Something is happening inside Amazon’s Buy Box engineering team, and it’s making some of the platform’s largest third-party sellers deeply nervous. According to multiple sources close to the matter, Amazon has been running an unannounced A/B test since approximately late March 2026 that reportedly restructures how the Buy Box eligibility algorithm weights its core signals — potentially demoting price competitiveness in favor of a composite “customer experience score” that folds in return rate, delivery speed variance, and AI-generated review sentiment analysis.
The alleged changes, which Amazon has not publicly acknowledged, are said to be code-named internally as “Project Meridian.” Ecommerce Times has not independently verified that name, and Amazon declined to comment for this story. But the chatter across private Seller Central forums, closed Slack communities, and at last month’s SellerCon event in Las Vegas was loud enough that several large-volume aggregators have reportedly begun emergency audits of their own account health dashboards.
What Exactly Is Amazon Allegedly Changing About the Buy Box?
The traditional Buy Box formula — which has long been an open secret among sophisticated sellers — heavily weights landed price, fulfillment method (FBA typically winning over FBM), seller feedback score, and shipping time. Sellers who have spent years optimizing their repricers around that model are reportedly alarmed that the goalposts may be moving without notice.
Sources familiar with the alleged test say the new composite scoring model introduces three new weighted variables:
- Return friction rate: How often buyers initiate returns specifically citing “item not as described” — a metric that has historically been tracked but not surfaced as a Buy Box signal.
- Delivery promise accuracy: Whether the seller’s actual delivery date matches the promised window at the time of purchase, scored at the ASIN level rather than account level.
- Sentiment delta: A proprietary AI-generated score measuring whether recent reviews trend more negative than the product’s historical baseline — potentially flagging sellers who win Buy Box by undercutting but then deliver inferior product batches.
If accurate, the implications would be significant. Sellers running aggressive repricing strategies — particularly aggregators managing hundreds of ASINs — could see Buy Box suppression on products where their operational metrics are softer, even if their price is the lowest on the listing.
Which Sellers and Aggregators Are Most Exposed?
The aggregator community is reportedly the most rattled. Several sources say that Thrasio, which has spent much of 2025 and early 2026 rebuilding its operational infrastructure after its bankruptcy restructuring, is among the brands closely watching their Buy Box share data for anomalies. Representatives for Thrasio did not respond to a request for comment.
Perch, the Boston-based aggregator that has quietly continued acquiring brands through 2025, is also said to be running internal diagnostics. One agency operator who manages PPC for multiple aggregator portfolios told Ecommerce Times, on condition of anonymity, that they’ve seen “unexplained Buy Box suppression on six ASINs across two client accounts since April” that doesn’t correlate with pricing changes.
“We’ve eliminated every known variable — price, IPI score, feedback rating — and we still can’t account for why two of our top performers lost Buy Box share in May. Something changed in the weighting, and nobody at Seller Central support can tell us what.” — Senior account manager at a mid-size Amazon aggregator, speaking anonymously
Sellers operating in competitive commodity categories — supplements, electronics accessories, home goods — are said to be disproportionately affected, because those are categories where multiple sellers are often within pennies of each other on price, meaning secondary signals would have outsized tiebreaker influence.
How Are Repricing Tool Vendors Responding to the Alleged Shift?
The rumored changes have reportedly sent ripples through the repricing software ecosystem. Vendors including Feedvisor, Informed.co, and BQool are all understood to be monitoring the situation. Feedvisor’s enterprise team is said to have quietly briefed several of its larger clients in May about the possibility of algorithm weighting shifts, though the company has not issued any public guidance.
Victor Rosenman, CEO of Feedvisor, did not respond directly to questions about Project Meridian, but in a statement provided to Ecommerce Times, he said:
“The Buy Box has never been purely a price race, and the most sophisticated sellers have always known that. Any evolution toward operational quality signals would be consistent with Amazon’s long-term direction. We’re continuously updating our models to reflect what the data actually shows.” — Victor Rosenman, CEO, Feedvisor
Sources say that Informed.co has internally flagged the anomalies to its data science team and is running comparative Buy Box win-rate analysis across its customer base to determine whether there’s a statistically significant pattern. An Informed.co spokesperson declined to comment on internal processes.
