Sellers talk about the Buy Box like it is weather. Some days you have it, some days you don’t, and nobody knows why.
That is not how it works.
The Buy Box – Amazon now calls it the Featured Offer – is not random, and it is not a lottery. It is an allocation decision made by a system with specific inputs, and every one of those inputs is something you can measure and influence.
If you understand what the machine is actually weighing, the Buy Box stops being a mystery and starts being a checklist.
What the Featured Offer Actually Is
A product detail page on Amazon is not a listing that belongs to one seller. It is a shared page for one product, and every seller offering that product competes to be the offer attached to the “Add to Cart” and “Buy Now” buttons.
That slot is the Featured Offer. Everyone else is buried under “Other Sellers on Amazon”, a link most buyers never click.
This matters because of a number that should shape how you think about everything on the platform: over 80% of Amazon sales go through the Buy Box. On mobile, where the other offers are pushed even further out of view, the share is higher still.
Losing the Buy Box does not mean selling less. In most cases it means not selling at all.
So the real question is not “how do I get more traffic.” It is “when Amazon has to pick one offer to feature, why would it pick mine.”
What Amazon Is Optimizing For
Amazon does not award the Featured Offer to the seller who deserves it most. It awards it to the offer that is least likely to create a bad buying experience.
Fewer complaints. Fewer late deliveries. Fewer returns. Fewer refunds Amazon has to process.
Every input the algorithm weighs ties back to that goal. Once you see it that way, the individual factors stop looking arbitrary.
One structural note, because it is changing as this is written: Amazon historically ran a pass-or-fail eligibility check on seller performance before an offer could even compete for the Featured Offer. Starting July 2026, that standalone gate is being removed, rolling out globally through the end of the year. The metrics did not stop mattering – they moved inside the ranking formula itself, as direct inputs alongside price and delivery. Which is the point of this whole article: it is one scoring system, and everything below is an input to it.
The Real Inputs
1. Landed Price
Not your item price. Your landed price: item price plus shipping.
A $19.90 offer with $4.90 shipping loses to a $23.90 offer with free shipping, even though the first seller thinks he is cheaper.
But – and this is the part sellers get wrong – the lowest landed price does not automatically win. Price is weighed against everything else. A seller with excellent metrics and FBA can hold the Buy Box at a higher price than a marginal seller undercutting them. Amazon effectively grants a price premium to offers it trusts.
How large that premium is depends on the gap in the other inputs. It is rarely huge. It is real.
2. Fulfillment Method and Delivery Promise
This is the heaviest single factor, and it is the one sellers most want to argue with.
FBA offers carry Amazon’s own delivery promise. Amazon knows exactly when the package will arrive because Amazon is delivering it. That certainty is worth more to the algorithm than almost anything a merchant-fulfilled seller can offer.
FBM can compete, but only with fast, reliable settings: short handling time, premium shipping options, and a track record that proves the promises are kept. Seller Fulfilled Prime sits in between, with strict performance requirements precisely because Amazon is lending you its promise.
In practice: an FBM offer usually needs a meaningfully lower landed price to take the Buy Box from a comparable FBA offer. If your handling time is set to two days “to be safe”, you are paying for that safety in Buy Box share.
3. Seller Performance Metrics
Three numbers do most of the work here:
- Order Defect Rate (ODR) – negative feedback, A-to-z claims, chargebacks. Amazon’s threshold is under 1%, but the algorithm does not wait for you to cross it. A rising ODR quietly costs you Buy Box share long before you get a warning.
- Late Shipment Rate – orders confirmed as shipped after the expected ship date. The bar is under 4%. For FBM sellers this is the metric that most directly maps to Buy Box eligibility.
- Valid Tracking Rate – Amazon wants 95% or better. No tracking means Amazon cannot verify the delivery promise, and unverifiable promises lose to verified ones every time.
Cancellation rate and customer response time feed in as well. FBA makes most of this moot: late shipment and tracking are Amazon’s problem by definition, and fulfillment-related complaints are struck from your record. Product-related defects – chargebacks, item-not-as-described claims – still count against you. That is a large part of why FBA wins so often. You are not just buying logistics. You are buying a nearly clean scorecard on the metrics the algorithm cares about.
4. Stock Availability
An offer with no inventory is not an offer. Obvious, but the second-order effect is not: inventory depth matters, not just presence.
