The Amazon Buy Box Algorithm, Explained (2026)

The Amazon Buy Box algorithm is the automated system that decides which seller's offer is featured in the "Add to Cart" and "Buy Now" section of a product page. When several sellers list the same product under a single ASIN, the algorithm evaluates every eligible offer in real time and awards the Featured Offer to the one it judges the lowest risk for that shopper -- based primarily on landed price, delivery speed, and seller reliability.

That one paragraph is the whole game. Everything below is the detail that separates sellers who win the box consistently from sellers who keep guessing. If you want the foundational background first, start with our companion guide on what the Amazon Buy Box is. This article goes one level deeper: not what the box is, but how the machine behind it actually thinks.

The Core Thesis: Lowest-Risk Selection, Not Lowest Price

For years sellers treated the Buy Box like a price war. Undercut everyone by a penny, win the button. That mental model is not just outdated -- it actively loses you money, because it leads you to compete on the one axis Amazon cares about least in isolation.

Amazon's real objective is not to surface the cheapest offer. It is to protect the customer experience. Every time the algorithm features an offer, it is making a bet that the order will arrive on time, in good condition, without a cancellation or a complaint. The Featured Offer is simply the bet Amazon is most confident in. Price is one input into that confidence, but a low price attached to a slow, unreliable seller is a bet Amazon would rather not make.

Once you internalize "lowest risk" as the objective function, every confusing thing about the Buy Box starts to make sense: why a more expensive seller sometimes wins, why your share drops the instant your metrics slip, and why the same offer wins in Dallas and loses in New Jersey.

A Note on "A10" and the Featured Offer Rename

Two naming things trip people up. First, you will see the Amazon seller community refer to the "A9" or "A10" algorithm. Treat these as community shorthand, not official Amazon products. Amazon does not publish a versioned "Buy Box algorithm" with a changelog. The labels are useful as a way to say "the current behavior of the system," but anyone quoting exact percentage weights for A10 is guessing.

Second, Amazon has been steadily renaming the Buy Box to the Featured Offer across Seller Central, its help docs, and its APIs -- Amazon's own seller guide now leads with the new name. The button has not changed. The reports you rely on are simply relabeling "Buy Box percentage" as "Featured Offer" metrics. We use both terms here because the field does.

The Factors the Algorithm Actually Weighs

Nobody outside Amazon knows the exact weights, and they shift over time and by category. But Amazon's Becoming the Featured Offer help page names the inputs and says what each one rewards. The table below takes that list as given. The order is ours, based on seller-community observation.

1. Landed Price (Not Sticker Price)

The algorithm evaluates landed price -- the total cost to the customer, including product price, shipping, and handling. A $15 item with $5 shipping loses to a $19 item with free shipping, because the shopper pays less for the second one.

Price can also take an offer out of the running altogether. The Amazon staff post above points sellers to Pricing Health to "check if any of your offers have pricing issues that make them ineligible." In our observation, a price that jumps well above the listing's usual range is a common cause, even when it is still competitive with other current offers. So check landed price against both your live competitors and the listing's historical norm.

Two adjacent signals are worth separating out, because Amazon's Featured Offer help page lists them as factors in their own right rather than folding them into price: free shipping and free returns. On shipping, Amazon says an all-inclusive price is preferable: a $40 offer with free shipping beats a $35 offer with $5 shipping, even though the total is the same. On returns, Amazon says customers prefer free returns, "especially for frequently returned items like apparel," and an Amazon staff post calls free returns "especially important for frequently returned categories like apparel." Our suggestion: if you sell in a category with a lot of returns and you are losing the box to an offer at the same landed price, free returns are a cheap thing to check before you cut your margin.

2. Fulfillment Speed and the Shipment Proximity Index

Speed is no longer a tie-breaker -- it is a primary factor. Offers using Fulfillment by Amazon (FBA) or Seller Fulfilled Prime (SFP) usually win on this factor because they tend to produce fast, specific delivery promises. Amazon says the fulfillment method itself is not scored: it is the promise that counts.

