Amazon Perfect Order Percentage (POP): What It Measured and What Amazon Enforces Now

Published: February 27, 2017
Last updated: August 10, 2026

Perfect Order Percentage (POP) - also called perfect order rate - is a legacy Amazon seller metric measuring the share of orders completed with no negative incident: no cancellation, no late shipment, no refund, no negative feedback, no A-to-Z claim, no chargeback, no buyer-initiated message. One blemish and the order was no longer perfect.

POP is not what Amazon uses for enforcement, and hasn’t been for years. Two of its seven components will cost you money if you try to minimize them.

The concept is still worth ten minutes - not as a score to chase, but as a diagnostic that catches a failure mode today’s metrics are structurally blind to.

Table of Contents

• What Perfect Order Percentage Measured
• Why You Can’t Find POP in Seller Central
• Where Each POP Component Lives Now
• The Metrics That Govern Your Account Today
• The Threshold Trap: 1% Is Not the Target
• The One Thing POP Still Gets Right
• Two Components You Should Not Optimize
• FAQ
• What to Do Monday Morning

What Perfect Order Percentage Measured

The formula, as legacy sources describe it:

POP = (Perfect orders in the last 90 days ÷ Total orders received in that period) × 100

Seven events disqualified an order:

• Negative feedback
• A-to-Z Guarantee claims
• Chargebacks
• Cancellations
• Late shipments
• Refunds
• Buyer-initiated messages

Exclusions included pending orders the customer cancelled themselves, cancellations made through Amazon’s own order-cancellation option, buyer messages unrelated to an order, and “thank you” notes needing no reply.

Treat all of this as historical. The 90-day window and the widely repeated 95% target can’t be verified against current Amazon documentation, and the exclusion language in the underlying sources refers specifically to the Amazon UK site - a tell that this was regional legacy help content, not global policy.

Why You Can’t Find POP in Seller Central

Nothing in Amazon’s current documentation mentions it.

Your Account Health dashboard (Performance → Account Health) has three sections: Customer Service Performance, Shipping Performance, and Policy Compliance. Customer Service Performance consists of Order Defect Rate and Invoice Defect Rate. POP is not among them. Go looking for the Perfect Orders report itself and current catalogues of Seller Central’s fifty-odd reports don’t list it either, nor the Customer Metrics page it lived on - the exact download older guides tell you to pull.

So why does the metric refuse to die? Search for Perfect Order Percentage and you get a tight cluster of pages all restating the same 90-day/95% framing, none citing Amazon. Earlier versions of this page were part of that cluster.

Amazon never announced a removal, but nothing confirms it still exists either. The defensible position: POP is legacy, and the 95% bar is not an Amazon requirement.

Where Each POP Component Lives Now

POP’s seven components didn’t vanish. Five were redistributed across separately enforced metrics; two stopped counting as defects at all.

Legacy POP component	Where it lives in 2026
Negative feedback	Folded into Order Defect Rate
A-to-Z Guarantee claims	Folded into Order Defect Rate
Chargebacks	Folded into Order Defect Rate
Cancellations	Pre-fulfillment Cancel Rate (< 2.5%)
Late shipments	Late Shipment Rate (< 4%) and On-Time Delivery Rate (≥ 90%)
Refunds	Not a defect metric - tracked in returns reporting only
Buyer-initiated messages	Not a defect metric - messaging has its own response expectations

Five of those failures still cost you, scored on four separate scoreboards with their own thresholds and windows. The other two stopped costing you anything, which is where the old advice turns dangerous.

The Metrics That Govern Your Account Today

Your Account Health Rating (AHR) is a 0-1,000 score, updated close to real time, that summarizes deactivation risk.

AHR band	Score	What it means
Green - Healthy	200-1,000	Not at risk of deactivation on the policies in the score
Yellow - At Risk	100-199	Your account is at risk of deactivation
Red - Unhealthy	0-99	Eligible for deactivation, or already deactivated

And the performance thresholds underneath it:

Metric	Threshold	Window	Applies to
Order Defect Rate (ODR)	Below 1%	60-day rolling	All orders
Pre-fulfillment Cancel Rate	Below 2.5%	–	Seller-fulfilled
Late Shipment Rate (LSR)	Below 4%	–	Seller-fulfilled
Valid Tracking Rate (VTR)	Above 95%	–	Seller-fulfilled
On-Time Delivery Rate (OTDR)	90% or higher	–	Seller-fulfilled, tracked units

Four of the five apply only to seller-fulfilled orders - a routine source of confusion when FBA and FBM sellers compare notes.

Two ODR mechanics catch sellers out repeatedly. The 60-day window is rolling, and a defect attaches to the order date, not the date the complaint arrived - so a bad week keeps costing you for two months after you fixed it. And multiple defects on one order count once. Amazon already applies order-level logic; it just narrows it to ODR’s three components.

One number matters more than the thresholds: AHR 250, sustained for six consecutive months. That, plus a Professional selling plan and a valid emergency contact on file, qualifies you for Account Health Assurance - a specialist contacts you before a deactivation, and your account stays active as long as you respond within 72 hours. It stops short of immunity: listing-level deactivations aren’t covered, nor are the serious violation categories (counterfeit, payment fraud, identity fraud, child safety). Even so, it’s the highest-leverage account health target available, and it rarely shows up in anyone’s operating plan.

