How to Choose a Repricing Tool: A Feature Checklist

Every repricing tool's website says it's fast, protects your margin, and wins the Buy Box. They can't all be right. The problem isn't that you don't know what to look for, it's that the marketing page for a brilliant tool and a mediocre one are almost identical, and you can't tell which is which until you've been paying for three months.

This page is two things in one. A scoring sheet with a verification step for each feature that actually decides your outcome, because a claim you haven't tested is just a claim, and a method for testing your final choice properly, in Safe Mode, before it changes a single live price. Together they cover the whole decision: what to look for, how to weigh it against what you actually sell, and how to confirm it works before you trust it with your catalogue. New to automated pricing? Start with what repricing is and the repricing basics page first.

TL;DR

Eleven items decide whether a repricer makes you money or just keeps you busy, each with the question to ask the vendor and the test to run during the trial. Score your shortlist out of 33. Anything scoring low on floors or speed is out, regardless of what else it does well. Once you've picked a tool, don't switch it on live: run it in Safe Mode against real market data for a week or two first, compare the simulated result to what you're actually getting, and only enable live pricing once the simulation shows it's an improvement.

The Checklist

Score each item 0 to 3 for every tool you're considering: 0 means it doesn't have the feature, 1 means the vendor claims it but you couldn't verify it, 2 means verified and works adequately, 3 means verified and genuinely good. 33 points are available across 11 items. The first four are pass or fail: a tool scoring 0 or 1 on any of them is off your list no matter how it scores elsewhere.

Print the table below, or copy it into a spreadsheet if you're comparing more than two or three tools, and fill in a column per vendor as you go through demos and trials rather than trying to remember the details afterwards. Memory favours whoever gave the nicest sales pitch, not whoever actually has the better product, and the whole point of scoring against a fixed sheet is to take that bias out of the decision.

Scoring: roughly 29 or above is a strong fit. 22 to 28 is workable, with known gaps. Below 22, keep looking. Any 0 or 1 in items 1 to 4 is disqualifying regardless of total. (These bands are the original 9-item scoring proportionally scaled to 33 points, Mark to sanity-check before publish, see the change note.)

Now the reasoning behind each, because a checklist you don't understand is just homework.

Book a Demo and see how Repricer.com scores against your own checklist before you commit to a trial.

Items 1 to 4: The Pass-or-Fail Four

1. Net-margin floors

The single most important feature, and the one most tools fudge. A basic tool lets you type a minimum price per SKU. That number is frozen the moment you type it. It doesn't know Amazon raised FBA fees in January, that your freight went up, or that your return rate crept from 2% to 4%. It's a floor that was correct once.

A proper tool calculates the floor: landed cost plus referral fee plus fulfilment cost plus returns provision plus ad allocation plus your target margin. When an input moves, the floor moves. Take a product costing $8.50 landed, with a 3.18FBAfee,an8%referralrateanda20%targetmargin:thecorrectflooris(8.50 + $3.18) ÷ (1 − 0.08 − 0.20), which comes out at $16.22. A tool with a typed floor keeps that number fixed until someone remembers to redo the sum by hand. A tool with a calculated floor moves it the moment any input changes, without anyone having to notice first.

The verification that matters: change a cost input and see whether the floor recalculates. If you have to update floors by hand when Amazon's fees change, the tool doesn't really have net-margin floors, it has a spreadsheet with extra steps. Our net margin guide covers what belongs in the calculation, and Net Margin Repricing works off net position after fees.

2. Repricing speed

Ask for a number, then don't believe it. Every vendor says "real-time". The word is meaningless without a figure, and the figure is meaningless without a test. Get the number, then open a listing you're on, watch a competitor move, and time how long your price takes to follow. That single test tells you more than the entire features page.

Context for what you're measuring: Amazon's own free Automate Pricing tool documents price updates as usually processing in under 15 minutes, and up to an hour after a rule change. That's your baseline. A paid tool that isn't dramatically faster than the free one isn't worth paying for. The gap compounds too: a Buy Box window that opens because a competitor goes out of stock might last minutes, not hours, and a tool working on a 15-minute cycle simply isn't in the race for it. Repricer.com reprices in seconds, which is where it built its name as the fastest Amazon repricer to win the Buy Box, and the cost of a slow repricer covers what the gap costs.

3. Upward repricing

The half nobody checks. Plenty of tools are competent at moving your price down and hopeless at moving it back up, which means you've automated your losses and left your gains manual, and you won't notice for months because falling prices look like the tool working.

