Amazon India Account Health: The Five Metrics That Actually Get You Suspended

Most sellers we meet on Amazon India are watching the wrong numbers. They obsess over star ratings and the latest one-star review, refreshing the product page like it owes them money. Meanwhile the metrics that actually decide whether their account survives the quarter sit ignored in the Account Health dashboard. Amazon does not police sentiment. It polices behaviour. And the gap between those two things is where most suspensions are born.

This is the uncomfortable part. A brand can have a 4.6 average rating and still get deactivated overnight, because a buyer-facing rating and a seller-facing health metric are not the same instrument. One measures how customers feel. The other measures whether you are a liability to the marketplace. Amazon only cares deeply about the second one. Once you internalise that, account management stops being a guessing game and becomes a discipline.

Order Defect Rate is the metric that ends accounts

If you only watch one number, watch Order Defect Rate. ODR bundles three failures into a single percentage: negative feedback, A-to-z guarantee claims, and chargebacks. Amazon wants this comfortably under its threshold, and the threshold is low enough that a handful of bad orders in a slow week can spike it. That is the trap. ODR is a rolling percentage, so a seller doing modest volume is far more exposed than a high-volume one. Twenty defects on two thousand orders is noise. Twenty defects on two hundred orders is a deactivation email.

The defensible move is to treat every A-to-z claim as a fire, not a ticket. A-to-z claims hurt twice. They count against ODR directly, and an unaddressed claim signals to Amazon that you are not engaging. Respond inside the window, every time, even when the buyer is wrong. We have watched brands argue themselves out of reinstatement by being right and slow instead of pragmatic and fast.

Amazon does not suspend you for being disliked. It suspends you for being unreliable. ODR is simply the number that measures your unreliability.

Late Dispatch Rate punishes the back office, not the brand

Late Dispatch Rate is where good brands with bad operations get exposed. It measures the share of orders you confirm shipment on after the expected dispatch date. Easy mode is Fulfilled by Amazon, where the warehouse handles the clock for you. The exposure lives in self-ship and Easy Ship, where your own pick-pack discipline is on trial every day.

The pattern we see repeatedly is a brand that fulfils beautifully on weekdays and quietly bleeds late dispatches over weekends and festival spikes. The metric does not care that your packer took Sunday off. It just climbs. Diwali and the big sale events are when LDR quietly accumulates into a warning, because volume triples and the same two people are still packing. The 2025 Great Indian Festival made that risk concrete: Amazon reported a record 276 crore customer visits with the highest-ever number of sellers crossing the ten lakh mark, and 70 percent of traffic came from beyond the top metros, per About Amazon India. A demand wave that size is also a dispatch test. If you are scaling on Amazon India, your dispatch capacity has to scale before your ad spend does, not after. We cover this exact sequencing in our first 90 days playbook, because the brands that fail early almost always fail on fulfilment before they fail on demand.

Policy violations are the silent killers

ODR and LDR are visible. You can watch them trend. Policy violations are different. They arrive as a notification and a hit to your Account Health Rating, and they are the category most likely to deactivate you with no warning at all. The common ones on Amazon India:

• Intellectual property complaints, where a brand owner or another seller files against your listing for trademark or copyright infringement.
• Authenticity and inauthentic complaints, which Amazon treats with extreme prejudice because counterfeit is an existential risk to its marketplace.
• Restricted product violations, often from sellers who do not realise a category needs approval or a compliance document.
• Listing policy breaches, like prohibited claims, manipulated images, or detail-page hijacking.
• Used-sold-as-new complaints, where condition disputes escalate into authenticity flags.

The reason these are so dangerous is that a single authenticity complaint can outweigh a year of clean operations. Amazon’s logic is asymmetric on purpose. The cost of a false suspension to one seller is far smaller, in Amazon’s calculus, than the cost of one counterfeit reaching a buyer. So it errs hard toward deactivation and makes you prove your way back. Keeping clean invoices from authorised distributors for every SKU is not paperwork hygiene. It is your defence file. When a complaint lands, the brands that get reinstated fast are the ones who can produce sourcing documents the same day, which is also the spine of any credible suspension appeal and plan of action.

