Compare repricers on decision logic and floors, and order routers on marketplace coverage and failure handling. First decide whether you need one at all: repricers exist for resellers competing in shared buy boxes, while private-label brands own their listings and rarely benefit. Routers earn their fee once orders flow from several marketplaces or warehouses.
The short version
- Name your business model before shopping. Reseller and private-label operations need opposite things from pricing software.
- A repricer is only as safe as its floor. Cost-plus-fees floors per marketplace and currency are the whole game.
- A router is only as good as its failure alerts. Silent routing errors become oversells and suspensions.
- No vendor names or rankings here. Software changes monthly; the ten evaluation items below do not.
- Margin is made before the tool runs. Pricing room is created at product selection and sourcing, and no algorithm manufactures it afterward.
Start with the question the vendors skip
Are you competing for a shared buy box, or do you own your listings? A reseller carrying other people's brands fights dozens of sellers on identical pages, where price moves win or lose the sale hourly. That business needs a repricer the way a trader needs a terminal.
A private-label brand is different. You are the only seller on your listing, the buy box is yours by default, and "repricing" against yourself mostly means eroding your own margin faster. Pricing for a brand is a strategic decision, position, anchoring, promotion windows, made a few times a quarter, not an algorithmic one made hourly. Plenty of brand owners run repricers anyway, because the tool was sold to them, and the subscription quietly automates discounting nobody asked for.
Order routing is the opposite: it becomes necessary for both models the moment orders arrive from several marketplaces and ship from more than one warehouse or fulfillment program. Route by hand across five European marketplaces and a UAE store, and the error rate finds you.
The ten-point comparison checklist
For each item: what done properly means. Score any candidate tool against all ten before a demo impresses you.
- Model fit. Done properly: the vendor asks whether you resell or own your brand before quoting, and says plainly when their tool is wrong for you. A vendor selling hourly repricing to a single-seller listing failed this item in front of you.
- Floor logic. Done properly: floors are computed from landed cost plus marketplace-specific fees plus currency, per marketplace, and the tool refuses to price below them even mid-war. A single global floor in one currency is a margin leak wearing a settings page.
- Decision transparency. Done properly: every price change carries a readable reason, what triggered it, what rule fired, in an exportable log. If support cannot explain yesterday's price, you are not running the tool, it is running you.
- Objective control. Done properly: you can set different goals per product, velocity on launches, margin on mature stock, and the tool holds them independently. One global aggression slider means your whole catalog shares one strategy.
- Safety rails. Done properly: hard min and max prices per SKU, anomaly checks against fee changes and mapping errors, and an instant pause. Ask the vendor what their tool does when Amazon returns bad data. The pause matters more than the algorithm.
- Coverage. Done properly: every marketplace and warehouse you operate today, plus the two you plan next, supported natively rather than through promised roadmap items. Missing coverage discovered post-signature is the oldest trick in software sales.
- Routing rules. Done properly: orders route by explicit, editable rules, which warehouse serves which marketplace at what cost and speed, with stock-aware fallbacks. A black-box "optimal routing" claim you cannot inspect is a rule you cannot fix.
- Failure handling. Done properly: failed or stuck orders alert a human within minutes through a channel someone actually watches, with automatic retry and a manual override. Routing failures are silent by nature, and silence is how oversells happen.
- Reconciliation. Done properly: the router's record of orders, shipments and inventory matches marketplace reports, and exports cleanly for accounting. A router that cannot prove what it did creates work equal to what it saves.
- Exit path. Done properly: your rules, logs and history export in open formats, and the contract lets you leave monthly. Any tool that holds data hostage has priced your switching cost into its roadmap.
What the software companies will not tell you
The margin a repricer protects was created long before the software ran. It comes from picking a market with pricing room and sourcing the product at a cost that leaves some. I have watched this from the factory side: Flapen sources through our own Guangzhou studio, with frameworks built across 500+ brands, and the difference between a product with repricing room and one without is decided in those sourcing negotiations, not in any dashboard. A tool defending three points of margin cannot help the seller who gave away fifteen at the purchase order. That upstream work, market selection and product economics, is the research layer, and ours is public at how we pick products.
The second omission: automation compounds your errors at machine speed. A wrong fee assumption in a floor, replicated across five marketplaces by a diligent algorithm, loses money faster than any manual mistake ever could. Configure slowly, audit monthly.
Related answers
- Alternatives to ChannelAdvisor for Amazon multichannel
- Best tools for managing multiple Amazon marketplaces
- Affordable Amazon multi-market dashboards
- Top Amazon marketplace management platforms ranked
- Amazon marketplaces by geography: the complete guide
If you would rather buy the operating judgment than the software stack, that offer lives at Flapen.