Verify it: run the trial for two weeks and look at your price history. If every movement is downward, the tool is a discounting engine, not a repricer. Margin hides in the upward half, in the windows where a competitor goes out of stock or drops out of the rotation and nobody's forcing your price down anymore. A tool that can't price up in that window is leaving money that was genuinely available.

4. Competitor filtering

Twenty minutes of setup, and it's the difference between a repricer and a liability. Not every seller on your listing deserves a response. Someone with poor metrics, a different condition, no Buy Box eligibility, or a fire sale on dead stock can drag your price down while never actually threatening your rotation. If your tool can't ignore them, it will follow them, and our price war guide covers where that ends.

Verify it: exclude a specific seller, then confirm your price genuinely stops responding to them. Some tools offer the setting and apply it loosely, letting the excluded seller back in the moment their price shifts slightly, which defeats the point of setting the filter at all.

Three or more of items 1 to 4 scoring low is a reasonable trigger to walk away from a shortlist candidate entirely, regardless of how well it scores on items 5 to 11. These four aren't nice-to-haves stacked at the top of the list for emphasis, they're the four that determine whether the other seven matter at all.

Items 5 to 11: The Ones That Decide How It Feels to Own

5. Rule and AI logic together

Not one or the other. Rules give you explicit control on the SKUs that need it (MAP-restricted lines, flagship products, anything with a story). AI handles pattern recognition across the long tail where you'd never write rules anyway, because nobody sits down and writes 4,000 individual pricing rules for a catalogue that size. Tools that force the choice make you pick which half of your catalogue to underserve, either hand-tuning a few dozen SKUs while the rest run on autopilot with no real logic behind them, or letting an AI model loose on flagship products where a wrong price for even a day is a real cost. The rule-based vs AI breakdown covers which segments want which, and most sellers who've done this for a while end up running both at once rather than picking a side.

6. Bulk controls

Fine at 50 SKUs. Decisive at 5,000. The test is simple: ask them to show you applying a strategy to 500 SKUs at once, and time it. If the answer involves a CSV round-trip and a support ticket, imagine doing that every time your costs change, every time Amazon updates a fee schedule, every time you add a new product line. A tool that handles this in a few clicks turns a recurring chore into a non-event; one that doesn't turns every cost change into a project.

7. Reporting you'd act on

Most tools show you price changes. That's a log, not a report. What you need is Buy Box win rate against profit per SKU, in one view. Win rate alone is a vanity metric, you can buy it by pricing at your floor all day and watching the percentage climb while your margin quietly disappears. Seeing both together is what tells you whether the tool is working or just busy, and it's the difference between a dashboard you glance at occasionally and one you actually use to decide what to fix next. Analytics and reporting pairs them deliberately, and the Buy Box win rate guide covers how to read the numbers and turn a percentage into an actual figure worth acting on.

8. Safety behaviour

The question nobody asks until it's 3am on Black Friday. What does the tool do when Amazon's API drops mid-session? Does it hold your last price, revert to a default, or keep firing updates into the void based on stale data? Ask for the documented failure mode, in writing, not a verbal reassurance on a sales call. A vendor who can't answer this crisply hasn't thought about it, and you'll find out when it matters, usually during the highest-traffic weekend of the year, which is exactly when an API hiccup is most likely and most expensive.

9. Onboarding

Underrated, and the difference between value in a fortnight and value in a quarter. The software setup is under an hour. Costing your SKUs properly and writing sensible rules across thousands of them is the part that never gets done, because it's tedious and nobody on a small team has a free afternoon for it. Ask directly: who configures my first 500 SKUs, and when, and is that person doing it with me or handing me a knowledge-base article and wishing me luck? Managed setup being included rather than an upsell is a meaningful signal about how the vendor sees the relationship, whether they want you live and profitable quickly or just signed up.

10. Buy Box prediction

A tool that only reacts to price changes leaves you finding out you lost the Buy Box after it's already gone. A tool with Buy Box prediction flags a listing that's about to lose rotation before the sale disappears with it. The mechanism worth asking about is a scoring model that reads the same signals Amazon's own algorithm weighs, feedback count and percentage, fulfilment method, dispatch time, and where the item ships from, and uses them to flag which of your SKUs are likely to win or lose the Buy Box against their current competitive set, not just which ones already have.

