You lower your price. Within seconds, so do they. You raise it. They follow. You try a larger cut to see whether they will break ranks, and they match that too, immediately, at any hour of the day or night.
It feels personal, and it almost certainly is not. A competitor who mirrors your every move within seconds is not watching you. They have configured a rule to position relative to a reference, and you are currently setting it. There is no decision being made on the other side at all.
That reframe matters, because it rules out the responses that only work against a human.
What the Behaviour Tells You
Three things, usefully.
They are automated, and faster than manual. Response within seconds or a few minutes means a live integration with a short evaluation cycle. Negotiating, signalling, or waiting them out on a price move will accomplish nothing — there is nobody reading the signal.
They are referencing something you can identify. If they consistently land a fixed amount below you, they are referencing the lowest offer or the Featured Offer price and you are it. If they match you exactly, they are configured to match rather than beat. The pattern tells you the setting.
They have a floor, and you can find it. This is the most actionable fact available to you. No automated competitor is unbounded. Somewhere below your current price is a number they will not pass, and knowing roughly where it sits changes your options entirely.
The Test That Gives You Their Floor
Carefully, and only on a SKU where you can afford the experiment.
Lower your price in modest steps, letting each one settle for a full cycle. Record their response at each level. At some point one of three things happens: they stop following, they follow but stop short, or they match and both of you are now below sustainable pricing.
If they stop following, you have found their floor, and that is valuable information — it is a boundary you can price just above indefinitely, capturing the position at the best price available to you rather than at the bottom.
If they follow you all the way down past the point where either of you can make money, you have learned something different and equally useful: their floor is wrong, or their motive is not profit. Either way, the price war is unwinnable and you should stop running the test.
Two cautions. Do not run this test below your own contribution floor — the information is not worth selling at a loss to obtain. And run it on one SKU, not across a group, because a group-wide descent is expensive and the answer from one SKU usually generalises.
Response 1: Stop Being the Reference
The most effective move, and it is counter-intuitive.
If they are configured to position relative to the lowest offer and you are consistently the lowest offer, you are the input to their pricing. Every move you make is an instruction you are issuing to a competitor’s system.
Stop being the input. Price against the Featured Offer price rather than the bottom of the list, position relative to it based on your non-price strengths, and let someone else occupy the lowest-offer slot. You will often find your price recovers meaningfully and your placement does not change, because the lowest offer was never winning the position in the first place.
Response 2: Compete on What They Cannot Mirror
A price move can be matched in seconds. A delivery promise cannot be matched at all without changing their operation.
Price, delivery speed, seller performance and inventory signals all feed Amazon’s Featured Offer decision. Of those four, only price is instantly copyable. That asymmetry is the whole strategic argument for investing in the other three.
The 2026 change sharpens it. Amazon began removing the separate seller eligibility gate for the Featured Offer from July that year — the US rollout starting in early July, with the EU and UK following later in the month — so offers now enter consideration automatically rather than passing a performance check first. Consideration is not placement, and the ranking inputs remain. With more offers in the pool, the sellers differentiated on something other than price have a clearer advantage than before.
Practically: shorten handling time, tighten your transit promise, hold deeper stock so you are present when they stock out, and protect your defect and late-shipment metrics. Then price above them and find out whether you hold the position anyway. Frequently you will.
Response 3: Add a Deadband and Stop Feeding the Loop
If both of you are responding to each other with small moves, you have built a machine that only descends.
A minimum change threshold — your price will not move unless the required change exceeds a meaningful amount — breaks the cent-by-cent cycle. It costs you nothing, because cent-level positioning is not what separates offers, and it stops you supplying a continuous stream of new inputs to their rule.
Response 4: Wait, with a Defined Window
The response that requires the most discipline and frequently works.
An automated mirror is not a sustainable position for the seller running it. Matching a competitor’s every move means never pricing to their own economics, and sellers in that posture tend to discover the problem when they reconcile a month. They run out of stock faster than they planned, because low prices clear inventory faster than they fund replacement.
So: set your floor at true contribution, hold your price above the race, accept reduced volume, and set a date four to six weeks out to reassess. If they stock out — which mirrors frequently do — the price recovers and you are still there with inventory.
Decide the window in advance. Waiting without a date is not patience, it is drift.
The Response That Does Not Work
Trying to out-cut them.
You cannot win a contest decided by evaluation frequency, and theirs is measured in seconds. Every cut you make is matched before it generates a sale, so you have reduced the market price and gained nothing. Repeated enough times, you arrive at a floor — possibly theirs, possibly yours — with the same share you started with and less margin on every unit.
Guard Against Being the Mirror
Worth asking honestly: is your own rule configured to chase the lowest offer on the listing?
If so, you are somebody else’s mirror, and whichever seller on your ASIN is least disciplined is currently setting your prices. The fix is the same as response one, which is why it is the recommendation regardless of who you think is chasing whom.
The research team at PricingMag has written up the full approach for sellers who say there’s “a competitor matching me within seconds,” including how to run the floor-finding test safely. Their broader Amazon repricing coverage on PricingMag covers the related dynamics, including how shared reference settings turn two sellers into a descending loop.
A mirror is a configuration, not an opponent. Change what it reflects.
