The Amazon Online Arbitrage Sourcing Workflow
Online arbitrage (OA) is buying products from online retailers and reselling them on Amazon for a profit. The winners aren't hard to analyze. They're hard to find efficiently, because most candidate deals don't survive Amazon's fees. A repeatable workflow keeps you from wasting hours on products that were never going to work.
1. Generate candidates, don't hunt one by one
Sourcing one product at a time doesn't scale. Build candidate lists in bulk: retailer clearance and sale pages, brand outlet stores, and deal feeds. The goal at this stage is volume of candidates, not certainty; you'll filter hard in a moment. Capture them into a single spreadsheet with at least a product identifier and your buy cost.
2. Screen the whole list against Amazon economics
This is where most of the list dies, and that's good. For every candidate, calculate the Amazon fees and your true landed cost, then the ROI. Set a minimum ROI and drop everything below it. A few percent of a big candidate list clearing your threshold is a normal, healthy result.
3. Confirm demand and competition
A profitable price means nothing if the product doesn't sell or you can't win the sale. Check the sales rank for steady demand, the Buy Box to see whether Amazon dominates the listing, and the offer count for crowding. Keepa history is the fastest way to read all three at once.
4. Check sellability before you spend
Confirm you can actually list the product: brand gating, hazmat, meltable, and oversize restrictions all turn a “winner” into stranded cash. Screen these out as part of the same pass, not after the box arrives.
5. Buy, track, and repeat
Buy only the products that cleared every filter, record what you paid against your max buy cost, and feed what sells back into which retailers and categories you prioritise next time. OA is a numbers game won by tightening the funnel, not by analyzing harder.
Make the screening step instant
Steps 2–4 are the same calculations repeated across a long list. Running them in Gridwinner turns a candidate list into a graded, risk-flagged shortlist in one local pass, so the slow part of OA becomes the fast part. See how it handles a bulk OA list.