What to ask: "Does the tool flag which SKUs are likely to lose the Buy Box before it happens, and what does it recommend I do about it?" A vendor who can only show you a historical win-rate chart is describing analytics, not prediction, those are different things and only one of them gives you a head start. Analytics tells you what already happened. Prediction gives you a window to act before it does, whether that's tightening a floor, adjusting a rule, or simply knowing which handful of SKUs are worth a closer look this week rather than treating the whole catalogue as equally at risk.

Verify it on trial: pick a SKU the tool flags as at risk, and check whether the flag was right over the following few days. The Buy Box win rate guide covers how to read win rate once you're tracking it, and Repricer.com's own Buy Box Predictor is built around exactly this scoring approach.

11. Amazon country marketplace coverage

If you sell on more than one Amazon marketplace, a tool that only really understands .com is quietly costing you money on every other one. VAT, referral fee percentages, competitive density, and seasonal timing all differ by marketplace, and a copied .com rule set gets all four wrong at once. A floor calculated in USD and applied to a VAT-inclusive market like Amazon.co.uk or Amazon.de without grossing up for VAT looks profitable and isn't, because a chunk of the price a buyer pays is tax you owe, not margin you keep. The competitive set is different too: the sellers competing for the Buy Box on Amazon.de aren't the same population as on Amazon.com, and a rule tuned for one routinely misreads the other.

What to ask: "Does one rule engine cover every marketplace I'm in, with its own floor, ceiling and competitive set, or am I managing separate setups?" Verify it on trial: ask to see the configuration for a second marketplace and confirm it isn't your .com rules copied over with a currency symbol changed. If you're extending into a new marketplace for the first time, the same Safe Mode principle covered later on this page applies there too, run the new marketplace's rules in simulation for a week or so before going live, since a marketplace you've never sold on has different competitive event patterns you won't know in advance. Repricer.com covers all 21 Amazon country marketplaces from one account, with marketplace-specific floors, ceilings, and competitive set filters per region, and repricing across Amazon's European and international marketplaces covers the VAT and competitive-set adjustments in detail.

How the Features Stack Up

Here's all 11 items side by side: what each one protects against, the question that surfaces a weak answer, and where Repricer.com lands on it. Use it as a working document rather than a reference table, copy the four columns into your own shortlist comparison and fill in a real vendor's answer against each row, then you'll actually see the gaps rather than just reading about where they tend to be.

Read the table by row, not by total. A tool that scores well on eight of eleven but fails items 1 and 2 outright is worse for you than one that scores adequately across the board, because the two pass-or-fail items are the ones that determine whether every other row matters. If you're filling this in against a real vendor call, write down their actual answer in the third column rather than what the sales page implies, the gap between the two is usually where the real decision lives.

Nice to Have, Not Deal-Breakers

Don't let these swing your decision.

• Mobile app. Useful for a glance on the train. Nobody configures a repricer on a phone, and if you need to, something's wrong with how the rest of the setup was done.

• Pretty dashboards. Charts are not insight. A tool with beautiful visualisations and weak floor logic is a worse tool than a plain dashboard sitting on top of floors that are actually correct.

• Alerts. Genuinely handy, until there are forty of them and you've muted the channel entirely. Look for configurable, not abundant, since an alert nobody reads isn't a safety feature.

• API access. Matters if you have developers and a real integration need, connecting the repricer to a custom inventory system or a BI tool. Otherwise it's a line item you'll never use, and paying extra for it buys nothing.

The pattern: features that make a demo impressive are rarely the features that make a repricer profitable. A vendor who spends the first ten minutes of a call on the dashboard's colour scheme rather than how the floor is calculated is telling you something about where their own priorities sit.

What to Prioritise by Seller Type

The eleven items don't weigh equally for everyone, and international coverage only matters if you actually sell across borders.

If you sell across Amazon's international marketplaces, marketplace coverage (item 11) stops being a nice-to-have and becomes close to a fifth pass-or-fail item: a rule engine that doesn't handle VAT, currency, and per-marketplace competitive sets separately will quietly misprice every market that isn't your home one. The seller types above aren't fixed categories either, most catalogues are a mix (a wholesale core with a handful of private-label lines, say), so the honest way to use this table is to weight each item against the actual share of revenue each segment represents in your own business, not against whichever label feels closest.

How to Run the Trial Properly: Test in Safe Mode Before Going Live

Repricing configuration errors are silent, they don't trigger an alert in Seller Central, they just show up later as a margin that's quietly narrowed. Testing before you go live is how you catch that before it costs anything.