In February 2026 Amazon narrowed OTDR enforcement from account-wide to item-level: rather than deactivating all your seller-fulfilled listings, it now deactivates the specific listings dragging the metric down. That shrinks the blast radius of one bad quarter without protecting anyone who underperforms consistently - chronic cases still face deactivation across the catalogue.

The Threshold Trap: 1% Is Not the Target

Stay under 1% ODR and you’re compliant. Compliant and safe are not the same state, and the distance between them is wider than most sellers assume.

Practitioner reporting puts strong sellers under 0.3% ODR, with FBA sellers commonly running 0.1-0.3%. That’s a 3-10x gap between the threshold and where healthy operations sit. A 0.9% ODR is technically fine and a long way outside normal - and because ODR moves in whole orders, a low-volume seller can cross 1% in one bad week.

Run it: at 300 orders in a 60-day window, three defects put you at 1%. Three. Two granted A-to-Z claims and one chargeback, and you’re having a conversation with Amazon.

Manage to 0.3%. Treat 1% as the cliff edge it is.

Now the part I can’t give you, because nobody can: how often this ends in suspension. Amazon publishes no deactivation rate, no median reinstatement time, no appeal success rate - and neither does anyone else, so treat quoted suspension percentages as invented. Prevention wins on asymmetry alone. It costs a weekly glance at four numbers. The alternative is an outage of unknown length during which you cannot list, sell, or win the Buy Box.

The One Thing POP Still Gets Right

Every current metric is scored on its own axis. POP scored the order. That difference creates a blind spot you can fall into.

Consider a seller-fulfilled operation with 1,000 orders in a 60-day window:

• ODR at 0.6% → 6 orders with defects
• Late Shipment Rate at 3% → 30 late orders
• Pre-fulfillment Cancel Rate at 2% → 20 cancelled orders

Every metric is green. Every threshold is met. Nobody at Amazon is calling.

Now count orders instead of rates. If those failures land on largely different orders, that’s roughly 56 - 5.6% of your customers had something go wrong. Your order-level clean rate is about 94.4%, which would have failed the old 95% bar while your entire dashboard reads healthy.

Keep the diagnostic and drop the rest - the score, the 95%, the report you can no longer download. What survives is the habit of dividing problem orders by total orders and looking at the number Amazon doesn’t show you. Pull your orders and returns data, dedupe by order ID, and you have it in twenty minutes.

Where does the idea break down? This is an operating number, with no bearing on your compliance standing. It won’t predict a deactivation and has no threshold attached - a seller in a category with structurally high return rates runs a worse figure permanently and stays perfectly healthy. Use it to track yourself over time and to surface the SKUs or ASINs causing disproportionate trouble. It says nothing useful about how you compare to other sellers.

Four metrics say you’re fine. Your customers disagree. Feedvisor connects fulfillment signals to your pricing and advertising data, so the orders going wrong surface as a pattern instead of as four separate rates that each look acceptable.

Two Components You Should Not Optimize

Refunds and buyer-initiated messages were POP components. Neither is a defect metric today, and suppressing them to protect your numbers will cost you.

Refunds are the expensive mistake. A buyer who can’t get a refund files an A-to-Z claim instead - and A-to-Z claims count toward ODR while refunds count toward nothing. Refusing a $30 refund to protect a number that no longer exists hands you a defect against a number that does. Processing returns and refunds cleanly is defect prevention in its own right.

Buyer messages are the milder version of the same error. A message is a customer asking a question. Answer it well and you prevent the negative feedback that does count.

FAQ

Is Perfect Order Percentage still used by Amazon?

Not for enforcement.

What is a good Perfect Order Percentage on Amazon?

There isn’t a published one, and the 95% figure circulating online can’t be verified against Amazon documentation. If you calculate the rate yourself, your own prior quarter is the only benchmark that means anything.

Does a high refund rate hurt my Amazon account health?

Not directly - refunds aren’t a defect metric and don’t feed your AHR. High return rates still signal listing or product problems, and they cost you in returns processing fees. But refusing refunds to hold the number down converts a harmless statistic into A-to-Z claims that carry real weight.

Where do I find the Perfect Orders report in Seller Central?

You most likely won’t - it belonged to a legacy Customer Metrics page. Use Performance → Account Health instead, and see Customer Metrics FAQs for how current reporting fits together.

Do FBA sellers need to worry about any of this?

Less than FBM sellers, and that’s the trap. Four of the five thresholds apply only to seller-fulfilled orders, so an FBA seller watching a green dashboard is really watching one metric - ODR - and product complaints, claims, and chargebacks all still land there.

If a low POP didn’t cause my warning, what did?

Check the notice - Amazon names the metric it acted on, and it will be ODR, a cancel or late shipment rate, or on-time delivery. Sellers tracking a composite often can’t tell which of the four slipped, because averaging them together is exactly what hides the one that moved.

What to Do Monday Morning

1. Check your AHR against 250, not 200. Sitting between the two means you’re healthy - and ineligible for the one program that catches you before a deactivation.
2. Set your ODR working target at 0.3%, and fix problems the week they happen - the 60-day window means today’s mistake is still on your record two months from now.
3. Calculate your order-level failure rate quarterly. Dedupe problem orders by order ID, divide by total, track the trend.
4. Stop treating refunds and buyer messages as defects. Handle both generously - it’s the cheapest way to prevent the claims and feedback that Amazon does count against you.