Every merchant we work with runs discounts. Almost none of them can tell you what the discounts are for.
So we measured it. Not a survey, not a vibe. We pulled 24 months of raw data from 10 of our retainer clients' Shopify stores: $285 million in gross sales, $29.4 million in discounts, and then went one layer deeper and dissected the discount anatomy of 1.14 million individual orders. Established brands, seven and eight figures, across five unrelated consumer verticals.
The headline: merchants gave back 10.3 cents of every gross dollar, and most of that spend can't be doing the job merchants assign to it. (One methodology note up front: we excluded draft and $0-total orders, which are manually priced orders, comps, and exchanges rather than promotions. They carried 29% of raw discount dollars, so every number below runs on the cleaned sample.) The full report with all twelve charts is on the State of Discounts 2026 page. Here's the short version.
Blanket discounting is literal
New customers get discounted on 41.6% of their orders, at $23.96 per order. Returning customers: 41.3%, at $24.12. That's not a strategy, that's a spray. And because returning customers place two-thirds of the orders, they collect 67.3% of the discount dollars, every one of which subsidized someone who had already decided to buy.
The bit that should sting: price-discounted orders have a lower new-customer share (30.2%) than full-price orders (32.6%). The "acquisition tool" acquires less than no tool at all.
Everyone picked 20%, and depth past 20 buys nothing
Median discount depth across 471,000 discounted orders: 19.96% of the basket. Five unrelated industries herding on one number nobody priced.
One chart ends the "20 or 30 percent?" meeting: pre-discount baskets are flat from 0% through 30% depth ($143 to $159), then fall. Deeper discounts attach to the same carts or smaller ones. Every point of depth past about 20 is margin donated to a basket that was coming anyway.
The leak is always-on, not Black Friday
November is 14.07% of annual discount dollars against 13.13% of annual gross. Barely above pro-rata. 86% of the leak is the other eleven months, led by evergreen codes: the biggest single code in the sample gave back $349K on its own and was live all twelve months. Merchants re-plan BFCM annually and never audit the codes that rival it.
Meanwhile 99.4% of the 132,555 distinct codes redeemed were used exactly once, and that swarm carries 56% of code dollars. Unique codes are often good practice. Half your code spend being individually unreviewable is not.
The one that made us build a product
Head to head across 1.14 million orders:
- Full-price orders: $111.61 net AOV, 32.6% new customers.
- Price-discounted orders: $115.65 net AOV, 30.2% new customers, at $24.35 per order of real margin.
- Gift-with-purchase orders: $172.88 net AOV, 46.5% new customers, at a booked "cost" that is the gift's retail sticker. The merchant's real cost is COGS, which we did not measure; for typical DTC margins it lands well below the price discount's $24.35.
Gifts carry the biggest baskets and half again the price cut's new-customer mix. (Free-shipping-only orders actually post the highest new-customer share of all, 50.1% at $7.50 a head: the broader lesson is non-price incentives.) And gifts get 7% of promo dollars, against 88% for price cuts.
Disclosure, stated in the report right next to that finding and repeated here: we build Promo Party, a gift-with-purchase app. That's why we had this data and this question. Judge the numbers: every stat is script-computed from raw order data, the methodology is published in full, and the selection-bias caveat on the gift findings (threshold campaigns gate gifts to bigger carts) is printed directly under the chart it weakens. And when cleaning the sample killed one of our own early findings, we cut it rather than shipping it; the same rule applies in both directions.
What we'd do, in order
- Cap depth at 20 to 25%. The data says deeper buys nothing.
- Differentiate new vs returning. Today, effectively nobody does.
- Audit your evergreen codes quarterly. At most stores, the top 10 codes are the bulk of code spend.
- Run gifts as a program, not a November stunt.
- Check your stacking rules: stacked orders take double their fair share of dollars.
The full report has the per-store breakdowns, the trend data (7 of 10 stores are already discounting less year over year, and the heaviest discounters are cutting fastest), the risks, the FAQ, and the complete methodology: State of Discounts 2026.
No customer-level data was collected at any stage, and stores appear only as letters. If you want your store's version of these numbers, that's an afternoon of work: the methodology section tells you exactly how.