The deeper mechanic is what sellers call the Shipment Proximity Index: how close your inventory sits to the specific shopper loading the page. Inventory spread across multiple regional fulfillment centers can promise same-day or next-day delivery to far more buyers than stock in a single warehouse. That wider fast-delivery radius is a direct ranking advantage, which is exactly why the Buy Box winner can change by ZIP code.

3. Account Health and the On-Time Delivery Rate

If you self-fulfill, your operational metrics are under a microscope. Amazon names the seller-performance signals it weighs for the Featured Offer: order volume, chargeback rate, and Voice of the Customer complaints. Separately, account-health requirements decide whether your listings stay active at all.

The metric with teeth is your On-Time Delivery Rate (OTDR). Amazon's February 2026 announcement describes a 90% on-time delivery rate requirement for seller-fulfilled listings: fall below it and the listings with the most impact on the drop are deactivated, and if you are significantly or repeatedly below it, Amazon may deactivate all of your seller-fulfilled listings. That is enforcement, not ranking, but a deactivated listing cannot win anything. Our recommendation is to treat 95% or higher as the working target rather than the bare minimum.

Order Defect Rate, cancellation rate, and valid tracking rate are account-health metrics too. Amazon does not list them among the Featured Offer factors, so treat them as the enforcement side of account health: they decide whether you keep selling, and in our experience they tend to slip together with the chargebacks and complaints Amazon does weigh.

4. How Fulfillment Method Shapes Your Odds

Your choice of fulfillment path shapes the delivery promise you can make, which is why it matters so much in practice even though Amazon does not score the method itself. FBA hands Amazon the delivery promise outright. Seller Fulfilled Prime lets you keep fulfillment in-house while earning the Prime badge. Standard FBM can compete when it ships fast from near the buyer.

One correction worth making, because a lot of guides state the opposite: Amazon's seller guide says fulfilling orders directly can be "just as effective at increasing your chances of becoming the Featured Offer" as fulfilling through Amazon. Its Featured Offer help page says that if all other factors are equal, FBA and seller-fulfilled offers are equally likely to be featured, and an Amazon staff post lists fulfillment method among the factors that don't matter: "FBA vs. Seller Fulfilled are treated equally." In other words, the fulfillment method is not itself a scored input. Our read of why FBA still wins the box so often: it usually produces the fast, specific delivery promise that Amazon does score. So a seller-fulfilled offer shipping quickly from near the buyer can compete, and chasing FBA will not fix a delivery promise that is slow for other reasons. FBA is also not instant: stock has to be shipped in and received, and Amazon staff say it can take up to 18 days after receipt for units to be available for immediate shipping.

5. Secondary Signals and Tie-Breakers

When two offers are closely matched, Amazon says it can rotate the Featured Offer between them rather than pick one. What tips a close race is not published. In our observation, stock depth matters here: offers running low tend to lose share before they actually sell out. One thing that does not tip it, according to Amazon, is customer reviews or star ratings; Amazon evaluates seller performance through the factors in the table above instead. These rarely override a clear price or speed advantage, but in a tight race they are the margin.

What Changed and When: A Buy Box Algorithm Timeline

Amazon does not publish a changelog for the Featured Offer, which is why so much Buy Box advice on the web is quietly years out of date. There is no version history to read -- but there is a public record, assembled from Amazon's own seller announcements and from regulators who have had to describe the mechanism on the record. Here is that record, dated.

1. December 2021

   Italy fines Amazon over Prime, FBA and the Buy Box

   Italy's competition authority fined Amazon EUR 1.128 billion, finding that Amazon kept sellers from using the Prime label on offers not fulfilled through FBA, and stating that the Prime label increases the likelihood of a seller's offer being selected as the Featured Offer in the Buy Box. In December 2025 the authority redetermined the fine at EUR 752.4 million in line with an Italian court ruling, in a decision published in January 2026.

   Why it matters: A regulator stating on the record that a seller's fulfillment choice fed Buy Box outcomes.

