Every summer, Prime Day lands like a tidal wave — a 3× single-day surge, roughly a quarter of the month's sales crammed into 48–96 hours. Brands pour in discounts and ad budget to catch it, then measure the splash — units sold — and call it a win.
But units sold can't answer the three big questions that actually matter to any brand:
- Was the gross-margin sacrifice worth it?
- How many genuinely loyal customers did we win?
- Did we earn a single point of lasting brand equity?
The splash isn't the story. Who's in the water, and where they go when it recedes — that's the story. We followed it brand by brand, shopper by shopper, and all three years told the same thing — which is why we'd expect Prime Day 2026 to repeat it.
Most Prime Day buyers are new to you — just like every other day
One thing we were especially keen to look at: of the customers a brand attracts on Prime Day, how many are genuinely new — buying it for the very first time?
We followed the same 15,000 shoppers across all three years, so at each Prime Day purchase we can look back over their full history and ask whether they'd bought the brand before. If not, it's new-to-brand; if so, repeat-brand.
The answer: across all three years, 70–82% of Prime Day purchases went to a brand the shopper had never bought before.
Here's the part that should unsettle you: a normal Tuesday looks exactly the same. Prime Day doesn't make shoppers more adventurous or change who you attract — it just turns up the volume on a marketplace where 7 in 10 purchases, every single day, go to a brand the buyer has no prior relationship with.
"Prime Day doesn't win you a new kind of customer — only more first-time ones. Whether that pays off comes down to how many come back."
Your landlord is your fiercest competitor
While you pay to acquire those strangers, the marketplace isn't a neutral venue — it's competing for the same cart, with its own products placed first.
To size it, we tagged every Amazon-owned brand in the data — Amazon Basics, Amazon Essentials, Echo, Fire, Kindle, Blink, and more — and tracked their share of total spend across the Prime Day window, measured against a matched baseline three weeks earlier so any lift reflects the event, not the season. That share jumps every Prime Day — roughly doubling in 2024 and 2025, and more than tripling in 2023. The size varies year to year, but the pattern holds across all three: structure, not coincidence.
And the doubling is only the visible part. Stack on the ~15% referral fee, fulfillment, and peak-auction ad inflation (industry figures), and the house takes a cut of your sale, sells you the visibility, and competes for the same cart — all at once.
Most customers buy once — and never come back
So Prime Day floods you with first-time buyers. Does the discount turn any of them into repeat customers?
To find out, we took every brand a shopper bought for the first time on Prime Day and followed that shopper forward to see whether they ever bought it again — then ran the identical check on brands first won on an ordinary, full-price day, so the discount's real effect would stand apart from normal repeat behavior. It barely registers: just 18% of brands won on Prime Day are ever bought a second time, and 82% never are. That's all but identical to the 17% repurchase rate at full price.
The markdown bought you volume and almost no extra loyalty. You discounted to win first-time buyers, Amazon took its cut, and eight in ten never came back.
"A discount that doesn't change retention isn't acquisition. It's a giveaway with a receipt."
What the data shows: a volume event, not a loyalty event
Put the three observations together. Prime Day delivers a 3× surge of mostly first-time-brand buyers, the platform's own brands roughly double their share of that spend, and the great majority of those new brand relationships are never repurchased.
That is the event the data describes: a high-volume, low-retention switching market in which the platform captures a growing slice of the spend. It held in 2023, 2024, and 2025 — the same picture three years running — which is the strongest reason to expect it again in 2026. A different read from the one a GMV recap shows, and a useful one heading into Prime Day 2026.
Methodology: how we read it
The findings draw on Ario's consented panel of U.S. Amazon shoppers — millions of SKU-level purchases with dates and prices, spanning 2023 through 2026. For each year we isolated the Prime Day window (2023, 2024, 2025) and compared it against a matched baseline three weeks earlier, so any lift reflects the event itself, not the season.
A purchase counts as new-to-brand when the shopper had not bought that brand earlier in their history. Retention tracks whether a brand first bought on Prime Day was ever purchased again in the months that followed. Amazon's own-brand share covers its private-label and device brands — Amazon Basics, Amazon Essentials, Echo, Fire, Kindle, Blink, and more.