Everyone on the listing is undercutting everyone, the price has fallen 18% in three weeks, nothing about the product or the demand has changed, and every seller involved is losing money relative to where they started.
The standard explanation is that someone is being aggressive. Usually nobody is. A race to the bottom is what happens when several sellers independently run reasonable automation configured in the same reasonable way — and the result is a machine that converts a healthy market into a thin one without any person deciding to.
Understanding the mechanism is what makes the exits visible.
How the Machine Works
Four conditions produce it, and they are all defaults.
One: everyone references the lowest offer. If each seller’s rule positions relative to the cheapest price on the listing, then one price cut is read by everyone as the new reality. Each responds, the response becomes the new lowest offer, and the next round begins. Nobody is targeting anybody. The system is simply recursive.
Two: nobody’s rule goes up. Most rule sets describe descent in detail and say nothing about recovery. So the price ratchets: down when anyone cuts, never back when the cutter disappears. Over months this is the single largest cause of permanent price erosion on stable ASINs.
Three: at least one floor is wrong. Someone’s minimum price is built on gross margin rather than contribution, or predates a fee change. That seller can go lower than their economics actually allow, and the whole group follows them down to a price that is unprofitable for everyone including them.
Four: the speeds are mismatched. Automated sellers move in minutes; manual sellers move in days. Manual sellers therefore spend most of their time out of position, cut harder when they finally act to compensate, and inject a large downward step that automation amplifies.
None of these is aggression. All of them are configuration.
Exit 1: Stop Referencing the Bottom
The highest-return change available, and it is one dropdown.
Reference the Featured Offer price instead of the lowest offer on the listing. You stop reacting to outliers — the liquidator, the abandoned listing, the seller with a broken floor — and start reacting to the price at which Amazon’s ranking has concluded an offer is actually working.
On a listing where the Featured Offer consistently trades above the lowest offer, this change alone recovers margin immediately, usually without costing you placement, because the offer you were chasing was never taking your placement in the first place.
Exit 2: Add the Instruction to Go Back Up
If your rule has no upward condition, your price can only ever ratchet downward, and every temporary price cut by anyone becomes a permanent feature of your pricing.
Configure recovery: when the competitive reference rises, or when the offer that set the low price goes out of stock, your price should move back. Both Amazon’s own Automate Pricing and third-party rule engines support this. It is simply not on by default in most people’s configurations, because sellers set up automation while worrying about being too expensive and never revisit it.
Expect this to feel uncomfortable the first time. Prices rising is the sensation of margin returning.
Exit 3: Compete on Something That Cannot be Matched in Ninety Seconds
A price advantage lasts as long as it takes competing automation to notice. A delivery advantage lasts as long as your competitor’s logistics.
This is the structural reason price competition is a poor long-term strategy on a shared listing: price is visible and instantly copyable, and everything else is not. Delivery speed, inventory depth and seller performance all feed Amazon’s Featured Offer decision, and none of them can be matched by a rule.
The 2026 context sharpens this. Amazon began removing the separate seller eligibility gate for the Featured Offer from July 2026 — US first, EU and UK later that month — so offers now enter consideration automatically. Consideration is not placement, and price, delivery speed, performance and inventory still determine the winner. With more offers in the running, the sellers who have invested in the non-price inputs hold a clearer differentiator than the ones holding only a price.
Practically: shorten handling time, tighten the transit promise, hold deeper stock on your core SKUs, protect your defect and late-shipment metrics. Then price above the race rather than inside it, and find out whether you hold position anyway. Many sellers discover they do.
Exit 4: Leave, Temporarily
The exit nobody considers, and often the correct one.
If the prevailing price on a listing is below your contribution floor, participating is not a strategy. Every sale is a loss, and automation will keep making them on your behalf for as long as you leave it running.
Setting your floor at your real contribution minimum and letting your price sit above the market has a specific consequence: you stop selling that SKU for a while. That is not failure. That is declining to buy market share with your own capital.
Two things usually follow. The sellers sustaining the low price run out of stock, because low prices clear inventory faster than they fund replacement — at which point the price recovers and you are still there. Or they do not, in which case you have learned that this ASIN is structurally unprofitable for you and the right move is to stop restocking it rather than to keep funding the discovery.
Both outcomes are better than the alternative, which is participating in a market you cannot make money in because leaving feels like losing.
What Not to do
Do not match a below-cost competitor on principle. You cannot win a contest decided by who can afford to lose longest, and the participant with a liquidation motive has no floor to reach.
Do not cut deeper to “end it quickly.” Nothing ends. Automation on the other side treats your cut as input.
Do not assume the lowest-priced seller is making money. Frequently they are not, and their floor is wrong rather than low. Following a mistake does not make it correct.
The Reframe
A race to the bottom looks like a competitive problem and is nearly always a configuration problem — a reference setting, a missing upward condition, a floor built on the wrong arithmetic. The sellers who escape them are rarely the ones who fight hardest. They are the ones who stopped reacting to the bottom of the offer list and started being willing to sit out a price they could not make money at.
The research team at PricingMag has mapped all four exits for sellers in the position when everyone on your listing is racing to the bottom, including how to judge whether a low price is sustainable for the seller holding it.
The machine only runs while everyone keeps feeding it. The exit is usually one setting away.