.46% → 0.95% (-35%) over 7 days on $1,155. A new hook on the same angle is requested from the Creative strategist.
Frequency ≥ 3 and rising, CTR down ≥ 25% between the first and last 3 days, spend ≥ 1 × target CAC. Keep it live until the new version beats it.
B wins: Sticky add to cart button on mobile product pages
Sticky add to cart button on mobile product pages · day 12 · A 4.80% vs B 5.43% · +13% · confidence 98%. This test runs outside the engine: apply the result in the theme or the testing tool.
+13% conversion
B wins at ≥ 95% confidence with ≥ 30 orders per variant and ≥ 7 days; no winner after 21 days stops the test.
Code MAYA20 costs 31% of its sales: move to the standard terms
Code MAYA20 costs 31% of its sales (target 25%, 77 orders, $4,932 gross over 4 weeks). Move to the standard terms: 15% discount, 10% commission. Same volume, cost 23% of sales.
+$393 margin a month
Cost = discounts $934.80 + commission $575.13 = 30.6% of gross. SCALE at 22% or less with 80+ orders, HOLD up to 25%, FIX up to 30%, KILL above. Partners on legacy terms are shown for information: the standard terms are proposed at renewal.
Run a 21-day geo holdout on California to measure what ads really bring
Stop ads in California (18% of revenue over 28 days) from 2026-10-15 to 2026-11-04 on every ad account, keep everything else unchanged, then compare their sales with the other regions.
Saves ~$3,331 over 21 days
Platform attribution overstates and understates. A holdout of 18% of revenue for 21 days gives the real effect of ads. Expected readout: platforms credit ads with 63% of sales in these regions; if that is real, sales there fall by about 63% against the others.