TL;DR
Repricing automation works when the setup is right, and the setup is mostly done before you touch the software. Work out your true cost per SKU, build your floor from that number rather than typing one in, start with a test group instead of the whole catalogue, then leave it alone for a fortnight before judging it. Most repricing failures aren’t tool failures. They’re a floor that didn’t know about fees and a seller who panicked on day two.
Most guides on repricing automation explain what it is and stop there. You already know what it is. What you want is the sequence: what to do first, what to set, what to watch, and when to stop fiddling.
So here’s the actual walkthrough. Eight steps, in order, with the mistake that lurks at each one. Follow it and your first fortnight will be boring, which is exactly what you want from a repricer.
New to the concept entirely? Start with what repricing is, then come back. Still deciding whether to automate at all? The manual vs automated comparison covers where the line sits.
Before you start: the one thing that decides everything
Your floor is the whole game.
Nearly every repricing horror story (the race to the bottom, the Q4 that sold out at a loss, the SKU that bled for a month) traces back to a minimum price that was wrong. Not to the algorithm. To the number underneath it.
So resist the urge to connect your account first. Get your costs right, and the rest of the setup is straightforward.
The setup, step by step
| Step | What you do | Time | The mistake to avoid |
| 1 | Cost every SKU properly | Longest step | Forgetting returns and ad spend |
| 2 | Build your floor from that number | 30 min | Typing a flat number instead |
| 3 | Set a ceiling | 20 min | Leaving it off entirely |
| 4 | Connect Amazon, pick a test group | 15 min | Switching on the whole catalogue |
| 5 | Choose a strategy per segment | 30 min | One rule for everything |
| 6 | Filter your competitors | 20 min | Matching sellers you shouldn’t |
| 7 | Switch on, watch 48 hours | 2 days | Judging it on day two |
| 8 | Read the metrics, then scale | 2 weeks | Watching price instead of profit |
Step 1: Cost every SKU properly
Before any software, you need a real number for what each unit costs you to sell. Not what you paid for it. What it costs to sell it.
That means landed cost (unit price plus freight plus duty plus prep), plus Amazon’s referral fee, plus your FBA or fulfilment cost, plus a returns provision based on your actual return rate, plus an ad allocation if that SKU is running PPC. Our net margin guide covers what belongs in the calculation, and the seller fees guide covers what Amazon takes.
The mistake: stopping at product cost plus referral fee. Returns and ad spend are the two everyone forgets, and they’re often the two that turn a “profitable” SKU into a loss-maker.
This is the longest step, and it’s the one people rush. Don’t.
Step 2: Build your floor from that number
Now your floor writes itself: total cost per unit, plus the margin you actually want.
The important part is that the floor should be calculated, not typed. A typed floor is frozen the day you enter it. It doesn’t know Amazon changed FBA fees in January. It doesn’t know your freight went up. A floor built from your cost inputs moves when they move, which is why Repricer’s minimum price floors work off net position after fees rather than a flat number.
The mistake: a round number that felt safe. “£24.99 sounds about right” is not a floor, it’s a guess wearing a floor’s clothes.
Step 3: Set a ceiling
Less glamorous, more valuable than people expect.
Your ceiling is the highest price you’ll charge. Set it from what the market actually bears, not from optimism. It matters because a good repricer prices up when you hold the Buy Box and demand supports it, and without a ceiling it has nowhere to aim.
The mistake: skipping it. If you only ever configure the downward half, you’ve automated your losses and left your gains manual.
Step 4: Connect Amazon and pick a test group
Now you can touch the software. Connect your Amazon account via API, let your listings sync, and then resist the obvious temptation.
Do not switch on the whole catalogue.
Pick a representative test group instead: 20 to 50 SKUs that look like your business. Include a competitive line, a slow mover, and something with healthy margin. That mix tells you more in a week than 3,000 SKUs will.
The mistake: going all-in on day one. When something looks wrong across 3,000 SKUs, you have no idea which rule caused it. With 30, you do.
Step 5: Choose a strategy per segment
Different SKUs need different logic. Your thin-margin wholesale line and your private-label hero should not share a rule.
Broadly:
- Contested wholesale lines want Buy Box-focused logic with a firm floor.
- Slow movers with storage cost want velocity logic that eases price down to clear stock before surcharges land.
- Uncontested or private-label SKUs want ceiling-hunting logic that tests upward rather than defending downward.
- Seasonal SKUs want rules that reflect the curve, and our seasonal repricing guide covers the patterns.
The repricing strategies page walks through the common ones, and the rule-based vs AI breakdown covers when to hand a segment to the algorithm and when to keep the rules explicit. If you’re self-fulfilling, FBM repricing strategies covers what changes without the Prime badge.
The mistake: one rule for the entire catalogue. It’s the fastest setup and the worst outcome.
Step 6: Filter your competitors
This step takes twenty minutes and saves more margin than any other single setting.
