What is promotion net effect?
Promotion net effect asks the only question that matters about a discount code: after the giveaway, did the orders it touched still make money? It is the contribution margin of a promotion's orders — a figure that already has the discount taken out of it, the same way it has COGS, shipping and fees taken out. A negative net effect flags a promotion whose orders, even counting all their margin as the promotion's doing, did not earn enough to cover what it gave away.
Formula
net effect = contribution margin of the orders the promotion touched
(that margin already nets out the discount given on them)
Worked example
A 20%-off code touches 340 orders, giving away €8,600 of discount, and those orders' contribution margin — after COGS, shipping, fees, AND that discount — comes out negative. The promotion paid customers to take margin away, and that is the generous reading, since some of those 340 orders would have happened at full price anyway. The worst-promotions widget lists exactly these negative rows, worst first, alongside how much was given away so the two figures can be read side by side.
How Saldo Metrics computes it
v_promotion_order_effect joins each discount application to its promotion and
order at (promotion, order) grain, carrying the discount amount and the order's
contribution margin under the order-grain cost rule: an order with any uncosted
line has unknown margin (NULL), not zero, and its revenue is reported as uncosted
rather than blended in. That order margin already has the order's discount
allocated out of it at the line level (the same contribution_margin_base every
other margin figure reads), so net effect is that margin as-is — not margin minus
the discount again, which would count the giveaway twice and made a break-even
promotion look like a money-loser equal to its full discount (fixed 2026-09-28).
The worst-promotions widget aggregates the view per promotion over the selected
period and keeps only promotions whose (already-net) margin is negative; the
alert metric evaluates the same figure per promotion. An order touched by two
promotions appears under both — the effect is per promotion, not a partition of
the order book.
Why it matters
Promotions are the easiest place in e-commerce to lose money while revenue charts go up. Discount reporting usually stops at "revenue generated"; measuring against contribution margin, net of the giveaway, is what catches codes that work brilliantly on volume and terribly on profit — stackable codes and storewide percentages being the usual offenders.
Common mistakes
- Crediting the promotion with all touched-order margin. Many of those customers would have bought anyway; the true incremental effect is worse than the reported one. Treat the net effect as an upper bound.
- Judging a launch-week code on its first period. Acquisition promotions can run negative on first orders and pay back on repeat purchases — check the cohort economics of customers it acquired before killing it.
- Ignoring uncosted orders. A promotion selling mostly uncosted SKUs shows less margin than it earns; the view reports that revenue as uncosted rather than as zero margin, but the net effect is still incomplete until costs land.
Where you see this in the app
The Worst Promotions dashboard widget, and the promotion net-effect alert metric.