   Sources: AGCM press release, AGCM Bulletin 2/2026 (PDF)

2. December 2022

   The EU makes equal treatment in the Buy Box a binding commitment

   The European Commission accepted commitments from Amazon to treat all sellers equally when ranking offers to select the Buy Box winner, and to display a second competing offer when an offer from a different seller is sufficiently differentiated on price and/or delivery. The Prime and second-offer commitments run for seven years; the rest run for five. They cover Amazon's EEA marketplaces, but the Buy Box and Prime commitments exclude Italy.

   Why it matters: This is why European product pages can show a second offer beside the Buy Box winner. We know of no US equivalent.

   Source: European Commission

3. September 2023

   The FTC and 17 states sue Amazon

   The FTC and 17 state attorneys general sued Amazon for monopolization. The Buy Box is central to the complaint's anti-discounting allegations: it alleges that Amazon punishes sellers who offer lower prices elsewhere, including by knocking them out of the Buy Box. It separately alleges Project Nessie, an algorithm it says Amazon used to raise prices by predicting when other online stores would follow an Amazon price increase.

   Why it matters: Our read: Nessie is a pricing algorithm, not the Featured Offer selection algorithm, and the Buy Box claims are about how Amazon uses the box, not a claim that the selection logic itself is illegal.

   Sources: FTC press release, FTC complaint (PDF)

4. September 2024

   The federal claims survive dismissal

   On September 30, 2024 the court denied Amazon's motion to dismiss the Sherman Act and FTC Act claims, and dismissed a small number of state-law claims.

   Why it matters: Our read: the monopolization claims that carry the Buy Box allegations are headed for trial.

   Source: Court order (PDF)

5. July 2026

   Amazon removes the seller-eligibility step

   Amazon began removing seller eligibility requirements for the Featured Offer, saying the first seller eligibility step was no longer delivering additional value to customers. It also said it is not changing how the Featured Offer is selected: offers are still evaluated on criteria such as competitive pricing, delivery speed, and performance. The rollout is gradual across all Amazon stores and completes by the end of 2026. Sellers need take no action.

   Why it matters: Our read: the biggest practical change in years. More offers can now compete on each listing, and the usual diagnosis for an empty Buy Box shifts from 'I am not eligible' to 'I am eligible and losing'.

   Sources: Amazon seller announcement, Amazon Featured Offer help

6. March 2027

   The FTC case is scheduled for trial

   The trial is scheduled to begin on March 29, 2027 in the Western District of Washington, as a bench trial on liability. In September 2026 Amazon said it plans to ask to postpone it, so treat the date as provisional.

   Why it matters: Our expectation: nothing about the algorithm changes because of a trial date, but trial exhibits could be the most detailed public look at the Buy Box sellers have had.

   Sources: MLex (September 2026), Court order (bench trial)

Two things fall out of that record, in our reading. First, the most consequential change is the most recent one, and it is the one least reflected in the guides we checked: since July 2026 there is no separate seller check in front of the ranking, while the ranking itself, by Amazon's account, is unchanged. Second, the written constraints on how the Buy Box winner is picked that we could find are regulatory, chiefly the EU commitments. If you want to know how the algorithm behaves, Amazon's own help pages and competition filings are better sources than another listicle.

Why the Buy Box Rotates and Shares

One of the most common questions is how often the Buy Box changes. The answer is: potentially on every page load. Because the algorithm re-evaluates offers in real time, the winner shifts as prices update, stock levels move, and delivery estimates change throughout the day.

When several sellers are genuinely comparable, Amazon does not pick one and freeze it. It shares the box, rotating it among the strong offers. The split is not even -- it is weighted toward the lower-risk sellers. If you and three competitors are closely matched but your metrics are slightly better, you might hold the box 40% of the time while the others divide the rest. This is why a healthy reseller target is roughly 60-70% rather than 100%.

