How to Read Your Amazon Account Health Dashboard

Most sellers meet the Account Health page the hard way. An email arrives about a listing removal or a warning, and they open the dashboard for the first time in months. By then the clock is already running. Account Health works better as a gauge than as an alarm. It shows how Amazon is measuring your account against its selling requirements, which numbers are drifting, and which issues are waiting for a reply. This guide explains how to read it, how to decide what to fix first, and which habits catch trouble before it becomes a notice.

What the dashboard measures

In Seller Central, Account Health sits under the Performance menu. It collects the measures Amazon uses to judge whether you meet its selling requirements and groups them into a few families. The layout differs by marketplace and by whether you ship orders yourself, and Amazon renames and rearranges sections from time to time, so treat the groups below as a map, not a copy of your screen.

• Customer service performance: order defect rate, which combines negative feedback, A-to-z Guarantee claims and chargebacks, plus cancellation rate, late shipment rate and tracking or delivery measures for orders you fulfill yourself.
• Policy compliance: complaints and violations tied to listings, such as suspected intellectual property issues, product authenticity, product condition, product safety, restricted products and review policy breaches.
• Account Health Rating: a summary score and status that rolls up the policy items, with more serious violations weighted more heavily than minor ones.
• Open notices: warnings, listing removals and requests for documents, usually with an action button and sometimes a response deadline.

Next to each metric you will see your current value, Amazon's target and a status. Always read the target on your own screen rather than from an article, including this one. Amazon sets the targets and the window each metric covers, and both can change.

Read it in order of consequence

A page with a dozen sections invites you to look at whatever is red or whatever is familiar. A fixed reading order prevents that. Move from what can stop you selling to what merely needs tidying.

• Start with open notices that carry a deadline. A document request or a removal with a response date is the one kind of item that can expire while you are busy with something else.
• Then read the account-wide metrics. Order defect rate, cancellation rate and late shipment rate describe your whole account, so a bad reading threatens everything you sell.
• Then read the policy items, which usually attach to a single listing. Open each one and note the ASIN, the reason Amazon gives and the action it asks for.
• Finish with trends. A metric that is still inside its target but has moved the wrong way for three weeks deserves a note today and a fix next week.

Two questions sort almost every item you find: how far does it reach, and is there a deadline? The grid below turns that into a default response.

One trap deserves a warning: rates are ratios, so low order volume makes them jumpy. If the window holds 100 orders, a single defective one moves a rate by a full percentage point. In a window of 1,000 orders, the same event moves it by a tenth of a point. That does not make the problem less real, because Amazon still scores the rate, but one bad order deserves a calm look at the cause, not panic.

Root cause, fix, prevention

Every item you open should leave a three-line record. The first line is the root cause: why it happened, described in terms of your own process. 'Amazon flagged the listing' is a symptom. 'The copy went live without a claims check' is a cause. The second line is the fix for this instance: what you changed and what you sent. The third line is prevention: the change to your process that makes a repeat less likely.

Here is a hypothetical. A listing is removed because its bullets make a claim the category does not allow. The fix is to edit the claim and follow Amazon's steps to reinstate the listing. Prevention is a short list of claims your categories restrict, checked before any copy is published, plus a rule that copy from an agency or freelancer passes the same check. Without the third line, you will be back at this page when the next batch of listings goes live.

The same pattern fits a rising late shipment rate. A root cause might be a handling time that promises more than your warehouse can deliver on Mondays. The fix is a longer handling time or an extra packing shift. Prevention is a weekly check that your handling time matches the real dispatch times in your order data.

When Amazon offers an appeal or asks for a plan of action, the same three lines are the skeleton of what you submit. Reviewers want evidence that you understand the problem and have changed something. Write from your record, stay factual and attach proof. No wording guarantees a result.

Early warnings that arrive before the red status

• Trend the numbers yourself. Once a week, copy each metric, its target and its status into a simple log. The dashboard shows the present; your log shows direction.
• Watch the inputs. Returns, negative feedback, A-to-z claims and buyer messages that use words like fake, used or damaged often appear before an authenticity or condition complaint becomes a policy item.
• Watch the supply side. Stock-outs on merchant-fulfilled items, stranded inventory and listings that have gone inactive tend to come first, with late shipments and cancellations close behind.
• Route notifications to a shared inbox. If warnings reach only one person's email, a holiday can turn a short response window into a missed one. Check notification preferences and user permissions so more than one person can see Account Health, and give each person a separate login instead of sharing one.

Mistakes that make small issues bigger

• Sending a generic reply to a specific notice. The reviewer compares your answer with what was flagged, so answer that exact issue.
• Fixing the listing but not the process, so the same violation returns on the next ASIN.
• Letting low-urgency policy items pile up. They feed the summary rating even when none of them is urgent alone.
• Judging one day's reading. Compare the same metric week over week before you decide anything has changed.
• Sharing one login across a team, which hides who changed what and leaves no one accountable when something goes wrong.

Where to start

• Open Account Health today and write down every metric, its target and its status as your baseline.
• Read the open notices and policy items, and sort them with the triage grid.
• Give each open item a root cause, a fix and a prevention step, even a one-line version.
• Create the weekly log and put a fixed time on the calendar.
• Check that notifications reach a shared inbox and that at least two people have appropriate access.
• Book a monthly review of repeat issues so the prevention list keeps growing.