What percent discount is actually good? A category-by-category benchmark

“20% off” is a great deal on a fridge. It’s an everyday baseline on a sweater. “50% off” is unusual on a phone, expected on home decor, and suspicious on a luxury handbag. Discounts only mean something against a category baseline — and the baselines vary dramatically.

The category benchmarks

Why the baselines vary so much

The biggest factor is margin. Categories with high markups have room to discount aggressively:

• Fashion typically has 60-70% gross margins. A 50% discount still leaves the retailer profitable.
• Mattresseshave legendary margins — which is why every mattress store has a permanent 50% off sale that’s never actually 50% off the “real” price.
• Electronics have ~5-15% gross margins (consumer brands) or ~5% on big-box retail. A 20% discount means the retailer is selling at cost or below — these discounts only appear as loss leaders during holiday sales.
• Apple products have intentionally narrow discount channels. The first-party Apple Store rarely discounts more than ~5%; the deepest discounts come from Best Buy, Amazon, and Costco on last-gen models.

How to spot a fake discount

1. Check the price history.Browser extensions like CamelCamelCamel (Amazon), Keepa, or Honey show you the actual price over the last 90 days. A “50% off” sale that’s been running for 6 months is the normal price — the “original” is the fake.
2. Compare to MAP.If multiple retailers all show the same price and the “sale” is identical, you’re at MAP. That’s the floor; no further discount is coming.
3. Beware “everyday X% off.” The original price was set with that discount baked in. The discount isn’t a discount — it’s the baseline price.
4. Compare across retailers.If the same SKU costs $800 at Retailer A with “30% off” and $600 at Retailer B with no discount, Retailer B is your better deal.

When to wait for a deeper discount

Some discount patterns are seasonal and reliable:

• Last year’s electronics: 6-9 months after new model launch. Apple drops the previous iPhone by ~$100 when the new one ships.
• Fashion: 2-3 months after season change. Spring → late summer. Fall → late winter.
• Home goods: post-holiday clearance, late January.
• Mattresses: Memorial Day, July 4th, Labor Day (US). Beyond that, mattress discounts run year-round.
• SaaS annual plans: Black Friday, end-of-fiscal-quarter promos.

Worked example: stacking a coupon on top of a sale

Scenario from a US outdoor retailer in November 2025: a $189.00 rain jacket marked “30% off”, a loyalty-program 15% coupon code, and 7.25% California sales tax. What do you actually pay, and what’s the effective discount?

1. Sticker: $189.00
2. After 30% off: $189.00 × 0.70 = $132.30
3. After 15% loyalty coupon: $132.30 × 0.85 = $112.46
4. Plus 7.25% sales tax: $112.46 × 1.0725 = $120.61

Effective pre-tax discount: ($189.00 − $112.46) / $189.00 = 40.5%. Notice that’s less than the 45% you’d get by adding 30 + 15. Stacked percentage discounts always multiply, never add — see our discount calculator for the multi-stack tab.

If the loyalty code had instead been “$15 off” flat, the result would have been $132.30 − $15.00 = $117.30, a worse outcome by ~$4. The general rule: on items over $100, flat-dollar coupons usually beat percentage coupons; on items under $50, percentage coupons usually win. Cross-over is roughly at coupon-value ÷ discount-rate.

Common mistakes when evaluating “deals”

• Comparing to the “original” price rather than the typical price.Most online listings show MSRP as the strikethrough “was” price. The real reference is the median price over the last 90 days. The US FTC’s 16 CFR Part 233 explicitly prohibits using inflated reference prices, but enforcement is sporadic and the burden of proof is on the regulator.
• Forgetting the return-shipping asymmetry. A 30%-off fashion item with $15 return shipping kills the discount if it doesn’t fit. Always price-in the expected return cost when sale-shopping clothing you can’t try on.
• Buying in bulk on a “buy 2 get 1 free” for items you didn’t need. The headline 33% discount is real (3 for the price of 2), but if you only wanted one, you paid 100% more than necessary for the two extra units.
• Ignoring credit-card cashback in the comparison. A 2% cashback card on the full-price item often beats a 5% sale at a retailer where your card earns 0%. Stack both when possible.
• Falling for “up to X% off” headlines. The advertised maximum is the steepest discount on the deepest-discounted SKU in the sale. Median sale-item discount is typically 30-50% of the headline number, per FTC and CMA guidance.

When the “good deal” benchmark does NOT apply

Three categories where the standard table breaks down:

• Luxury and limited-edition goods. Hermès, Rolex, Patek Philippe never go on sale at first-party retail. The “30% off Rolex” listing is almost certainly grey market with no warranty, or counterfeit. The right baseline for these categories is “0% off list, possibly with waitlist” — any discount is the suspicious data point.
• Goods with secondary-market price discovery.Trading cards, sneakers (StockX), and collectibles trade above retail at launch. A 30% retail discount on a launch-day Jordan model often still leaves the retail price below resale. The “is this a good deal” question collapses into “is the resale floor above the discounted price?”
• Loss-leader staples.Costco rotisserie chicken at $4.99, Trader Joe’s bananas at $0.19 — deliberately below cost to drive traffic. Comparing their “regular price” to anything else isn’t informative because they’re not priced to be compared.

The math, regardless of category

Use our discount calculator for the actual sale price and savings. For stacked discounts (10% promo + 15% loyalty), remember the math is multiplicative, not additive — 10% + 15% stacked is 23.5% effective, not 25%. See our finance methodology for the formula.

Loyalty programs and effective discount rate

Annualising the value of a loyalty programme:

• Sephora Beauty Insider:Rouge tier (spend $1,000/year) earns ~10-12% back in points and freebies. Effective discount: ~11% on the categories you’d have bought anyway.
• Starbucks Rewards: ~5% back in stars. Below the 2% cashback of a standard credit card.
• Amazon Prime: $139/year for free shipping + Prime Video + other benefits. Break-even for a household making 8-10 Prime-eligible purchases per year; below that, a Prime-free workflow with $35-threshold free shipping is cheaper.
• Costco Executive: $130/year for 2% cashback on most purchases. Break-even at ~$3,250/year of Costco spending; positive return beyond that.
• Airline status programmes: typically worth 5-10% in upgrades, lounge access, and waived fees — but only if you fly enough to maintain status. Below the threshold, the programme is a discount of zero.

Discount math for compound and stacked offers

Three rules for the math that actually appears at checkout:

• Sequential discounts multiply. “30% off, plus an extra 20% off at checkout” isn’t 50% off; it’s 1 − (0.70 × 0.80) = 44% off. The two discounts apply sequentially.
• Tax stacks on the discounted price. Sales tax computes on the net price after discount. A $100 item at 40% off in California (7.25% tax) settles at $60 × 1.0725 = $64.35, not $107.25 × 0.60 = $64.35 (the math happens to agree because multiplication is commutative, but mental ordering matters).
• Free shipping has a dollar-equivalent. Most retailers charge $5-15 for shipping on small orders. A “free shipping at $50” threshold is effectively a 10-30% discount on a $50 order that you wouldn’t have placed without the threshold — factor this into discount comparisons.

The seasonal-discount calendar (US)

Most US categories follow predictable annual cycles. The sharpest discounts in each category, by month:

Two practical rules from this calendar: buy seasonal goods at the end of their season (winter coats in March, patio furniture in September), and buy gift items just before or just after the gifting peak, never during it. The exception is Black Friday electronics — genuine doorbuster pricing on TVs and laptops typically beats any other point in the year for the same SKU.