Most Amazon repricing strategies reduce to three parameters: the floor you refuse to go below, the set of competing offers you allow the rule to react to, and how often the rule is permitted to move. Pick those three deliberately and the choice between a rule-based and an AI repricer stops being the important decision. Pick them by default and no tool will save the margin.

This page is about how to set the parameters themselves. If you are still choosing an engine, the Amazon repricing tools hub covers the tool landscape, and best Amazon repricer compared works through the head-to-head selection. If you want general price-point theory — bundle pricing, charm pricing, launch discounting — that is a different problem, covered in Amazon pricing strategies . Repricing is the narrower question of what an automated rule does to a live price, every day, without you watching.

The three parameters every repricing rule comes down to

Strip the marketing off any repricer, including Amazon’s own free one, and you are configuring the same three things.

ParameterWhat it decidesFailure mode if left on default
Floor (and ceiling)The band the price may move insideA rule that follows a competitor below your true break-even
Competitor filterWhich rival offers count as “the competition”Chasing offers that were never going to win the placement anyway
CadenceHow fast and how often the price may changeReacting to noise, or reacting a day late

Amazon’s own Automate Pricing tool exposes exactly this shape. Its rule types are described on the official page (checked 2026-08-02) as “Competitive price-based rules” that “Compare against the Featured Offer, lowest price, or lowest external price”, “Sales-based rules” that “Change your price based on your sales volume in a specified period of time”, and business-price rules; the setup step is “Manage your margins by setting a minimum price and an optional maximum price.” The tool itself is free — Amazon states it “is included for free in the Professional selling plan” (Amazon Automate Pricing , checked 2026-08-02). Third-party repricers differ in speed, in how the band is calculated for you, and in how much of the decision they take over — not in the underlying anatomy.

Parameter 1: build the floor from landed cost, not from a number that feels safe

A floor typed in by hand is a guess about your unit economics. A floor derived from cost is a statement about them. The difference shows up the first week a competitor decides to dump inventory.

The line items that belong in the calculation:

  • Unit cost from the supplier, at the exchange rate you actually paid
  • Inbound freight and duties, allocated per unit
  • Prep, labelling and any 3PL handling
  • Amazon’s referral fee for the category, as a percentage of the final price — which means the floor is a fixed point, not a subtraction (the fee moves when the price moves)
  • Fulfilment fee, plus monthly and long-term storage allocated across expected sell-through
  • A returns and damage allowance, taken from your own return rate rather than a category average
  • Advertising cost per unit, if the ASIN only sells with ads running
  • The margin you are actually trying to defend

Two of those are the ones sellers most often leave out, and both bite in the same direction. Referral fee as a percentage means a floor computed against your current price is wrong at any other price. Advertising cost per unit means the same ASIN has two different break-even prices: run the inputs in the worked example below without the ad line and it breaks even at $15.47, run them with $2.40 of ad cost per unit and it breaks even at $18.29 (both at that table’s 15% referral fee, before any target margin) — and a repricer that does not know this will happily undercut into a loss that only appears in next month’s P&L.

A worked illustration — the figures below are hypothetical inputs chosen to show the arithmetic, not category benchmarks, so substitute your own from a Seller Central fee preview:

Line itemPer unit
Supplier cost$6.20
Freight, duty, prep$1.35
Fulfilment fee$4.75
Storage allocation$0.30
Returns and damage allowance$0.55
Ad cost per unit sold$2.40
Subtotal before referral fee and margin$15.55
Referral fee15% of price
Target contribution margin12% of price
Floor$15.55 ÷ (1 − 0.15 − 0.12) = $21.30

The last row is the part worth internalising: because two of the costs scale with price, the floor is the subtotal divided by what is left of the dollar after them, not the subtotal plus a markup. Setting the floor at $17.50 “to be competitive” would put every sale below cost while the dashboard still shows the Featured Offer being won.

Ceilings get less attention and cost less money, but they matter on thin-competition ASINs where a rule can drift a price up until conversion quietly stops. Amazon’s guidance on its own tool is to “Set a maximum price, or let us automatically ensure your offers stay competitive while never going significantly higher than recent prices” (checked 2026-08-02). Third-party tools take a similar position: Aura’s page states “Protect your margins with automated price floors and ceilings” and that “Maven can even set your maximum prices for you — learning where each listing tops out as the market tests the ceiling” (goaura.com, checked 2026-08-02; see our Aura review for how that plays out in user reports).

