One lost box, zero impressions

In July the story was that Amazon removed the performance gate on the Featured Offer. Nine weeks on, the story is that the change reaches your account on a day nobody tells you, and something expensive is sitting directly downstream of it.

Amazon described the change plainly. Until July, winning the Featured Offer was two stages: an account-level performance check decided whether you could compete, then offers from everyone who passed were ranked. Amazon is removing the first stage, saying that step was no longer delivering additional value to customers.

Amazon is also explicit that it is not changing how the Featured Offer is selected. Price, delivery speed and performance still decide the winner. Performance did not stop mattering — it stopped being a door and became a term in the ranking.

Analysts describe it as gate-then-rank becoming rank-only, with account-health signals now weighted inside a single score. The weights are not published.

Here is what nine weeks have clarified. The rollout runs in waves through the end of 2026 — US from early July, EU and UK from July 20 — with no per-account notification. No banner, no email, no date. Sellers who previously could not compete on a given ASIN become able to, at some point, on Amazon's schedule.

Which means the number of offers contesting one box on your listing can rise on a Tuesday you find out about from a report.

And if you own a brand, notice who that population is. The sellers who could not previously compete were, by definition, sellers failing performance criteria. For most brand owners, that overlaps considerably with the sellers already sitting on your listings uninvited.

Now the part that makes this a budget question rather than a share question. Sponsored Products generally requires your offer to hold the Featured Offer in order to serve. That is not new and it is not part of the July change — it has been true for years, which is exactly why it goes unexamined. But it means a widened candidate pool now sits directly upstream of ad delivery.

Lose the box on an ASIN and it does not just cost you the organic conversion. The advertising you built around it stops appearing. Spend collapses, which looks like efficiency in a weekly report, and impressions collapse with it, which does not.

And the honest limit on all of this. Nobody can attribute a Buy Box movement to the gate removal. Competitors reprice, stock runs out, delivery promises change. Amazon says selection did not change, and it may be right. The defensible statement is narrower: a variable you did not control just became more volatile, on a schedule you cannot see, upstream of a line item you budget monthly.

THE BRIEF

Seattle, next week. Amazon Accelerate runs September 22–24, with speakers covering Selling Partner Experience, Fulfilment, Trust and Store Integrity, and Support. Read the agenda as a forecast whether or not you go.

Amazon is now the largest US retailer by gross merchandise value, passing Walmart in June — the first change at the top since the 1980s. Roughly 60% of that GMV is third-party inventory.

Assistants are moving conversion. Kohl's reported marketplace GMV up while net sales fell, crediting shoppers using its AI assistant. Attribute quality is becoming a distribution question on marketplaces you have never logged into.

Owner: whoever owns pricing or ads.

Two reports, and one honest limitation to set up front.

Because the rollout has no per-account date, nobody can run a before-and-after on it. What you can do is establish the baseline you should already have had and read it forward — worth more anyway, since the rollout runs to year-end.

Start with the Sponsored Products advertised product report, last 30 days. Sort by spend, take the top 15 ASINs. Ad spend is the exposure here, not revenue.

Then open Business Reports, Detail Page Sales and Traffic by Child Item, which carries Featured Offer percentage per child ASIN. Look up only those 15 — a lookup, not a merge.

Pull two four-week windows, and do not use June. Prime Day sits in it, and a Prime Day window compared against a normal one tells you about Prime Day. Use four weeks in May as your clean reference, and the last four completed weeks as your current read. Keep both windows short; the report is happier that way.

Now the fork. If you are the only offer on an ASIN, its Featured Offer percentage sits near 100 and does not move, and none of this applies to it. Set those aside — two minutes, and a real answer. What is left, anything with competing offers, is where this change lands.

Then drop what you cannot read: anything out of stock in either window, anything you repriced, anything under a few hundred sessions.

What survives is your watch list. Rank by ad spend, keep the top five somewhere you look weekly. You are not looking for a verdict today. You are looking for movement over the next eight weeks, on the five ASINs where movement costs you the most.

Then do nothing else this week, and this is the actual instruction. Two reflexes to resist. Do not cut price to win a box back — in September that also lowers the price history capping your Black Friday deal depth, and you would trade a November discount for a share point you have not diagnosed. And do not raise bids on those ASINs: if you are not holding the Featured Offer, Sponsored Products is not serving, and a higher bid on an ad that is not running buys nothing.

TRANSLATION

Amazon removed a door and told everyone it changed nothing about the room. Both things are true. What changed is that the number of people in the room now moves without warning, and your ad delivery is standing on the same floor. Measure it before you pay to fix it.

See you Friday.

Have any questions? Grab a 15-minute slot here: book a time here.

Subscribe to The Elite Edge here.

— Dan