Practice — Hot · Solution

1 · "Bigger orders," both readings — verified

By revenue per order, the claim is false — trade orders are the smallest, $105 below Consumer. By units per order it is barely, meaninglessly true: 3.040 versus 3.033, a difference of less than one item per hundred orders. Trade designers do not buy bigger baskets; they buy the same baskets at a discount, which mechanically drags their revenue-per-order below everyone else's. (Basket size needed no new measure: [Total Units] and [Order Count] against the Segment dimension — filter context did the per-segment work.)

2 · What the discount costs

The trade discount costs about 6.5 margin points versus Consumer — every trade dollar earns roughly 53 cents of gross profit against Consumer's 59. Since baskets are the same size, that margin gap is the discount's entire visible effect on order economics.

3 · The steelman

The strongest true fact for the program: 13,244 orders and $9.22M of revenue — more than twice Corporate's volume — from a channel that presumably costs little to serve and may repeat-purchase reliably. The program moves real volume; what the data does not show is whether that volume required the discount, and that is the honest limit of this analysis.

4 · A verdict for the CFO

"The bigger-orders claim doesn't hold: trade baskets contain the same three items as everyone else's — 3.04 units per order against Consumer's 3.03 — and because of the discount, trade orders are actually our smallest by revenue at $696 against Consumer's $802. The discount currently costs about six and a half margin points. What the program genuinely delivers is volume: thirteen thousand orders and $9.2M we might not otherwise see. But deepening the discount to chase bigger baskets would spend margin on a behavior this data says the discount doesn't drive. Before any deepening, I would want to know what trade customers do when the discount changes — retention and order frequency, which this dataset doesn't yet answer."

Reflection answers

• The trigger phrase is any comparative without a measure attached — "bigger," "better," "biggest," "growing." The reflex: bigger by what measure? — asked before the first formula, every time. Three lessons, three instances: units vs dollars, line vs order, revenue vs units-per-order.
• The week's real work happened in the definitions. Once AOV, margins, and counts existed as named measures, the entire trade-program trial was drag-and-drop against the Segment column — filter context multiplied the family across a new dimension for free. Good measures are written once and answer questions forever; that is the compounding return this module is building.