Menu Performance and Experiment Workspace — example input brief

Purpose: build a menu performance diagnosis, experiment and approval workspace.
This is a tutorial input brief, not a complete data export or production dataset.
Project: https://cd422b3721b541ad977d070135fb8b6b.prod.enter.pro/

Input tables: Items; Stores; Channels; Daily Item Sales; Recipe Costs; Promotions; Stockouts; Experiments; Reviews; Observations; Users.
Sales fields: date, store_id, item_id, channel, daypart, promotion_flag, units, gross_sales, discounts, refunds, ingredient_cost, marketplace_fees, stockout_slots.
Rules: net_sales = gross_sales - discounts - refunds; contribution_after_ingredients = net_sales - ingredient_cost; contribution_after_fees = contribution_after_ingredients - marketplace_fees; food_cost_pct = ingredient_cost / net_sales. Compare the current period with an equal-length baseline on the same scope; neither contribution measure is net profit.
Workflow: like-for-like comparison -> item diagnosis -> experiment proposal -> finance review and operations readiness -> launch -> 7/14/28-day observations -> scale, iterate, roll back or archive.
Roles: Menu Director, Finance Analyst, Operations Manager. A proposal stays "not applied" until both reviews approve.

Observed example, 2026-10-12 (data snapshot Sep 1, 2026): Harborline Kitchen, stores S01-S24, channels Dine-in, Takeout and Marketplace. Current period 2026-08-04 to 2026-08-31; baseline 2026-07-07 to 2026-08-03. Truffle Mac & Cheese (F002, $16.95): 21,376 units and $362,323.20 net sales in both periods; ingredient cost per unit $5.25 -> $7.15; contribution after ingredient cost and fees $235,185.60 -> $194,571.20 (-$40,614, -17.3%). EX-F03 Price Test: proposed +$1.00 in 6 dine-in stores, Sep 8 - Oct 5, 2026; no concurrent control group, no significance test; Finance review and Operations readiness pending; no outcome recorded.
If additional demo data is needed, generate clearly labelled synthetic records with consistent IDs and dates.