Valid Tracking Rate and the metrics behind the metrics

Below the headline numbers sit a second tier that most sellers never open. Valid Tracking Rate measures whether your shipments carry trackable, scannable information. On-Time Delivery Rate measures whether parcels actually arrive in the promised window. Pre-fulfilment Cancel Rate measures how often you cancel an order before shipping it, usually because you oversold stock you did not have.

These rarely suspend an account on their own. They do something quieter and arguably worse. They throttle you. A poor Valid Tracking Rate erodes Amazon’s trust in your fulfilment and can cost you the Buy Box, faster shipping badges, and category eligibility. Pre-fulfilment cancellations are a direct symptom of broken inventory sync, and they tell Amazon you are listing stock you cannot deliver. The brands that treat these as early-warning lights fix the operational rot before it metastasises into an ODR or policy problem. The brands that ignore them wonder why their visibility quietly evaporated.

How the five metrics actually interact

Here is the part the dashboards do not spell out. These five are not independent. They are a chain. Oversell your stock, and Pre-fulfilment Cancel Rate rises. The orders you do ship go out late, so Late Dispatch Rate rises. Late and missing parcels generate A-to-z claims and negative feedback, so Order Defect Rate rises. Frustrated buyers escalate, condition disputes turn into authenticity complaints, and now you have a policy violation. One operational failure, inventory accuracy, propagated through all five metrics and arrived as a suspension that looks, on the surface, like it came from nowhere.

This is why we argue that account health is an operations problem wearing a marketing costume. You cannot review your way out of a fulfilment failure, and you cannot advertise your way out of a policy strike. The work is unglamorous: accurate inventory, fast dispatch, clean invoices, same-day responses. That is the actual job inside Marketplace Account Management, and it is why a sharp account manager pays for themselves long before they ever touch your ad budget. We make the case for that math in our piece on how a marketplace account manager earns their fee.

What changed recently

Two shifts in the last year make this discipline more urgent, not less. The first is on cost. In March 2026 Amazon India expanded zero referral fees to over 12.5 crore products priced under one thousand rupees across 1,800-plus categories, a more than tenfold expansion from the previous year, and cut Easy Ship fees by more than 20 percent for products under three hundred rupees, according to About Amazon India. YourStory framed it as the market leaning toward zero-commission models after Flipkart’s own waiver. The operator read is blunt: when the marketplace stops taking a cut on entry-price SKUs, more sellers and more SKUs flood in, and the only lever Amazon has left to manage that crowd is enforcement. Cheaper fees plus heavier policing is the trade.

The second shift is the policing itself. Account Health Rating enforcement has moved from reactive to proactive, with more listings flagged and accounts placed at risk even during periods of moderate sales, and the rating now surfaces as colour-coded risk levels rather than a single buried number. None of the five metrics above changed in definition. What changed is the tolerance. The window between a metric drifting and Amazon acting on it has narrowed, which means cadence is no longer a nice-to-have.

Watch the metrics on a cadence, not in a panic

The single behavioural change that separates stable sellers from suspended ones is cadence. Sellers who only open Account Health when something breaks are perpetually reacting. Sellers who review the five metrics on a fixed weekly rhythm catch the trend while it is still a trend and not yet a threshold breach. A spiking ODR seen on Monday is a problem you can solve. The same spike discovered in a deactivation notice is a problem you can only appeal.

Build the rhythm into your week. Pull ODR, LDR, the policy-violation log, Valid Tracking Rate and Pre-fulfilment Cancel Rate, and read them as one story rather than five disconnected gauges. If a number moves, ask which operational root caused it, because it almost always traces back to inventory or dispatch. We have written a full structure for this in the monthly account health audit every serious seller should run, and pairing it with our Marketplace Account Management work and broader Marketplace Growth support is how brands stay live through the events that suspend everyone else.

None of this is exotic. It is just attention pointed at the right numbers. Stop refreshing your reviews. Open the dashboard Amazon actually reads, and manage the five metrics that decide whether you trade tomorrow.