Which tools automate cost-derived floors rather than flat ones: Seller Snap’s own description says its repricer “is designed to consider all costs—like fees, shipping, and minimum ROI—to make sure your pricing strategy stays profitable” (sellersnap.io, checked 2026-08-02), and Repricer.com gates a “Net Margin repricing” feature to its upper plans (repricer.com/pricing comparison table, checked 2026-08-02). Whether the premium is worth it is a tool question — see the Seller Snap review and the hub.

Parameter 2: the competitor filter decides what the rule is even looking at

A repricing rule with no filter treats every competing offer as a threat. In practice a large share of the offers on a crowded listing are not competing for the same buyer at all, and reacting to them transfers margin for nothing.

Conditions worth setting explicitly:

  • Fulfilment method. If you are FBA and the price you are chasing is a merchant-fulfilled offer with a longer handling time, you are matching a price the buyer is not comparing you against on equal terms.
  • Seller rating and feedback count. Amazon states plainly that price alone is not sufficient: asked whether automating price guarantees the placement, its FAQ answers “No. Lowering your price can help you become the Featured Offer, but you also need to meet performance-based requirements in areas like inventory availability, fulfillment, and customer service” (checked 2026-08-02). It follows that an offer from a seller who cannot clear those requirements is not a threat worth pricing against — though how far below your own metrics the cut-off should sit is a judgement call, not a published threshold.
  • Condition. Used and refurbished offers on a listing where you sell new should normally be excluded.
  • Stock depth. An offer that will be gone in two days can drag your price down for two weeks if the rule reacts to it and nothing pushes the price back up.
  • Your own other offers. On listings where you hold more than one SKU or sell in more than one condition, the filter has to exclude you, or the rule ends up competing with itself.

There is also a reference-point choice hiding in the filter. Amazon’s rule types offer three different anchors — the Featured Offer price, the lowest price on the listing, or the lowest external price — and the three behave very differently on a listing where the cheapest offer never wins the placement. Anchoring to the Featured Offer means you track the price that is actually converting; anchoring to the lowest price means you track whoever is most desperate.

The public complaint record is a useful reminder that filters are not merely theoretical. One Repricer.com reviewer on Trustpilot (2025-12-11) reported: “I spent over three hours trying to get the Cross ASIN repricing feature to work, and it simply doesn’t. The function is advertised, but it’s not functional in practice.” Whatever the resolution in that particular case, the lesson generalises: confirm during the trial that the filter conditions you depend on actually fire, rather than assuming a checkbox in a feature table works.

Parameter 3: cadence is a real setting, not a vendor benchmark

Speed is the most heavily marketed dimension of repricing and the one most often over-bought. What actually matters is the relationship between how fast prices move on your listings and how fast your rule is permitted to answer.

Two published data points, both checked 2026-08-02:

  • Amazon’s own tool: for existing rules, “price updates are usually processed in less than 15 minutes, but may take longer under certain circumstances”; for a rule you have just changed, “it can take up to one hour for all products assigned to the rule to re-price according to the new parameters” (sell.amazon.com/tools/automate-pricing, checked 2026-08-02).
  • BQool describes its speed as event-driven rather than interval-driven: “Instant repricing refers to price adjustments that occur immediately after BQool’s system receives an SQS (Simple Queue Service) notification from Amazon” (bqool.com/price/repricing-central/, checked 2026-08-02), with the caveat on the same page that “occasional delays may occur if there is limited activity on the Amazon Marketplace” (see the BQool review for what that means in the entry tier).

Aura markets a “10-Second Repricing Speed” on its homepage (checked 2026-08-02). Whether ten seconds beats fifteen minutes for your catalogue is an empirical question about your listings, not a general truth: on an ASIN where three FBA sellers rotate the placement hourly, latency is decisive; on a private-label ASIN with no competing offers, it changes nothing at all. The honest framing is that repricing speed is a ceiling on responsiveness, and buying a higher ceiling than your listings need is a common way to overspend.

