Reprice your most competitive Amazon listings in 1 second with Overdrive.

Learn More
← Back to blog
AI repricer setup: a seller's step-by-step guide
GuidesAmazonWalmartRepricing

AI repricer setup: a seller's step-by-step guide

Koby Kasnett

Founder and CEO at Informed Repricer


Getting started with AI repricing software can feel like handing the wheel to someone you've just met. The upside is real: Informed Repricer customers report an average 63% increase in Buy Box ownership within their first two weeks. But that result doesn't happen by plugging in your catalog and walking away. It happens because the setup is deliberate, the guardrails are tight, and the pilot is measured.

This guide walks you through exactly that process: connecting the right data, setting margin guardrails before repricing touches a single price, running a 30-60 day pilot, and reading the KPIs that tell you whether it's working.

Who this is for: Amazon and Walmart sellers ready to activate an AI repricer for the first time, or sellers who've tried automated repricing before and want a cleaner, more controlled approach.

Prerequisites: An active Seller Central or Walmart Seller Center account, a basic understanding of your product costs, and roughly 2-3 hours for initial setup.

Difficulty: Moderate. The decisions matter more than the technical steps.

Prepare your data feeds before anything else

The quality of your repricing outcomes is only as good as the data feeding into the machine learning repricer. Before you flip any rules on, work through this checklist.

Inventory feed. Confirm each SKU/ASIN is correctly mapped in your repricer. Check that on-hand quantity and available (sellable) quantity are separate fields, and know how often they sync. If your inventory feed updates every few hours, stale stock data can cause the AI to reprice listings you can't actually fulfill.

Cost data (landed cost). This is the most common onboarding mistake, and it's costly. Your cost input must include item cost plus shipping and fulfillment costs. If you only enter the product cost, your margin floor is calculated incorrectly and the repricer may push prices below profitability. Tools like InventoryLab integrate with Informed Repricer specifically to pull clean landed cost data.

Competitor/market signals. In practice, these are price changes happening around your offer on the same listing. Your AI repricing software monitors those shifts and responds within your defined rules. Informed Repricer's 1-second repricing means competitor moves get matched almost instantly, but only within the boundaries you set.

Validation step. Before activating repricing, pull 10-20 representative SKUs covering different margin levels, sales velocities, and competition intensities. Spot-check that costs are accurate, min/max prices make sense, and no SKU has a missing or zero-cost entry.

Set guardrails before you activate a single rule

AI repricing is not the same as "always be the cheapest." Amazon's Featured Offer (Buy Box) eligibility isn't based on lowest price alone (Amazon Ads Help Center, July 2026). Price is one factor among several, including fulfillment method, seller metrics, and availability. That means racing to the bottom can cost you margin without winning the Buy Box.

Your three non-negotiable guardrails:

  1. Minimum price: the floor below which repricing never goes, regardless of what competitors do.
  2. Maximum price: the ceiling, so upward repricing doesn't trigger a suppressed listing.
  3. Margin/profit floor: a calculated floor based on landed cost plus your target margin percentage. Seller Assistant's repricing floor guidance (August 2026) makes clear that minimum price alone isn't enough because cost structures change. A dynamic margin floor catches those gaps.

The concept to internalize: you compete inside your safe margin band, not below it. If a competitor drops below your floor, you hold. If they raise their price, your repricer should raise yours too. That's what Informed Repricer means by "reprice upwards."

Three rule templates for the pilot

Template 1: Balanced profit + Buy Box. Use this for your core catalog. Compete to match or beat the Buy Box price, but only within your margin band. When the market is volatile or the competitor set is uncertain, hold price rather than chase.

Template 2: Aggressive Buy Box win. Apply this to high-velocity, high-competition SKUs where Buy Box ownership directly drives volume. Narrow your competitor scope to the most relevant offers, hold your margin floor firm, and accept tighter margins in exchange for share. Still protected by the floor.

Template 3: Private label or low-competition. For listings where you're the only or dominant seller, use an optimal price approach. You're not reacting to competitors, you're finding the price that maximizes revenue per unit. Avoid setting this to react to every minor price signal, it'll cause unnecessary movement.