Not every seller on your listing deserves a response. Someone with poor metrics, a different product condition, no Buy Box eligibility, or a fire sale on dead stock can drag your price down while never actually threatening your rotation. Exclude them.
Price wars usually start because somebody matched a competitor they should have ignored. Our price war guide covers the patterns that trigger spirals, and understanding the Buy Box algorithm tells you which sellers Amazon actually takes seriously.
The mistake: matching everyone by default, then wondering why your price keeps falling.
Step 7: Switch on and watch the first 48 hours
Turn it on. Then watch, but don’t touch.
The first two days are about checking nothing weird is happening at the edges: suppressed listings, locked variants, SKUs you meant to exclude, floors that look wrong when you see them live. That’s a sense-check, not a performance review.
Read your floors out loud once before you go live. A misplaced decimal is the cheapest expensive mistake in this business.
The mistake: treating day-two noise as a verdict.
Step 8: Read the right metrics, then scale
After the fortnight, look at four numbers:
- Buy Box percentage. Usually the first to move, because the tool reacts faster than you did.
- Profit per unit. The one that matters. If Buy Box is up and profit per unit is down, your floor is wrong.
- Sales velocity. Tells you whether the extra rotation is converting.
- Price change frequency. If it’s thrashing, your rules are fighting each other.
Analytics and reporting is built around win rate and profit impact rather than a raw log of price changes, which is the difference between data you act on and data you scroll past.
Once the test group behaves, expand in stages rather than all at once. And when Amazon next changes a fee, revisit step 1. Costs move; floors should move with them.
The mistake: watching whether your price moved instead of whether your profit did.
How long until it’s working?
Roughly:
- 48 hours: you’ll know nothing is broken.
- 1 to 2 weeks: Buy Box percentage starts to move.
- 2 to 4 weeks: the pattern is clean enough to judge.
- Ongoing: revisit whenever fees or costs change.
In Repricer’s own customer survey, 98% of sellers won more Buy Boxes after switching and 90% said the rules helped protect their margins. Your own numbers depend on your catalogue and category, but the direction is consistent. Repricer’s sales analysis covers how performance developed for customers over a longer run.
What automation won’t fix
Worth saying plainly before you start.
- A sourcing problem. If your landed cost leaves no margin, automation just finds the floor faster.
- A failing account. Order Defect Rate and account health sit upstream of price.
- A bad listing. Repricing gets you into the rotation. Images and reviews still do the converting.
- An empty SKU. The Buy Box doesn’t go to zero stock.
Repricer’s AI Buy Box optimizer targets the box on more than price, and the Buy Box Predictor uses competitor stock, fulfilment method and feedback score to call outcomes before they happen. Neither fixes the four things above.
If configuration is the part you’d rather hand over, free managed setup puts someone who does this daily on steps 1 through 6 with you. And if you want the wider argument for automating at all, the benefits beyond price piece covers what changes past the pricing itself.
FAQ
How long does it take to set up repricing automation? The software part takes under an hour: connect your Amazon account, set your floors and ceilings, choose a strategy, select your SKUs. The part that takes real time is step one, costing your SKUs properly, and that’s the step that decides whether any of the rest works. Budget an afternoon for a small catalogue.
Should I automate my whole catalogue at once? No. Start with 20 to 50 representative SKUs including a competitive line, a slow mover, and a healthy-margin product. Confirm the rules behave, then expand in stages. Switching on 3,000 SKUs at once means you can’t diagnose anything when it looks wrong.
How long before I can tell if it’s working? Give it two weeks minimum, four to be confident. Buy Box percentage usually moves first, within one to two weeks. Judging it after 48 hours is the most common setup mistake, and it’s why sellers yank settings before any pattern emerges.
What’s the most common repricing automation mistake? A flat minimum price instead of a calculated one. A typed floor doesn’t know Amazon changed its fees, so it quietly stops protecting your margin the moment your costs move. Build the floor from landed cost plus fees plus target margin, and let it update when the inputs do.
Can I set different rules for different products? Yes, and you should. Contested wholesale lines, slow movers with storage exposure, and uncontested private-label SKUs all want different logic. One rule for the whole catalogue is the fastest setup and the worst result.
Does repricing automation reduce my profits? Only when the floor is wrong. Your minimum price is a hard limit the tool won’t cross, so a floor calculated from real costs means automation protects margin more reliably than manual pricing, which only applies a floor when you remember to check.
Where to start
Don’t open the software today. Open a spreadsheet and cost ten SKUs properly, including returns and ad spend. If those numbers surprise you (and they usually do), you’ve just found the reason your last repricing attempt didn’t work.
Get the floor right, test on thirty SKUs, and leave it alone for a fortnight. That sequence alone puts you ahead of most sellers who automate.
If you’d rather have someone walk your catalogue through steps 1 to 6 with you, see what the setup looks like on your own listings.