What Safe Mode Is and What It Shows You

Safe Mode is Repricer.com's simulation environment. It runs your complete repricing configuration, every rule, floor, ceiling, and competitive set filter, against real Amazon marketplace data, without submitting a single price change to Amazon. When Safe Mode is active on an ASIN, Repricer.com observes the real-time competitive data on the listing, evaluates your configured rules against that data as if they were live, and records what price the rule would have set and what the simulated Buy Box outcome would have been. It then compares this to what actually happened on the listing, your real price, your real Buy Box share. Amazon never receives a price update during the test, so it can run indefinitely without any risk to your live listings.

Safe Mode produces three comparison metrics for each ASIN: simulated average selling price versus your actual ASP, the primary measure, where a higher simulated ASP at a comparable win rate means the configured rules produce better margin outcomes than what you're currently getting; simulated Buy Box win rate versus your actual win rate, where a simulated rate more than 15 points below your actual rate at the same or lower ASP suggests the floors are set too conservatively, holding the price too high for the competitive range rather than too low; and floor breach events, which record every moment the rule tried to set a price below your minimum and was stopped, a useful diagnostic on whether the floor itself is set correctly or is out of step with where the market actually sits. None of the three tells the whole story on its own, which is why the evaluation step below asks for all three read together rather than any single number treated as a verdict.

The Three-Step Safe Mode Test

Safe Mode testing takes 7 to 14 days. Three steps get you through it.

1. Configure your rules in Safe Mode. Set the floor from landed cost plus FBA fee plus referral fee plus target margin, verify it sits above your actual break-even. Set the ceiling from the 90-day historical high on that listing. Choose the rule type for the listing's competitive density, and set the competitive set filter (fulfilment method, feedback threshold, stock minimum). Enable Safe Mode on the ASIN once configured, the simulation starts recording immediately.

2. Run the simulation for 7 to 14 days. High-velocity ASINs with 10 or more competitive price events a day generate enough data in 7 days. Slower-moving ASINs need the full 14 to produce a comparison worth trusting. Your actual live prices stay untouched the entire time.

3. Evaluate the results. Pull the Safe Mode report per ASIN and compare simulated ASP to actual ASP, simulated win rate to actual win rate, and the floor breach count. Mark each ASIN improve, neutral, or investigate. Don't enable live pricing on anything that scores worse in simulation than what you're already getting.

How to Read the Results

A positive result looks like a higher simulated ASP at a comparable or only slightly lower win rate, for example a simulated ASP of $23.10 against an actual $20.85, at simulated win rate three points below actual. At a typical margin structure, that ASP gain outweighs the small win-rate cost, and the net effect is more margin per day even though you're winning slightly fewer sessions.

A negative result is the one worth understanding properly: simulated ASP lower than your actual ASP. This means the configured rules would have produced lower prices than what you're currently getting, without enough of a win-rate improvement to justify it, and it's a sign the configuration is chasing Buy Box share at the expense of margin rather than optimising for profit. Don't enable rules that show this. Check two things first: whether the ceiling is set below the current market price, which forces the rule to drop toward it unnecessarily, and whether the rule type is an undercut rule on a crowded listing, which spirals prices down as every automated competitor responds to every other one. Fix whichever applies, then rerun the simulation before touching anything live.

The floor breach count tells you a third thing on its own. A breach price close to your floor means the floor is set correctly and the rule was rightly stopped from going lower. A breach significantly above your floor means your competitive set filter is too loose, letting a seller who isn't genuine competition trigger a response that drives the price down for no reason. No breaches at all across the test period is itself useful information: it means the competitive field stayed above your floor the whole time, which confirms the floor is correctly set for that listing's current dynamics, though it's worth re-running the check periodically since competitive conditions on a listing rarely stay fixed for long.

Worth doing this evaluation per ASIN rather than as a single average across the whole test group. An average that looks positive can hide a handful of ASINs where the simulation performed badly, and those are exactly the ones that need a rule review before going anywhere near a live price. Sort by the ASINs where simulated ASP came in lowest relative to actual, and start your investigation there.

Going Live and the First Week

Enable live repricing per ASIN, not catalogue-wide, starting with the 5 to 10 ASINs where Safe Mode showed the clearest improvement. Run those live for 7 days, then compare the actual Buy Box win rate and ASP against the Safe Mode simulated values. A match within 10% confirms the simulation was accurate and you can extend live pricing to the next group, repeating the same 7-day check each time rather than switching on the rest of the catalogue in one move. A gap larger than 10% usually means competitive conditions shifted since the test ended, a new seller entering the listing or a seasonal demand shift, or the configuration behaves differently under real conditions than in simulation, either way it's worth investigating before scaling further rather than assuming it will settle on its own. Keep any "investigate" ASINs in Safe Mode until a rule change fixes the issue and a fresh simulation confirms it, there's no cost to leaving something in simulation for longer while you work out what's wrong with it.