Start with your top 50-200 SKUs by revenue and competition level. Don't try to run your entire catalog through the pilot.

Run a structured 30-day pilot

  1. Select your pilot cohort. Pick 50-200 SKUs spanning different margin tiers and Buy Box difficulty levels. Mix of templates 1, 2, and 3 is ideal.
  2. Connect feeds and run a sync test. Confirm inventory, cost, and competitor data are flowing. Check for missing ASINs or cost fields.
  3. Create your two primary strategies (balanced and aggressive) with margin floors locked in for both.
  4. Configure competitor scope. Decide which sellers you're competing against. A tighter competitor set reduces noise and overreaction.
  5. Enable upward repricing and Smart Max features. This prevents races to the bottom and keeps listings from being suppressed. Informed Repricer includes these controls specifically to protect against that outcome.
  6. Days 1-7: review daily. You're looking for feed health, expected price movement, and any SKU breaching its margin floor. If a SKU is moving when you don't expect it to, investigate the rule match before changing anything else.
  7. Make one change at a time. Don't adjust competitor scope, margins, and price bands in the same session. You won't know which change caused what.
  8. Day 30 check: if margins are stable and Buy Box win rate is improving, you can consider widening the competitor scope or expanding to more SKUs. If margins are slipping, tighten the floor first.
  9. Day 45-60: use your KPI data to decide which SKU groups graduate to full automation and which need manual review.

The KPIs that actually matter

Buy Box win rate / Buy Box percentage. This is the share of page views where your offer holds the Buy Box position (SellerEngine, June 2022). On Amazon, find it under Reports > Business Reports > the "Buy Box Percentage" column. On Walmart, download the Buy Box report via Seller Center (Walmart Marketplace Learn, March 2025). The Walmart report gives one row per SKU including your current item price and shipping price, which is useful for validating your feed math.

Gross margin per SKU. Track this weekly against your margin floor. Any SKU trending toward its floor needs attention before it crosses it.

Sales velocity. Units sold per day or week for the pilot cohort. Control for any promotions or external events that could skew the window.

Conversion rate. Are competitive prices translating into purchases? If Buy Box win rate climbs but conversion stays flat, the issue may not be price at all.

A simple tracking table helps here:

SKU Group Start Buy Box % Week 2 Buy Box % Day 30 Buy Box % Margin stable? Decision
High-comp , , , Yes/No Expand/Adjust
Private label , , , Yes/No Expand/Hold

When something goes wrong

Prices aren't updating. Start with feed connectivity. Then check SKU/ASIN mapping. Then confirm the rule is enabled and applied to that SKU group. If all three check out, look at the price calculation output for a specific SKU to see if the min/max bounds are blocking movement.

Buy Box win rate isn't improving. Your min price may be set too high relative to the competition. Alternatively, you may be competing against the wrong seller set. Check who holds the Buy Box and whether that seller is in your competitor scope.

Margin drops unexpectedly. First, audit your landed cost inputs. A missing fulfillment fee is the most common culprit. Second, check whether upward repricing is enabled. If it's off, the repricer can only go down.

Prices are overreacting to noise. Narrow your competitor set, add rate limits on how quickly price changes can fire, and consider smoothing rules by velocity so low-volume SKUs don't swing on a single competitor move.

Inventory mismatch. If your feed shows incorrect available quantity, repricing may behave unexpectedly. Make sure your feed reflects sellable stock, not total on-hand.

For any issue you can't resolve in 5 minutes: check feed health, confirm the rule is matched to the right SKU, look at the price calculation output for a specific ASIN, note any SKU exceptions, then contact support with that data in hand. Informed Repricer's live chat support is set up to triage exactly this kind of structured diagnostic, which speeds resolution considerably.

The goal of the first 60 days isn't to automate everything. It's to prove the system works safely with a bounded set of SKUs, validate that margin floors hold under real conditions, and build enough confidence in the data to expand. Once you hit your Buy Box and margin targets consistently, expanding to the full catalog becomes a straightforward decision rather than a leap of faith.