Prime Day 2026, day by day: a slow start with a stronger finish.

We pulled the day-level numbers from the latest Prime Day and compared them to last year’s event. As a whole, Prime Day was down 8.6% YoY, but the real story hides under that number: a weak start yet with a strong closing, with advertising holding its efficiency even as budgets shrank. Here is what the four days actually looked like, and why we think it's worth looking beyond the numbers.

The four-day arc: an upward trend from a low start

Total portfolio revenue landed at -8.6% YoY for the full event, but that blended figure hides how much the picture changed day to day: the decline shrank every day it ran. Day one was the weakest point by far, while day four ended above last year’s equivalent day.

We read this as an event that corrected itself as it went, not one that was weak throughout. Whatever weakened demand on day one, be it price positioning, timing, or something else, it eased steadily and reversed by the close. Anyone judging this Prime Day only by day-one numbers alone would have called it a much bigger miss than it turned out to be.

What was moving underneath the revenue line

The revenue recovery lines up with a clear shift in shopper behaviour. Fewer people browsed: glance views were down to -14.3% YoY across the four days. Lower traffic impacted the overall performance of the deal event. Our hypotheses? The latest heat wave hitting Europe and the ongoing FIFA World Cup may have had an effect on shoppers’ interest.

But of the people who did land on a page, far more turned into sales. Conversion rate was up more than 20% YoY, strongest on the opening day, easing a little through midweek, then ticking back up on day four.

That conversion gain came with a cost. Average selling price was down almost as much as conversion was up. Deeper discounting may have contributed, but the product mix likely played a role as well The net effect: fewer shoppers, converting far more often, at lower prices, still added up to more units moving than last year.

Advertising: efficiency held even as budgets tightened

On the paid side, teams pulled back hard. Ad-attributed revenue and ad spend both fell by roughly the same percentage YoY, so return on ad spend (RoAS) barely moved, holding within a couple of percentage points of last year’s level. What does this mean? Less money in, proportionally less revenue out, efficiency essentially unchanged.

What did move was quality: click-through rate rose by double digits YoY, and on- ad conversion improved too, even though the number of ad impressions bought was down by roughly a third and clicks fell by roughly a quarter. Fewer, better-targeted impressions did more work per euro than last year’s high-volume approach.

The advertising week had its own daily curve

Day one of the event was the strongest day for advertising by a clear margin: it captured roughly 30% of the whole event’s ad revenue, at the best RoAS of the four days. Day two and day three cooled off, with day three being the lowest point for both ad revenue share and RoAS. Day four partially recovered, though it did not fully match day one.

Click-through rate told a slightly different story: it actually peaked mid-week, on days two and three, even as RoAS was falling. Clicks kept coming at a good rate, they just converted a little less efficiently as the week wore on and the discounting deepened.