The Double-Discount Trap: Why 20% + 30% Off Isn't 50% Off
A jacket is 20% off. At the register, a coupon takes an extra 30% off. Half price, right?
No — you’ll pay 56% of the original: a 44% discount. And the gap between the 50% your instinct computed and the 44% the register charges is not a rounding error. On a $200 jacket it’s $12, paid by everyone who adds percentages that should be multiplied.
The arithmetic
Each discount is a multiplication of whatever price is currently on the item:
- 20% off means × 0.80 → the $200 jacket drops to $160.
- An extra 30% off applies to $160, not $200: × 0.70 → $112.
Chained: 200 × 0.8 × 0.7 = 112, so you saved $88 of $200 — 44%. The second discount acts on an already-shrunk base, which is why stacked discounts always total less than their sum. The general rule for any stack, in either direction:
Convert each change to its multiplier (1 − d/100 for discounts, 1 + p/100 for increases), multiply the multipliers, and read the total change off the product.
Three stacked discounts of 10% are × 0.9³ = × 0.729 — 27.1% off, not 30%. The percent off calculator does the chain for you: run the first discount, feed the result into a duplicated row with the + button, and read the final price.
This error is measured, and retailers know it
The definitive study is Chen and Rao (2007), aptly titled When Two Plus Two Is Not Equal to Four. Across lab experiments and a field test, they found that people process sequential percentage changes by adding them — judging “20% off plus an extra 25% off” as more attractive than an economically identical single discount of 40%, because 20 + 25 = 45 feels bigger. In their field experiment, the sequential-discount framing measurably increased sales of the same product at the same effective price.
Read that from the retailer’s side: a double discount is a way to advertise a bigger number than the one you receive, at zero cost, legally, using only your arithmetic against you. “Extra X% off clearance” tags are the purest specimen — clearance prices are already discounted, so every additional percentage acts on the smallest base in the store.
A fair caveat: Chen and Rao studied consumer promotions specifically, and the effect size varies with how easy the computation is made for the shopper. The direction of the error, though, is stubbornly one-way — additive estimates always overstate stacked discounts, so the surprise at the register is always in the store’s favor, never yours.
The 10-second defense
You don’t need exact math in the aisle — you need the bias corrected:
- Add, then knock a chunk off. The true stacked discount is always less than the sum; for two mid-size discounts, subtract about a tenth of their product-of-tens (20% and 30%: sum 50, correction 20 × 30 ÷ 100 = 6, true discount 44%). That correction term is the exact formula: d₁ + d₂ − d₁d₂/100.
- Or think in what-you-pay terms. “Pay 80%, then pay 70% of that” — 0.8 × 0.7 = 0.56 — and 56% of a round number is mental-math-friendly.
- At a screen, don’t estimate. Duplicate a row in the calculator and chain the discounts exactly; click the result to copy the final price.
The same trap, mirrored, applies to increases — stacked raises and compounding returns exceed their sum — and to the famous asymmetry where a 50% loss needs a 100% gain. Those get their own treatment in why percentages confuse us, and the general habit that prevents all of them is the one worth keeping: percentages multiply; only your instinct adds.