If your Brand Store looked like it fell off a cliff in January and nobody on your team had touched it, you were probably reading a measurement change rather than a performance change. We had several sellers arrive convinced they had broken their own store, and in each case the store was fine.
This is the sort of thing that costs people a fortnight of unnecessary panic, and occasionally costs them a store redesign they did not need. So it is worth understanding exactly what changed, where the old numbers went, and which comparisons are now meaningless.
What Amazon changed
On 1 January 2026, Amazon Ads moved Store ad reporting to what it calls a shopping-signal enhanced last-touch attribution model, and applied "a shorter attribution window" to reflect how shoppers behave in a store. That is Amazon primary, published on its own advertising site on the effective date.
Before that, ad views were credited within a fourteen day window. The change affects Sponsored Brands and Sponsored Display placements billed on viewable impressions, and Amazon DSP store placements. Purchases, Sales and ROAS calculated the new way became the standard reporting numbers, across more than twenty five countries.
Where your old numbers went
They still exist, which is the part most people miss and the part that ends the panic fastest.
Amazon kept the fourteen day view based figures available for eligible campaigns, reported separately as "all views" metrics through the unified reporting interfaces and the APIs. So the comparison you actually want, old basis against old basis, is still possible. You are not being asked to accept a drop on faith.
Worth pulling those before you conclude anything, because they answer the only question that matters in the first week: did the store change, or did the ruler change.
Why this reads as a performance drop
A shorter window credits fewer purchases to the same advertising. Nothing about shopper behaviour changed on New Year's Day, but a sale that happens nine days after an ad view now falls outside a window it used to fall inside, so it stops appearing in that column.
The effect is largest exactly where you would expect: considered purchases, higher price points, categories where people research for a week before buying. If your product is bought slowly and deliberately, more of your credited sales were sitting in the tail of that fourteen day window, and more of them left when the window shortened.
So two brands can see very different sized drops from the same change, and the size of your drop says more about your buying cycle than about your store.
What this does not mean
It does not mean you sold less. Total sales are measured separately from advertising attribution, so the first thing to check is whether your actual ordered product sales moved at all over the same period. In most of the accounts we looked at, they had not.
It also does not mean the advertising stopped working. The purchases that fell outside the window still happened. They are simply no longer credited to the ad view, which is a change in bookkeeping rather than in outcome.
The one thing it genuinely does change is any decision you make from the reported number, which is why the rest of this matters.
The comparisons that are now wrong
Any chart that crosses 1 January 2026 on the new basis is comparing two different measurement systems, and it will show a decline whether or not one occurred. That includes the year on year comparisons that most reporting defaults to, which is an unfortunate collision, because year on year is exactly the view brands lean on in Q1.
Three habits worth adopting for the next year:
- When comparing across the boundary, use the all views metrics on both sides, so the basis is consistent.
- When judging the store itself, use total ordered product sales and conversion, which the attribution change does not touch.
- When setting targets for the year, rebase them on the new numbers rather than on last year's, or you will spend the year missing a target that was measured differently.
Section level insights, and what they are actually good for
Separately, Amazon introduced section level engagement reporting for Stores in January 2026, giving renders, viewable impressions, clicks and click through rate for each section of the store rather than for the store as a whole.
This is the more useful development of the two, and it went largely unnoticed because it arrived alongside the attribution news. Store optimisation has historically been guesswork dressed up as design opinion, because you could see that a store page performed without seeing which part of it did the work. Section level data replaces the argument with evidence: which section people actually engage with, and which beautifully designed block everybody scrolls past.
The practical use is ordering. If the section that earns the clicks is sitting below three that do not, that is a layout decision you can now make on data instead of taste.
What we would do first if this were our account
- Check total ordered product sales across the period. If they held, the store did not break and the rest of this is a reporting exercise.
- Pull the all views metrics for the same campaigns and compare like against like, so you know the true size of any real change.
- Rebase your reporting and your targets on the new basis, once, deliberately, rather than re-explaining the drop every month.
- Open the section level insights and read the store from the top down. The first weak section above a strong one is the cheapest fix available.
- Leave the design alone until the data says which section is the problem. A redesign triggered by an attribution change is a very expensive misunderstanding.
If the store has not been looked at since it was built, the section data usually makes that obvious quickly. Rebuilding around what it shows is what our Brand Store design and build service does, and the audit that precedes it is free.
Related guides
Common questions about the Store attribution change
Did my Brand Store actually stop working in January 2026
Almost certainly not, if the only thing you saw was a drop in attributed sales. Amazon changed how it credits purchases to ad views on 1 January 2026 and shortened the window, so the same performance produces a smaller reported number. The way to confirm it on your own account is to compare total ordered product sales across the period, since that figure is not affected by the attribution change.
Where can I still see the old fourteen day numbers
Amazon kept them available for eligible campaigns as separate all views metrics in the unified reporting interfaces and the APIs. They are not the headline figures any more, so you have to go and get them, but they are there and they are what you want for any comparison that crosses the start of 2026.
Why did my drop look bigger than another brand's
Because the size of the effect depends on how long your customers take to decide. A shorter window removes credit for purchases that happen well after the ad view, so categories where people research for days lose more than impulse categories do. A larger drop is usually a signal about your buying cycle rather than about your store.
Should I redesign the store because of this
Not on the strength of the attribution change alone, and this is the expensive mistake we would most like people to avoid. Establish first whether real sales moved. If they did not, a redesign is solving a problem that does not exist. The section level engagement data Amazon released in the same month is a much better basis for deciding whether the store needs work.
What are section level insights and do I have them
They are engagement metrics reported for each individual section of your store rather than for the store overall, covering renders, viewable impressions, clicks and click through rate. Amazon introduced them in January 2026. They are useful because they show which part of the page is doing the work, which is something store reporting could never tell you before.
Does this change affect Sponsored Products too
The change Amazon published covers Store ad placements, specifically Sponsored Brands and Sponsored Display placements billed on viewable impressions and DSP store placements. Sponsored Products is a click based product and works differently. If your Sponsored Products numbers also moved in the same period, that is worth investigating on its own terms rather than attributing to this.