The first week needs daily attention, and it's worth treating that as non-negotiable rather than something to get to when there's time. A five-minute daily check confirms no ASIN has triggered an out-of-bounds alert (a price move more than 20% from the previous price in one event), that win rate sits within five points of the Safe Mode baseline, and that no live ASIN has dropped to zero Buy Box share, which usually signals a floor set above the competitive range rather than anything wrong with the rule itself, worth checking before assuming the worst. A mid-week check going a level deeper, pulling the analytics dashboard for the live-enabled ASINs and looking at whether ASP is stable, rising or declining against the Safe Mode baseline, and whether any ASIN is consistently selling within fifty cents of its floor, which signals the competitive field has settled at the floor and the rule isn't ceiling-hunting when it should be. At the end of week one, a fuller review calculating profit per Buy Box session for each live ASIN, daily margin divided by daily Buy Box sessions held, confirms whether the configuration is actually optimising for margin, not just for win rate. Most of the genuinely common live-repricing problems, a floor sitting too close to the competitive range or a rule type spiralling on a crowded listing, show up in this first week if they're going to show up at all, which is the whole argument for checking daily rather than weekly at the start.

Run the trial itself on your actual catalogue, not demo data: twenty to fifty representative SKUs including a contested line, a slow mover, and a healthy-margin product, with your floors fixed before you start so a bad floor doesn't get blamed on the rules. Give it a fortnight rather than a few days, since Buy Box data is noisy over shorter windows and a few days is often just long enough to misread a competitor's temporary stock-out as a verdict on the tool. Score what you see as you go rather than trusting memory afterwards, filling the checklist table in during the trial rather than reconstructing it from impressions a week later, and test one tool at a time on the same SKUs so you can actually tell what caused what, running two repricers against the same products at once makes every result impossible to attribute cleanly to either one.

Book a Demo and score a tool built around fee-aware floors, seconds-fast repricing, and a Safe Mode that shows you the result before it's live.

FAQ

1. What's the most important feature in a repricer? Net-margin floors, and it isn't close. A floor calculated from landed cost plus fees plus target margin updates when your costs move; a typed minimum price is frozen the day you entered it. Every other feature builds on top of that, because speed and Buy Box targeting only help if the wins are profitable.

2. How do I check a repricer's speed claim? Don't take the number from the website. Open a listing you're active on, watch for a competitor to change price, and time how long yours takes to follow. Amazon's own free tool documents updates as usually under 15 minutes, so that's the baseline any paid tool should beat comfortably.

3. How much should I budget for repricing software? Entry-level tools start around $25 to $100 a month, mid-tier sits nearer $100 to $300, and enterprise plans are usually quote-based. Most vendors offer a trial of a fortnight or so. Judge it against what you currently lose to slow pricing and stale floors, not against zero, the free option has a cost too, it just doesn't appear on an invoice.

4. What is Safe Mode in a repricer? Safe Mode is Repricer.com's simulation environment. It runs your complete repricing configuration, floors, ceilings, rule types, competitive set filters, against real-time Amazon competitive data, without changing any live prices. You compare the simulated ASP and win rate to your actual outcomes, then enable the rules only once the simulation confirms they're an improvement.

5. What happens if my Safe Mode results show lower simulated ASP than my actual ASP? It means the configured rules would have produced lower prices without enough win-rate improvement to justify it. Don't enable them live. Check whether the ceiling sits below the current market price, or whether an undercut rule on a crowded listing is spiralling prices down, fix whichever applies, and rerun the simulation before going live.

Where to Start

Take your current tool and score it out of 33 before you look at anyone else's. Most sellers discover they're already paying for something scoring roughly half the available points, and that the gap sits in items 1 and 2, the two that decide whether repricing makes you money or just makes you busy. It's worth doing this even if you're happy with your current tool, since a checklist you score honestly against your own setup sometimes surfaces a gap you'd stopped noticing because it had been there long enough to feel normal.

If it scores well, you've saved yourself a migration, and you can stop reading here with a clear answer instead of a vague feeling that things are probably fine. If it doesn't, you now know exactly what to ask the next vendor, exactly which of the eleven items to press them on, and how to test whatever you choose in Safe Mode before it touches a single live price.

Book a Demo and score a tool built around fee-aware floors and repricing in seconds.