Cadence also has a second, slower dial that almost nobody sets: how often you change the rules. Amazon’s own hour-long propagation window for parameter changes is a hint — a rule you keep editing never produces a clean read. Pick a review interval, hold the parameters still between reviews, and change one thing at a time.

The Buy Box slice: what a pricing rule can and cannot do

This is worth stating precisely, because it is where repricing advice tends to overclaim. Amazon publishes eligibility factors for the Featured Offer and states that price is one input among several; it does not publish how those inputs are weighted, and it explicitly denies that price alone decides the outcome (see the FAQ answer quoted above). Everything beyond that — how much a fulfilment upgrade is worth in cents, how quickly a metrics dip is forgiven — is inference, including the inferences in this article.

What follows from the published position, what is only widely observed, and what does not follow:

  • Follows: a repricing rule can only move one of the inputs. If your account metrics or inventory availability are the constraint, a more aggressive floor buys nothing and costs margin.
  • Widely observed, not published: sellers commonly report that the placement rotates between eligible offers. That is a third-party observation, not an Amazon statement — neither the Automate Pricing page nor Amazon’s Featured Offer explainer describes a rotation rule either way. Where it does hold on your listings, a short observation window will misattribute a rotation to your rule change.
  • Does not follow: that undercutting by a cent is a reliable mechanism. That is a widely repeated inference, not an Amazon statement, and on a listing where several sellers run repricers it mostly produces a downward ratchet.

For the eligibility side of the question — who qualifies, what disqualifies an offer, how the placement is presented — how the Featured Offer is decided covers the mechanics in full, and our earlier Buy Box guide is the older long-form treatment on this site.

Three parameter sets, by cost and inventory structure

Seller category is a poor predictor of the right parameters. Cost structure and turnover are better ones, because they change which of the three dials carries the weight.

High-turnover, thin-margin inventory. The floor is close to the ceiling by definition, so precision in the floor calculation matters more than anywhere else — a two-percent error is most of the margin. The filter should be tight, because chasing an irrelevant offer costs a disproportionate share of the contribution. Cadence should be as fast as the tool allows, since the whole strategy rests on holding the placement at a price a few cents from break-even. Event-driven repricers and the AI tiers of rule-based ones are the class of tool built for this.

Long-tail, high-margin inventory. Here the band is wide, and the ceiling is the parameter that pays. A rule that only ever moves downward leaves money on ASINs where competition thins out seasonally. Cadence can be slow; a rule that reprices every ten seconds on an ASIN that sells four units a week is optimising noise. Rule-based engines with good ceiling logic are usually sufficient — RepricerExpress , now part of Repricer.com, is the long-standing example of that class.

Aged inventory being cleared. This is the one case where the floor should be allowed below your target margin, and the parameter that matters is the exit condition: a date, a units-remaining threshold, or a storage-fee trigger that returns the rule to normal. Amazon’s own “Sales-based rules” — “Change your price based on your sales volume in a specified period of time” (checked 2026-08-02) — map onto exactly this shape, and are free with a Professional plan. Time-box it explicitly; a clearance floor left in place after the inventory clears is the most expensive kind of forgotten setting.

Testing an Amazon repricing strategy before trusting it

Repricing changes are unusually hard to evaluate because the market moves underneath the test. A workable protocol:

  1. Split the catalogue. Hold a control group of comparable ASINs on the old parameters. Without it, a seasonal shift reads as a strategy win.
  2. Change one parameter per cycle. Floor, filter, cadence — one at a time. Combined changes produce results you cannot attribute.
  3. Run past one full rotation cycle. On contested listings, days are noise. Two to four weeks is a more defensible window, and longer for slow-moving ASINs.
  4. Record contribution margin, not just placement share. Placement share can be bought at any price; that is the failure mode the floor exists to prevent.
  5. Verify the features you are relying on during the trial. Most vendors offer a window for exactly this — Repricer.com’s page states “14-day free trial. No credit card required.” and BQool’s shows “14 DAY FREE TRIAL” (both checked 2026-08-02). Use it to test the specific filter and floor behaviour you plan to depend on, not the demo path.

Then leave it alone between reviews. The most common way a well-designed set of repricing parameters degrades is continuous adjustment that never lets any version of the rule produce a readable result.