News
Amazon Prime Day Hit $26.4B, But Seller Margin Tells a Different Story
Amazon Prime Day 2026 (June 23-26) drove a record $26.4 billion in US online spend, up 9.3% year over year, per Adobe Analytics. For sellers, a 9.3% rise in gross spend is not 9.3% more units. A record Prime Day can still erase contribution margin once you net out the deal discount, referral and FBA fees, and the ad-cost spike.
Key Takeaways
- Prime Day 2026 (June 23-26) hit a record $26.4 billion in US online spend, up 9.3% year over year, per Adobe Analytics. Day one alone did $8.3 billion, up 5.3%.
- The growth was inflation-driven, not volume-driven. Numerator found nearly half of shoppers said inflation made them more likely to shop the deals, so the lift is higher AOV and deal-hunting, not 9% more units sold.
- Gross marketplace spend tells a seller nothing about its own margin. On Prime Day a brand stacks a deal discount, a referral fee, an FBA fee, and a spike in ad cost on the same four days.
- The trap is a record-GMV Prime Day that loses money per incremental unit, because you discounted demand you would have captured at full price anyway.
- Run Prime Day on contribution margin and incrementality, not GMV. Set a per-SKU margin floor, feature the items that actually acquire new customers, and treat it as a CAC event only if the LTV math works.
If you sell on Amazon, the Prime Day headline this week was easy to feel good about: a record $26.4 billion in US online spend across June 23 to 26, up 9.3% on last year. The number is real, and it is genuinely big. It also tells you almost nothing about whether your Prime Day made money. Gross marketplace spend is a demand statistic. Your margin is a different number entirely, and it lives several lines further down.
This is the gap we spend Prime Day week walking brands through. For the full playbook see our CFO guide to Amazon Prime Day, and for how a fractional CFO for ecommerce frames event economics, read on.
What happened
Adobe Analytics put US online spend over the four-day Prime Day window at $26.4 billion, up 9.3% from the July 2025 event and slightly ahead of its own $26.3 billion estimate. Day one, June 23, did $8.3 billion, up 5.3% and the single biggest ecommerce day of 2026 to that point. Shoppers leaned into electronics, appliances, children's items and everyday essentials.
The more interesting data point came from Numerator: nearly half of shoppers said inflation made them more likely to shop the Prime Day deals, while only about 20% said it made them less likely. In other words, the growth was powered by people hunting for relief from higher prices, not by a sudden surge of new disposable income.
| Prime Day 2026 | Figure |
|---|---|
| Dates | June 23-26, 2026 (four days) |
| Total US online spend | $26.4 billion (Adobe Analytics) |
| Year-over-year growth | +9.3% vs July 2025 |
| Day-one spend | $8.3 billion (June 23) |
| Day-one growth | +5.3% |
| Adobe pre-event estimate | $26.3 billion |
| Inflation effect | ~50% of shoppers more likely to shop; ~20% less likely (Numerator) |
Source: Adobe Analytics via Fortune and Chain Store Age (spend); Business of Fashion and Numerator (inflation sentiment). Figures are aggregate US online spend across all retailers in the window, not Amazon's own sales.
Returns are quietly eating your margin. See by how much.
Get our Real Cost of Returns calculator: plug in your numbers, see the true hit per return.
Check your inbox. We'll send the Real Cost of Returns calculator shortly.
A 9% number is not a 9% number
Here is the first thing to do with the headline: take it apart. A 9.3% rise in dollars spent over a fixed four-day event is the product of three different forces, and only one of them is good news for a seller.
The first is price. Inflation has lifted ticket prices across most categories, so the same basket costs more this year. The second is mix and average order value, as deal-seekers trade up to the discounted big-ticket items (the appliances and electronics Adobe called out). The third, and the only one that scales your unit economics, is genuine incremental volume: more items actually leaving the warehouse.
Numerator's finding tells you the balance is tilted toward the first two. When half your shoppers say they are buying because prices are high, you are looking at demand that is defensive and deal-led, not expansionary. The dollars are up. The units are up by much less. And for a seller, units, not dollars, are what your costs scale against.
The seller's cost stack on Prime Day
Now put yourself on the supply side of that $26.4 billion. When a brand runs a Prime Day deal, it stacks four costs on the same units, all at once:
| Prime Day margin stack | What it costs |
|---|---|
| Deal discount | Often 20% or more off list, plus a Lightning Deal or Prime Exclusive fee |
| Referral fee | ~8% to 15% of the item price, by category |
| FBA fulfillment | Per-unit pick, pack and ship, often higher in peak periods |
| Advertising | ACoS spikes as every competitor bids the same keywords on the same four days |
Source: Amazon fee schedules and Eightx client Prime Day reviews. Discount and ACoS ranges are typical, not universal.
Any one of these is survivable. Stacked together on a discounted unit, they are how a brand posts a record Prime Day and still loses money on the margin line. We have sat with founders who celebrated a 40% revenue jump over the event, then watched the contribution-margin number go negative once the deal discount, the higher ACoS and the fees were netted out. The dashboard said triumph. The P&L said they had paid customers to take inventory.
