Discount Impact on Margin
Date: 2026-08-16
A discount comes entirely out of contribution, so a small percentage off price is a large percentage off profit. The arithmetic of how much extra volume it must generate to break even is the calculation nobody does before running the promotion.
What it is
The margin impact of a discount is the reduction in contribution per order, and the volume increase required to offset it.
The mechanism, and it’s the whole note:
full price 10% off
revenue £50.00 £45.00
− cost of goods £27.50 £27.50 ← unchanged
− variable costs £7.50 £7.40 ← only fees fall
──────── ────────
contribution £15.00 £10.10
contribution lost £4.90 = 33%
A 10% discount removes a third of the contribution. The discount comes out of margin alone, because the cost of the goods doesn’t change.
The break-even volume
contribution volume uplift needed
10% off £10.10 +49%
20% off £5.20 +188%
25% off £2.75 +445%
30% off £0.30 effectively unachievable
A 20% discount needs orders to nearly triple just to stand still. That almost never happens, and it’s why blanket discounting destroys profitability while looking like a successful campaign in a revenue dashboard.
In plain terms: the lower your margin, the more devastating a discount is. At 30% contribution margin, a 10% discount costs a third of your profit per order.
Cannibalisation makes it worse
The volume calculation above assumes every discounted sale is incremental. Most aren’t:
1,000 orders during a 10% promotion
700 would have bought anyway, at full price
300 genuinely incremental
cost of the 700 700 × £4.90 lost = £3,430 given away
gain from the 300 300 × £10.10 = £3,030 earned
────────
−£400
A promotion that lifted orders by 43% lost money, because most of the discount subsidised customers who needed no discount.
This is why Incrementality Testing matters more for promotions than for almost anything else — and why a holdout on a promotional email is one of the highest-value tests available.
Targeted beats blanket
The arithmetic above is the argument for discounting narrowly:
- Lapsed customers only — no cannibalisation, because they weren’t buying — Winback Campaigns
- Basket-threshold discounts, which require incremental spend to unlock
- Time-limited on slow-moving stock, where the alternative is a write-off
- First-order codes, accepting they raise CAC — paid to every new customer, including those who’d have come anyway — and lengthen payback — Payback Period
And against discounting habitually: customers learn. A predictable promotional calendar trains people to wait, which converts full-price demand into discounted demand permanently — Promotional Cadence, Discounting Strategy.
Before running one
Four questions, and the first is the one that’s usually skipped:
- What volume uplift breaks even? Compute it. If it’s above +50%, the promotion is very unlikely to pay
- How much is incremental? Hold out a random slice and find out
- What does it do to mix? Discounting a low-margin category is worse than the headline suggests — Basket Composition
- What’s the next one? Each promotion sets the expectation for the following quarter
Worked through end-to-end in Guide - Unit Economics and Commercial Decisions.