Finding The Maximum Discount Depth Before Per-Order CM Goes Negative
The walk-down method merchandising leads use to find the precise discount % where per-order contribution margin flips negative — and the guardrail that comes out of it.
Quick answer
Walk your contribution margin calculator down in 5% discount increments at your real AOV, COGS %, pick/pack, payment fee, and return rate. The last increment where per-order CM is still positive — minus one step for safety — is your discount floor. For most apparel stores with 20-25% return rates, that floor lands between 25% and 30% off; for beauty with low returns and bundle economics, it sits closer to 40-45%.
Maximum Discount Depth Before Negative Contribution Margin
The largest sitewide or sticker discount you can run before per-order contribution margin crosses zero at your real cost structure.
Maximum discount depth is the cliff edge: the percentage off where the next 5% step turns a per-order contribution-margin-positive promo into a loss-making one. Finding it requires a walk-down — plug your real AOV, COGS %, pick-and-pack, payment processing, shipping, and forecast return rate into a CM calculator, then drop the discount in 5% steps until CM hits zero.
The answer is not one number. It moves with return rate, free-shipping policy, and whether the discount is realized or sticker. The output is a written guardrail your merchandising and growth teams agree on before a promo calendar gets approved.
Most stores discover this cliff the wrong way: a Black Friday post-mortem where revenue looks great and CM looks terrible. The walk-down moves the discovery to before the promo, not after.
Why the cliff exists (and why it moves)
Contribution margin per order is what's left after variable costs: COGS, pick-and-pack, payment fees, outbound shipping, and the prorated cost of returns. A discount cuts revenue but doesn't cut any of those costs — every variable cost stays the same in absolute euros while the top line shrinks.
That's why the cliff feels non-linear. The first 10% off only eats your gross margin headroom. The next 10% starts eating into the fixed-in-absolute-terms variable costs, and CM drops faster than the discount %. By the time you're at 35-40% off, one extra step can flip the entire order negative.
Return rate is the silent multiplier
A returned order costs you the inbound shipping, the outbound shipping, the pick-and-pack on both legs, and often the unit itself if it can't be resold. At a 25% return rate, your effective discount on the cohort is meaningfully deeper than the sticker — see realized vs sticker discount for the gap. This is why apparel hits the cliff earlier than beauty.
How to run the walk-down
Start with last quarter's actuals, not your target P&L. Pull AOV, blended COGS as a % of revenue, average pick-and-pack cost, payment processing %, shipping cost per order, and the return rate for the category you're about to discount. Use the category-specific return rate — sitewide averages hide the apparel problem.
Then walk the discount from 0% down in 5% steps. Record per-order CM in euros and as a % of discounted revenue at each step. The last step where CM is still comfortably positive (we use ≥ €4 or ≥ 8% of discounted revenue, whichever is lower) is your operating floor. The step after that is the cliff.
Illustrative walk-down: apparel store, €85 AOV, 48% COGS, €4.20 pick/pack, 2.9% payment fees, €6 shipping, 25% return rate
| Discount % | Discounted AOV | Per-order CM (€) | CM % of revenue |
|---|---|---|---|
| 0% | €85.00 | €18.40 | 21.6% |
| 10% | €76.50 | €11.90 | 15.6% |
| 20% | €68.00 | €5.40 | 7.9% |
| 25% | €63.75 | €2.15 | 3.4% |
| 30% | €59.50 | -€1.10 | -1.8% |
| 35% | €55.25 | -€4.35 | -7.9% |
| 40% | €51.00 | -€7.60 | -14.9% |
In this example, 25% off is the last positive step and 30% off is already CM-negative. Your written guardrail becomes: "no sitewide discount past 25% without an exception sign-off." That's the artifact you actually need.
Common detection signals you're already past the cliff
Three signals show up in the analytics before the finance team flags it. First: discounted-cohort AOV trending up but blended CM trending down month over month. Second: promo days posting record revenue but worse-than-baseline contribution per visitor. Third: return rate on promo orders 3-6 points higher than full-price orders — discount hunters are also the most aggressive returners.
If you're stacking free shipping on top of a sitewide percentage, you have a second cliff layered on the first. The two costs compound because shipping is a flat euro amount on top of the % off — see why stacking free shipping on a sitewide % off collapses CM twice for the mechanism.
Category matters more than you think
An apparel store at 25% return rates hits the cliff around 28-30% off. A beauty or skincare store with 4-6% returns and bundle economics often has a cliff closer to 42-45% off. Don't import another vertical's discount floor — run the walk-down on your own numbers.
How to fix it without killing the promo
Switching from sitewide % off to threshold-gated free shipping lifts AOV and protects CM at the same time, because the discount only triggers on orders that already cleared a margin bar. Tiered discounts (10% off €60+, 15% off €100+, 20% off €150+) do the same — see sitewide % off vs threshold-gated free shipping at the CM cliff for the trade-offs.
If you genuinely need to go past the cliff — competitor-led BFCM pressure, end-of-season inventory you can't carry — use first-order CM-adjusted LTV to justify a deeper discount cliff. The math only works if you have hard evidence the acquired cohort repurchases at full margin within 90-120 days. Otherwise it's a story, not a calculation.
Frequently asked questions
With typical 48-55% COGS and 20-25% return rates, the CM cliff sits between 25% and 30% off. Written guardrails of 25% sitewide max, with exceptions requiring a CFO sign-off, are common in the €1M-€15M revenue band.
Yes, but less than you'd think. Higher AOV spreads the flat-euro costs (pick/pack, shipping) over more revenue, so the cliff moves slightly deeper. A €120 AOV apparel store might hit the cliff at 30% instead of 27%. COGS % and return rate move the cliff far more than AOV does.
Realized. Customers stack codes, qualify for free shipping thresholds, and return at higher rates on promo orders. Realized vs sticker discount routinely shows a 5-8 point gap, and the realized number is what your CM actually experiences.
Each return costs you roughly 2x the variable cost of a kept order (inbound + outbound + double pick/pack + restock or write-off). A category with 25% returns vs 10% returns will see the cliff move 8-12 points shallower. See how return rate shifts the CM-negative discount cliff for the detailed mapping.
Per-order CM is the operational guardrail — it's what the calculator outputs and what you can hold a discount calendar against. Cohort CM (with repeat purchases factored in) is the strategic view used to justify exceptions. You need both.
Bundles let you discount the sticker more deeply because COGS % drops with batch fulfilment, and returns on bundles are typically 2-3% vs 5-6% on singles. The cliff for a 3-item skincare bundle can sit at 40-45% off, well past the single-SKU floor.
Quarterly at minimum, and any time COGS shifts by 3+ points or shipping rates change. Most stores recalculate before the BFCM planning meeting in September and again after the Q1 returns wave clears in February.
Economically, yes. A €6 shipping subsidy on a €60 order is a 10% effective discount. When you stack free shipping on top of a sitewide percentage you collapse CM twice — the percentage cut and the absorbed shipping euros.
If the realized discount is materially lower than the sticker (e.g. minimum order thresholds, exclusions, single-use codes), a 40% off banner can outperform a 25% off banner on conversion while landing at the same CM. Anchor-price perception is the lever — see anchor-price perception: why a 40% off sticker outperforms 25% off even when CM is the same.
Cap the depth on hero SKUs only, gate the deepest tier behind a minimum spend, and budget the CM hit as a paid-acquisition cost — but only if cohort LTV evidence supports it. See what to do when BFCM pressure demands a discount past your CM floor for the playbook.
Test ideas before you ship them
Run unlimited A/B tests, attach hypotheses to outcomes, and build a searchable archive of what works — and what doesn't.