If you are down to your last few units, or your FBA stock is sitting in a single fulfillment center far from the buyer, your effective delivery promise weakens and your share can drop before you technically stock out. And after a stockout, the Buy Box does not snap back the moment inventory lands. The system re-establishes confidence gradually.
Consistent availability is a ranking input, not a hygiene task.
Rotation: Why “Winning” Is Never 100%
When two or more offers are genuinely close – similar landed price, both FBA, clean metrics – Amazon does not pick one and stop. It rotates the Featured Offer between them, splitting time roughly in proportion to how strong each offer is.
This is why your Buy Box percentage in Seller Central is a percentage and not a yes/no flag. 70% Buy Box share means the algorithm considers your offer better than the alternatives most of the time, not all of it.
It also means small improvements move the needle. You do not need to beat a matched competitor decisively. Shaving your handling time or closing a 2% price gap shifts the rotation in your favor, and the revenue difference between 60% and 85% share is substantial.
Suppression: When Nobody Wins
Sometimes the buttons disappear entirely and the page shows “See All Buying Options” instead. No seller has the Buy Box. That is suppression, and it is the algorithm saying: none of these offers meets the bar to be recommended.
The usual triggers:
- Price above the reference range. If every offer is priced well above the product’s recent history or its price elsewhere on the web, Amazon would rather feature nothing than feature what looks like a bad deal.
- No offer worth featuring. Every offer on the listing scores badly – weak metrics, long delivery promises, thin history.
- Pricing errors and instability. Wild price swings – often caused by misconfigured repricers – can trigger suppression on their own.
A suppressed Buy Box is a conversion killer. Buyers who have to click through a list of offers abandon at a much higher rate. If you sell private label and your own listing is suppressed, it is almost always a pricing signal problem – fix the price relative to your own history and off-Amazon presence first.
Repricers: Tool or Treadmill
Repricers exist because the Buy Box is dynamic. Competitors change prices, stock out, and come back, and the winning conditions shift hour by hour.
Used well, a repricer does two things. It wins share when a competitor stocks out or drifts upward, by moving your price up as well as down. And it keeps you competitive without you watching the listing all day.
Used badly, it does one thing: it races you to the bottom. Two sellers running naive “beat the lowest price by $0.01” rules will walk a $24 product down to $16 in an afternoon, destroy the margin for everyone, and – because of the reference-price logic above – sometimes drag the listing toward suppression when prices later try to recover.
The rules that separate the two outcomes are simple:
- Set a floor based on your real landed cost plus minimum acceptable margin. Never let the tool go below it.
- Use a repricer that raises prices when competition thins. Downward-only repricing is margin donation.
- Compete against comparable offers. Matching a marginal FBM seller’s price when you are FBA with clean metrics is giving away the premium the algorithm already grants you.
If you are a reseller on shared listings, a good repricer is close to mandatory. If you are private label, you probably do not need one at all – which brings us to the last point.
Private Label vs Reseller: Two Different Games
The Buy Box conversation splits cleanly depending on who else is on your listing.
If You Sell Private Label
You are usually the only seller on the listing, so you are not competing against other offers. You are competing against the system’s quality bar. Your checklist:
- Stay in stock. Depth, not just presence – stockouts cost you the Buy Box now and momentum after.
- Keep your price stable relative to your own history and any off-Amazon channels. Big promotional swings and cross-channel undercutting are the main causes of private-label suppression.
- Use FBA or meet SFP standards, so the delivery promise is never the weak point.
- Watch for hijackers. The moment another seller appears on your listing, you are in the reseller game whether you like it or not.
If You Resell on Shared Listings
You are in the competitive version of the game. Your checklist:
- Know your landed cost per SKU and set repricer floors from it. No exceptions.
- Prefer FBA on contested listings, or run FBM with genuinely short handling times – not padded ones.
- Guard ODR, late shipment rate, and valid tracking rate like the ranking inputs they are.
- Track your Buy Box percentage per SKU weekly. A falling share is the earliest visible symptom of a metric or price problem.
- Stop fighting for listings where the math does not work. A Buy Box won below your floor is not a win.
Closing
The Buy Box is the single highest-leverage mechanism on Amazon, and it runs on inputs you control: landed price, delivery promise, performance metrics, and availability.
Sellers who treat it as luck react to losing it. Sellers who treat it as a system prevent losing it.
The machine is not picking names out of a hat. It is scoring offers. Make yours the one that is easiest to trust, and the rotation takes care of the rest.