Below Standard on eBay, Order Defect Rate on Amazon

There is a specific kind of bad morning in this business. Sales are fine, reviews are fine, nothing is broken — and the marketplace has moved your account down a level. Search placement drops, fees go up, and the reason is a percentage you had never looked at.

The frustrating part is that these systems are not secret. Both eBay and Amazon publish their thresholds, and both measure roughly the same thing: whether the shipping paperwork was done on time.

That is the whole point of this article. The metrics that decide your account are administrative, they are measured over a long window, and for a small seller they are counted in orders, not in percent. Which means they are entirely preventable, and almost nobody checks them until the email arrives.

The left column is judged on one day a month; the right one moves as things happen.

eBay seller performance standards: Top Rated, Above Standard, Below Standard

eBay scores every seller on the 20th of each month and puts them in one of three levels: Top Rated, Above Standard, Below Standard.

To stay Above Standard — which is simply "not in trouble" — you need a transaction defect rate of no more than 2% of transactions, and no more than two cases closed without seller resolution (or 0.3% of transactions).

Top Rated is a different bar: transaction defect rate at or under 0.5%, cases at or under 0.3%, late shipment rate at or under 3%, and tracking uploaded on time and validated on at least 95% of orders. That status is worth real money — Top Rated Plus listings get a discount on final value fees — so losing it is a price increase you did not agree to.

Note which number lives where, because this is the detail sellers get backwards. A late shipment rate does not, by itself, put you Below Standard — it costs you Top Rated. What puts you Below Standard is the transaction defect rate and unresolved cases, and that is the level where the consequences get serious: items ranked lower in search, selling limits reduced, advertising tools blocked, and from the first of the following month, order funds held until tracking shows the item shipped.

Then the detail that catches small sellers. The evaluation looks back three months if you had 400 or more transactions in that period, and twelve months if you had fewer. Under a hundred orders a year, a single bad week does not average out — it is the average, for a year. And at that volume, 0.3% is three orders.

That asymmetry is worth sitting with. A high-volume seller who has a terrible fortnight can trade out of it: the next eight weeks of clean orders dilute the damage before the next evaluation. A seller shipping thirty orders a month cannot. The same two defects that a large account absorbs without noticing will still be sitting in a small account's numbers next August.

Amazon's order defect rate and the rolling 60-day window

Amazon publishes targets rather than levels, and they are strict: order defect rate under 1%, late shipment rate under 4%, pre-fulfillment cancellation rate under 2.5%, and valid tracking rate above 95%.

Order defect rate is the one to watch, because it is measured over a rolling 60-day window and it combines things you think of as unrelated — negative feedback, A-to-z guarantee claims and chargebacks all land in the same number. Under 1% of a small order count is, again, a handful of orders.

"Rolling" is the word doing the work there. eBay's window closes on a date; Amazon's window moves with you, so a defect stays in your order defect rate for sixty days from the day it happened and then falls out on its own. That is more forgiving than it sounds — nothing is held against you for a year — and less forgiving than it sounds, because there is no evaluation date to prepare for. The number is live.

On top of the individual metrics sits the Account Health Rating, a score that moves as things happen rather than on an evaluation date: green means healthy, and a low enough score makes an account eligible for deactivation. It reacts faster than eBay's monthly cycle, which is good news and bad news. The good news is you can see a problem the day it starts. The bad news is that nobody opens that page on a good day.

What late shipment rate and valid tracking rate actually measure

Read the two lists together and a pattern shows up: almost every metric is about the shipping record, not the product.

Valid tracking rate is the clearest case. A tracking number that was uploaded late, or that never scanned with the carrier, counts as no tracking at all — even though the buyer got the parcel and was perfectly happy. You can have flawless service and fail the metric on paperwork. This is the one that punishes good intentions hardest: sellers who buy a label, mark the order shipped, and then hand the parcel over two days later have a number that was valid at the moment they entered it and invalid by the time it was measured.

Late shipment rate is the same trick in a different field. It compares the day you marked it shipped against the handling time you promised on the listing. Sellers lose that number not by shipping slowly but by promising one-day handling on items they touch twice a week. Nothing about the buyer's experience changed; the promise did.

Transaction defects and cases are the expensive category, and the difference between them matters. A buyer opening a case is not yet a defect. A case that you resolve is not a defect either. It becomes the most costly single event in eBay's table only when the marketplace has to close it because you did not act — which is a deadline problem, not a service problem.

Which points at the fix, and it is unglamorous:

• Promise the handling time you actually keep, not the one that ranks best. Two days you always hit beat one day you usually hit.
• Upload tracking that scans. A number entered from a label that was never handed over is worse than no number, because it fails silently.
• Answer cases before the marketplace closes them. A case closed without seller resolution is the most expensive single event in eBay's table.
• Know your denominator. If you ship a few dozen orders a month, work in orders, not in percent: find out how many defects put you over, and treat that number as your budget for the year.

That last one is the habit that changes behaviour. "Keep the defect rate under 2%" is advice nobody can act on in the moment. "You can afford three more defects before August" is a number you remember while deciding whether to chase a supplier today or tomorrow.

The part nobody does daily

None of this is hard to understand. It is hard to remember — it is a page in a seller dashboard that gives you good news on all but about four days a year, so it stops being a page anyone opens.

That is the kind of watch worth handing to something that never gets bored. SellerClaw is a team of AI agents connected to the stores you already sell on, and one of the plainest things it does is read those numbers every morning across every marketplace you use, then say something only when a number is moving the wrong way — and tell you which unshipped orders are causing it, rather than just reporting the percentage.

The first check costs nothing beyond connecting a store: ask it how your account health looks and how many defects you are from the next threshold down. If the answer is "comfortable", you have lost five minutes. If it is a number smaller than you expected, you just found the cheapest problem you will fix this quarter — and the best time to fix it is before the season's volume arrives.