For where Amazon margin sits before you discount it, our Amazon vs Shopify contribution margin benchmarks and the ACoS by vertical benchmarks are the baseline to start from.
What to watch next
Three things separate a Prime Day that built the business from one that just moved inventory.
- Incrementality, not GMV. The question is not how much you sold, it is how much of it you would have sold anyway. Volume pulled forward from your own ready-to-buy base, discounted, is margin you set on fire. Volume from genuinely new customers, or a first subscription with real lifetime value, is what the discount is supposed to buy. Estimate the split before you call the event a win.
- Per-unit contribution after everything. Take the deal price, subtract the referral fee, the FBA fee, the deal fee and the incremental ad cost, and look at what is left per unit. If that number is thin or negative on your hero SKUs, you ran a marketing campaign, not a profit event, and it had better be paying you back in customers.
- The inflation squeeze on COGS. The same price pressure driving shoppers to deals is also sitting in your landed cost. A 20% discount cuts deeper against a thinner gross margin than it did a year ago. Re-check your floor against this year's COGS, not last year's.
The operator takeaway
A record Prime Day is a great demand signal and a terrible margin assumption. The $26.4 billion tells you consumers still show up for a deal, even, especially, when they feel squeezed. It does not tell you that showing up made you money.
Run the event the way you would run any other capital decision. Decide in advance what each discounted SKU is for, set a contribution-margin floor and hold it, and measure the result in contribution dollars and new customers, not in GMV screenshots. Treat Prime Day as a customer-acquisition event you are willing to lose a little margin on to win a buyer worth keeping, and only on the SKUs where that math works. Everything else featured on Prime Day is just a discount you gave your own base.
If you want to pressure-test your own Prime Day numbers, our contribution margin calculator and the discount-rate-by-vertical benchmarks are the fastest way to see whether the event made money or just made revenue.
Frequently Asked Questions
how big was Amazon Prime Day 2026?
Amazon's four-day Prime Day 2026 ran June 23 to 26 and drove a record $26.4 billion in US online spend, according to Adobe Analytics, up 9.3% from the July 2025 event. Day one alone did $8.3 billion, up 5.3% year over year and the single biggest ecommerce day of 2026 to that point. The total came in slightly ahead of Adobe's $26.3 billion pre-event estimate.
was Prime Day growth driven by inflation or by more sales?
Mostly inflation and deal-seeking, not unit volume. Numerator found nearly half of shoppers said inflation made them more likely to shop Prime Day, while only about 20% said it made them less likely. A 9.3% rise in gross spend on a four-day event is a mix of higher prices, higher average order value, and more people hunting deals, not proof that 9.3% more units moved. For a seller, that distinction is the whole game.
why does record Prime Day spend not mean record seller profit?
Because gross merchandise value is a top-line number and your margin lives several lines below it. On Prime Day an Amazon seller typically stacks four costs on the same units: the Prime Day deal discount (often 20% or more), the referral fee (8% to 15%), the FBA fulfillment fee, and a spike in advertising cost as every competitor bids up the same keywords on the same four days. A brand can post a record Prime Day in revenue and still lose contribution margin on the incremental units.
what is the most common Prime Day margin mistake?
Discounting demand you would have captured anyway. Much of a brand's Prime Day volume is pulled forward from existing, ready-to-buy customers who would have paid full price in July. If you give those buyers 20% off plus a deal fee, you have simply handed margin to people who were already going to convert. The event only creates value when the discount buys something you did not already have, a genuinely new customer or a first subscription, not a repriced sale to your own base.
how should a brand decide what to put on Prime Day deals?
Feature by margin and by acquisition value, not by what is easy to discount. Put forward SKUs with enough gross margin to survive a 20%-plus cut plus fees, and that tend to bring in new-to-brand buyers or start a subscription with real lifetime value. Hold back thin-margin items and pure repeat-purchase staples your existing customers buy anyway. Set a contribution-margin floor per SKU before the event and refuse to break it for vanity GMV.
does inflation make Prime Day worse for sellers?
It can, because it squeezes both ends. Inflation lifts the input costs behind your products, so your gross margin going into the event is often thinner than a year ago, which means the same 20% discount cuts deeper. At the same time it pushes shoppers toward deals, so volume skews to your most discounted, most price-sensitive SKUs. A 9.3% bigger headline can sit on top of a per-unit margin that is actually down year over year.
how do I measure whether my Prime Day actually worked?
Measure contribution margin and incrementality, not GMV. Take Prime Day revenue, subtract the deal discounts, referral and FBA fees, and the incremental ad spend, and compare the contribution dollars to a normal four-day window. Then estimate how many of the orders were genuinely new customers versus pulled-forward repeat buyers. If contribution dollars rose and you acquired customers worth more than the margin you gave up, it worked. If you simply discounted your own base, it did not.