Avg Units per Order = DIVIDE ( [Total Units], [Order Count] ) —
3.033 (259,421 ÷ 85,538). Pure composition: two existing measures and a safe
division. No new iteration was needed because both ingredients already existed as answers.
Avg Line Profit = AVERAGEX ( Sales, Sales[Quantity] * ( Sales[UnitPrice] *
( 1 - Sales[DiscountPct] ) - RELATED ( Products[UnitCost] ) ) ). The identity: an
average is the sum divided by the count, so Avg Line Profit × 165,042 lines must reproduce
Gross Profit ($38,714,371.79) to within rounding. If yours does not, the per-row expression
differs from the one inside Total Cost/Total Revenue — same math, or the identity breaks.
Free consistency checks like this are how measure families audit themselves.
Three sorts, verified: Revenue: Furniture leads ($24.80M).
Profit dollars: Furniture still leads — its revenue bulk outweighs its thin
margin. Margin %: Decor leads (63.95%), Furniture last (55.56%).
Gross Profit crowns Furniture: profit dollars = revenue × margin, and
Furniture's $24.80M at 55.56% (~$13.8M of profit) beats Decor's $5.02M at 63.95% (~$3.2M).
A high percentage of a small number lost to a low percentage of a huge one — the columns
explain each other, which is why the table shows all three.
Reflection answers
The decider: whether the answer's ingredients already exist as measures.
Basket size divided two finished answers. Average line profit needed arithmetic inside
each row before any aggregation — per-row math means an iterator's row context, every
time.
The honest first response is a question back: "best by which measure? Revenue and
profit dollars say Furniture; margin rate says Decor — and they lead for different reasons."
Third appearance of the course's oldest lesson: units crowned Lighting in Lesson 1, dollars
crowned Furniture, margin crowns Decor. The measure defines the winner; the analyst names the
measure.