The July 2026 change has a consequence here that is easy to miss. This part is our own reasoning, not something Amazon has said: because the seller-eligibility step no longer filters the pool, more offers can enter the rotation on a typical listing. Offers that the old check would have screened out are now evaluated like any other. For a strong seller this mostly changes nothing, since a weak offer still ranks below you. But on listings where you were winning partly because competitors were shut out, your share can drift down without any of your own metrics moving. If your Buy Box percentage slipped in the second half of 2026 and nothing in your account changed, check this first.

Why Your Buy Box Changes by ZIP Code

This is the single most under-appreciated consequence of the algorithm, so it deserves its own section. Because proximity feeds delivery speed, and delivery speed is a primary factor, the same ASIN can have different Featured Offer winners in different parts of the country at the same moment.

The practical problem: Seller Central gives you a single, nationwide Buy Box percentage. If that number reads 75%, you have no idea whether you are winning everywhere at 75% or dominating half the country and losing the other half completely. Those are very different businesses with very different fixes.

Our own tracking data puts a real number to how common this split is. Across 8,397 Buy Box positions monitored on the Buy Box Checker platform, covering 1,775 ASINs and 116 delivery locations, 20.6% of US products checked in two or more locations on a given day show a different Buy Box status across those locations that same day (Buy Box Checker platform data, August 15, 2026). Our reading is that the proximity mechanics described above drive most of that variance: the same ASIN, the same seller, the same price, but a different delivery estimate depending on where the shopper is standing. Some share of it is also checks that returned no clear status, which the figure counts as a difference.

The pillar guide on what the Amazon Buy Box is shows the full data breakdown, including suppression and lost rates. This is the mechanism behind those numbers: fulfillment geography is not an edge case. Across all 1,774 tracked products, 45.4% have at least one location where the Buy Box is lost or suppressed.

This is the specific gap BuyBoxChecker.com was built to close: it checks Buy Box ownership in specific ZIP codes so you can see the regional map behind your blended number, for free.

When No One Wins: Buy Box Suppression

Sometimes the algorithm declines to feature any offer. Instead of an "Add to Cart" button, shoppers see a gray "See All Buying Options" button. This is suppression, and it usually destroys conversion because it forces shoppers to choose an offer manually.

The most common triggers:

• External price competition. Amazon continuously scans the web. If it finds your product cheaper at Walmart, Target, or your own Shopify store, it may suppress the box to avoid featuring a price it considers uncompetitive.
• Price spikes above the historical range. A sudden jump well above the product's normal price can trigger suppression even with no competitors.
• Lost eligibility. An inactive listing, or an offer that is out of stock or cannot ship to the shopper, drops out of contention. Seller-performance metrics no longer do this on their own: Amazon began removing that gate in July 2026, so poor metrics now count against you in the ranking rather than keeping you out. An Account Health enforcement action that deactivates a listing still removes it.

Suppression is dangerous precisely because Seller Central does not alert you loudly. Your sales simply fall off a cliff. Catching it early is one of the strongest arguments for monitoring Buy Box status continuously rather than checking a weekly report. If you are already in this situation, the recovery checklist for a lost or suppressed Buy Box walks through the specific steps to diagnose the cause and get the Featured Offer restored.

How to Read the Algorithm's Signals as a Seller

You will never see the algorithm's internal scores. But it is not a black box you are helpless against -- you can observe its outputs and work backward. A practical reading routine looks like this:

1. Track ownership over time, not just a snapshot. A 70% that is trending down means something different from a 70% trending up.
2. Track it by location. Regional gaps point straight at a fulfillment or inventory-placement problem rather than a pricing one.
3. Correlate drops with changes. When share falls, check what moved: a competitor's price, your OTDR, a stock-out, or a suspected suppression event.
4. Separate "eligible but losing" from "suppressed." The first is a competition problem; the second is usually a pricing or listing problem. They need different responses.

Do this consistently and the algorithm becomes legible. You stop reacting to a mystery number and start diagnosing specific, fixable causes. For a side-by-side look at the tools that surface these signals automatically, see the comparison of Buy Box monitoring tools.

Frequently Asked Questions