Is Amazon Trying to Squeeze Out FBM Sellers Entirely?
A secondary layer of speculation among seller communities involves whether Project Meridian — if real — is also designed to further disadvantage Fulfilled by Merchant sellers in categories where Amazon’s own logistics network has high coverage. The delivery promise accuracy metric, in particular, is seen by some sellers as structurally biased against FBM operators who rely on third-party carriers and lack Amazon’s own last-mile infrastructure.
Kevin King, a well-known Amazon educator and host of the Billion Dollar Seller Summit, posted in a private mastermind group last month — a post that was subsequently screenshotted and circulated widely — that he’d heard from “two separate seven-figure sellers” that FBM accounts in their categories had seen Buy Box eligibility drop by as much as 15 percentage points since April without any corresponding account health warnings.
“If the delivery accuracy scoring is being measured against Amazon’s own logistics benchmarks, then FBM sellers are being graded on a curve that’s designed for them to fail. That’s not a quality improvement — that’s a funnel into FBA fees.” — Kevin King, host, Billion Dollar Seller Summit
Amazon’s financial incentive here is not subtle. FBA fees, which were restructured in early 2025 with the rollout of the tiered “Seller Fulfilled Prime 2.0” program, remain a significant revenue driver for the company’s third-party services segment. Any algorithm change that nudges sellers toward FBA enrollment would have direct revenue implications.
What Are Sellers Actually Doing Right Now?
In the absence of confirmed information from Amazon, sellers are taking a range of defensive actions. Agency leaders and brand operators report a surge in demand for account health audits, with tools like Helium 10’s Market Tracker 360 and DataDive being used to benchmark Buy Box win rates against category-level norms.
Among the tactics reportedly being deployed:
- Auditing return reason codes at the ASIN level to identify “item not as described” clusters that could flag poorly performing inventory batches.
- Stress-testing delivery promise accuracy by reviewing the gap between estimated delivery shown at checkout and actual carrier scan data, using tools like ShipStation’s analytics dashboard.
- Temporarily pulling aggressive repricing floors on high-volume ASINs to reduce the risk of winning Buy Box with inventory that carries elevated return rates.
- Escalating to Amazon Strategic Account Services (SAS) reps — for sellers enrolled in that program — to request any available account-level transparency on Buy Box suppression signals.
- Increasing FBA inbound shipments on key ASINs as a hedge against FBM exposure, despite the associated storage and fulfillment cost increases.
One Shopify-to-Amazon multichannel brand founder who operates a seven-figure home goods business told Ecommerce Times they’ve halted all FBM listings for SKUs with more than $10,000 in monthly Amazon revenue until “there’s more clarity on what’s actually being tested.”
When Will Amazon Officially Address These Changes?
That remains the central frustration for sellers. Amazon’s pattern with Buy Box algorithm updates has historically been to make changes quietly, observe outcomes, and only disclose changes retroactively — if at all — through updates to Seller Central help documentation.
Several attendees at the Amazon Accelerate pre-conference that took place in Seattle in late May reported that Amazon policy team representatives deflected direct questions about Buy Box weighting changes with language around “continuous optimization of the customer experience.” No formal announcement has been made, and no Seller Central communication has been issued.
What is clear is that the uncertainty itself is having a measurable effect. Multiple agency leaders report that clients are pausing planned catalog expansion decisions on Amazon pending clearer signals, and at least two mid-size brands with significant Amazon revenue are said to be accelerating their direct-to-consumer channel buildout on Shopify as a risk hedge.
Whether Project Meridian is a major structural overhaul, a limited A/B test that will be quietly rolled back, or simply a pattern of normal algorithmic noise being over-interpreted by anxious sellers remains genuinely unclear. But in a marketplace where Buy Box ownership is the difference between a product that moves and one that collects dust in an FBA warehouse, the anxiety is entirely rational — and Amazon’s silence is doing nothing to calm it.
