Amazon Account Health Every Metric That Can Suspend You in 2026

Most sellers who receive a suspension notice did not have a clean Account Health dashboard the week before. The signals were visible at sellercentral.amazon.com. An Order Defect Rate trending toward 0.8%, a cluster of product authenticity complaints accumulating across two or three ASINs, a Voice of the Customer score deteriorating. The suspension was not a surprise to the data. It was a surprise to a team not reading the data consistently.

Amazon Account Health is a running risk calculation that Amazon's systems review continuously. Every metric inside it has a threshold, and several have 2026-specific updates that changed the margin for error. This is every metric that can suspend an account and exactly where each threshold sits.

The Account Health Rating (AHR) is a score between 0 and 1,000 displayed at the top of the Account Health dashboard. It aggregates policy compliance signals from across the account and updates in near real-time based on new violations, resolved issues, and order volume.

Three operational zones define what the score means. A score of 200 or above indicates a healthy account. A score between 100 and 199 places the account in at-risk status, triggering performance notifications and making the account eligible for deactivation review. A score below 100 is critical: deactivation can occur without further notice.

The 250 threshold carries specific significance in 2026. Accounts that maintain a score of 250 or higher for at least six consecutive months become eligible for Account Health Assurance, a program under which Amazon contacts the seller before deactivating rather than acting without warning. That threshold is the practical target for any brand serious about protecting its Amazon channel.

One critical distinction: AHR reflects policy compliance only. It does not capture performance metrics. An account can carry a perfect AHR of 1,000 and still be suspended for an Order Defect Rate breach. Both systems operate in parallel and independently trigger deactivation.

Amazon publishes threshold values for five performance metrics. Breaching any single one can trigger deactivation regardless of AHR or policy compliance history.

Order Defect Rate causes more suspensions than any other metric because it has three independent contributors that compound separately: negative buyer feedback at 1 or 2 stars, A-to-Z Guarantee claims regardless of outcome, and credit card chargebacks. At low order volumes, the math is unforgiving: eleven defective orders against one thousand total puts ODR at 1.1%. The same eleven orders against three hundred puts it at 3.7%. A single category issue, a fulfillment error at the carrier level, or a cluster of returns concentrated in one week can push ODR through threshold before the next report surfaces the trend.

Late Shipment Rate is measured against both a 10-day and a 30-day window, and both windows matter independently. An account can pass the 30-day average while the 10-day window has already breached 4% during a volume spike or fulfillment disruption. The 10-day breach is the one that triggers action.

On-Time Delivery Rate, formalized at a 90% threshold in late 2024 and now a standard 2026 tracking metric, measures whether orders arrive by the promised delivery date rather than just whether they shipped on time. Late Shipment Rate and OTDR track different moments in the same order and can each breach independently.

Internal Alert Thresholds for Proactive Management
The practical internal target for all performance metrics is 50% of Amazon's published threshold. At 0.5% ODR, 2% Late Shipment Rate, and 1.25% Cancellation Rate, an account has a meaningful correction window if a fulfillment problem emerges. At 0.9% ODR, a single bad week eliminates that window entirely.

Policy compliance violations reduce the AHR score by a point value specific to each violation type. Amazon updated enforcement weightings in early 2026, and the practical consequence is that certain violation types now carry far greater impact than before.

A single unresolved restricted product warning now carries roughly the same AHR score impact as three Late Shipment Rate violations under the updated weighting. An unresolved critical violation, including review manipulation findings or suspected counterfeit determinations, can trigger a deactivation review within 72 hours of the flag being opened.

Violations are not permanent. Amazon applies a decay schedule over 180 days and points from older violations gradually reduce their impact on the AHR score. The system does not favor passive waiting, however. An account under a sustained complaint pattern cannot recover through inaction: new violations add negative points while older violations are still decaying, keeping the score suppressed. The only path to meaningful recovery is resolving open violations while preventing new ones.

Three violation categories warrant active management rather than reactive response:

IP complaints filed through Brand Registry reduce AHR significantly and typically cannot be resolved without supply chain authorization documentation or trademark clearance evidence. Repeat IP complaints from the same rights holder signal an upstream distribution or authorization problem that amazon brand protection work addresses at the source rather than through individual complaint responses.

Product authenticity complaints generate both AHR reduction and direct ASIN removal. Amazon's system can act on these proactively before the seller has responded, which makes early detection through the Voice of the Customer dashboard the primary prevention tool.

Listing policy violations accumulate faster across multi-ASIN catalogs when compliance is audited at the account level rather than the ASIN level. A prohibited claim or restricted product flag on a low-velocity ASIN carries the same AHR impact as the same violation on a flagship product.

Voice of the Customer replaced the Customer Reviews dashboard in October 2025 and now serves as the earliest available signal of incoming Account Health risk. The dashboard tracks buyer-reported product experience issues that have not yet been formally filed as complaints: Product Authenticity flags, Product Condition reports, Listing Accuracy complaints, and Product Safety concerns.

Each ASIN receives a CX Health rating from Excellent to Very Poor based on its Negative Customer Experience (NCX) rate. A "Poor" or "Very Poor" ASIN is generating buyer dissatisfaction that has not yet converted into formal feedback, A-to-Z claims, or policy complaints. It typically does convert, usually within 2 to 4 weeks. An ASIN showing sustained Product Authenticity flags in VoC is the earliest visible signal that counterfeit or unauthorized product may be reaching buyers through that listing.

Reviewing VoC weekly against every ASIN and treating "Poor" ratings as an early Account Health alert is the difference between catching a problem at the source and filing a Plan of Action after the account has already been affected.

The monitoring discipline required to keep Account Health genuinely clean across a multi-ASIN catalog is not technically complex. It requires a consistent weekly cadence, defined escalation thresholds, and dedicated operational ownership of the dashboard.

Most in-house teams at growing brands review Account Health reactively, in response to a performance notification from Amazon. By the time a notification arrives, the metric has either already breached threshold or is within days of doing so. The correction window has narrowed.

A structured amazon account management agency maintains a weekly review across every metric layer simultaneously: ODR against both the 60-day rolling window and its week-over-week trend, Late Shipment Rate against both the 10-day and 30-day windows, AHR trend including point impact of any new violations, and VoC NCX rates by ASIN. Escalation is triggered at 70% of threshold, not at breach.

The operational advantage beyond cadence is pattern recognition. An agency managing multiple brand accounts can identify a spike in A-to-Z claims correlating with a specific FBA batch, a cluster of product condition complaints tracing to a packaging change, or a timing pattern between promotional events and negative feedback surges. Those patterns are visible across dozens of accounts. They are invisible to an in-house team running them across one.