amazon referral fee 2026

There is a quiet con running across Indian marketplace selling, and most brands are paying for it without noticing. They hire an account manager, or retain an agency, and what they actually receive is a person who logs into Seller Central, exports a few reports, reformats them into a deck, and emails it over on Friday. The numbers in that deck went down or up entirely on their own. The account manager did not cause the movement. They narrated it. And narration, however well formatted, is not management.

We are blunt about this because we have inherited too many accounts where the previous manager was a reporting layer wearing a strategy title. The brand was paying for outcomes and receiving documentation. So here is the standard we hold ourselves to, and the one you should hold any partner to. A marketplace account manager earns their fee by moving three things together: account health, buy-box ownership, and advertising efficiency. If they can only show you one of those, and only describe rather than influence it, they are a cost, not an investment.

The reporting trap

Reports are necessary. They are not the work. The confusion between the two is where most account management quietly fails. A report tells you the order defect rate climbed last week. Management is the person who saw it climbing on Tuesday, traced it to a late-dispatch spike from a single warehouse, and fixed the pick-pack rota before the number crossed a threshold. The first is a historian. The second is an operator.

The tell is simple. Ask your account manager what they changed last month, not what they observed. If the answer is a list of insights, charts and trends, you have a reporter. If the answer is a list of actions and the numbers those actions moved, you have a manager. The best operators produce reports almost as a byproduct, because the report is just the receipt for work already done. We have written about what a report should even look like in our piece on a marketplace reporting dashboard that leadership will actually read, and the short version is that a dashboard nobody acts on is decoration.

You do not pay an account manager to tell you what happened. You pay them to be the reason something different happened.

Account health is the floor, not a metric

The first thing a real account manager protects is your right to keep trading. Everything else is moot if Amazon deactivates you on a Thursday. This means account health is not one tile on a dashboard. It is the floor the entire account stands on, and a competent manager watches it on a cadence rather than discovering problems inside a suspension notice.

That work looks unglamorous. It is the daily discipline of catching a rising order defect rate before it breaches, responding to every A-to-z claim inside the window, keeping clean sourcing invoices ready for the day an authenticity complaint lands. The metrics that actually end accounts are not the ones most sellers watch, which is exactly why we mapped them in detail in the five metrics that actually get you suspended. A manager who cannot name those five from memory, and tell you where each one sits today, is not managing your health. They are hoping it holds.

The deeper point is that account health is an early-warning system for operational rot. A spike in pre-fulfilment cancellations is not a health problem. It is an inventory-sync problem that will become a health problem, then a buy-box problem, then an ad-efficiency problem. The manager who reads the whole chain backwards from a single moving number is doing the job. The one who logs the number and moves on is filing it.

Buy-box ownership is where money is quietly lost

Here is the metric most reports skip entirely, because it is uncomfortable to show. Buy-box win rate. You can have healthy traffic, a clean account, and still be handing revenue to a competing seller on your own listing because you lost the buy box and nobody noticed. Every rupee of ad spend you push toward a listing where you do not own the buy box is partly funding someone else’s sale.

A real account manager treats buy-box ownership as a daily ledger, not a quarterly curiosity. They know which SKUs are slipping, why, and whether the cause is price, fulfilment reliability, stock position, or a hijacker on the listing. And critically, they fix it without simply racing the price to zero, because winning the box at a margin you cannot survive is not winning. We laid out that exact discipline in winning the buy box without racing to the bottom. The skill is holding the box on the strength of operations and reliability signals, not on desperation pricing.

This is also where the three levers reveal themselves as one system. Buy-box loss is frequently an account-health symptom in disguise. Poor valid tracking rate and late dispatch erode the fulfilment signals Amazon uses to award the box. So the manager who fixed dispatch to protect health just also protected the buy box, which just also protected ad efficiency, because now the spend lands on a listing the brand actually owns.

Ad efficiency is the third lever, not a separate department

The most common structural failure we see is the wall between account management and advertising. The ads sit with one team or tool, the account sits with another, and nobody owns the seam. So the ad team pours budget into a SKU that is out of stock, or losing the buy box, or sitting on a suppressed listing. The money burns and the report shows a rising ACoS that nobody can explain, because the explanation lives in a different team’s dashboard.

An account manager earns their fee precisely at that seam. They are the person who makes sure advertising spend only flows toward listings that are healthy, in stock, and buy-box-winning. The competence is integrative. It is knowing that the right move this week is to pause spend on a SKU losing the box and redirect it to one that is converting, a decision no ads-only specialist and no reports-only manager can make alone. The discipline that ties this together is what we mean by Marketplace Account Management, and it only works when one owner holds health, buy box and spend in the same head.

What changed recently, and why it raises the bar

The job got harder in 2026, and that is precisely why a passive reporter now costs you more than ever. Two shifts make the point.

First, the economics of the listing changed underneath you. From mid-March 2026, Amazon India removed referral fees on more than 12.5 crore products priced under one thousand rupees across over 1,800 categories, and cut referral fees on several higher-priced categories, with the company claiming sellers can save up to seventy percent on fees, per Outlook Business and Zee Business. A reporter will note the lower fee and move on. An operator reads it as a margin event: the breakeven ACoS on those sub-thousand-rupee SKUs just moved, which means the ad bids that were unprofitable last quarter may be profitable now, and the price you can defend the buy box at just changed. The fee cut is only money in your pocket if someone reprices, rebids and re-prioritises against it. Left alone, it quietly subsidises your competitors who did the math.

Second, account health enforcement is getting more structured, not more forgiving. Amazon has been piloting a Seller Challenge mechanism inside Account Health Assurance that lets eligible sellers request an enhanced review of specific listing-level enforcement actions, with a limited number of challenges per 180-day window, as reported by Amazon Sellers Appeal. The detail that matters: the challenges are rationed. A manager who burns them on weak cases, or fails to keep documentation ready to win one fast, has wasted a finite asset. That is exactly the kind of judgment a reporting layer cannot exercise, because it requires owning the account health story end to end rather than narrating it after the fact.

What earning the fee actually looks like

Strip away the titles and the decks, and a marketplace account manager earns their fee when they can show you a chain of cause and effect they personally drove. The honest test is whether they can complete this sentence with specifics: I noticed X, I traced it to Y, I changed Z, and here is the number that moved as a result. Everything short of that is reporting in a costume.

The things a manager worth paying for actually does in a month:

Catches a health metric trending toward a threshold and fixes the operational cause before it breaches, not after the suspension email.
Reviews buy-box ownership SKU by SKU and recovers boxes lost to fulfilment, stock or hijacking, without collapsing margin.
Reprices and rebids when fee structures move, so a referral-fee cut becomes recovered margin rather than a gift to competitors.
Redirects ad spend away from listings that are out of stock or losing the box, so efficiency improves without cutting budget.
Runs the account on a fixed monthly audit cadence so problems surface as trends, not emergencies.
Produces a report that is the receipt for work done, not a substitute for it.

This is also the simplest way to decide whether to keep a manager or replace one. Look at the last three months. If account health, buy-box ownership and ad efficiency all moved in the right direction, and your manager can tell you which of their actions moved each one, they are earning the fee. If two of the three drifted while you received beautifully formatted reports about it, you are funding a narrator. Fire them, and find someone who treats the account as a system to be operated, not a dataset to be described. That same operator instinct is what separates an in-house hire from a partner, a tradeoff we work through in in-house versus agency account management. The premise of our Marketplace Account Management and Marketplace Growth work is one owner holding health, buy box and spend in the same head, because the three levers were never meant to live in three different inboxes.