My exact system for building a sourcing database from scratch
Most people start OA by randomly browsing retailer websites hoping to find a deal. That works sometimes. But it is painfully slow and there is zero consistency to it. What changed everything for me was building a seller ID database. Once you have one, you are not guessing anymore. You are working from a list of products that REAL resellers are already selling profitably. Here is exactly how I do it. STEP 1: FIND CROWDED LISTINGS Open Keepa Product Finder and set the minimum FBA offers to something like 8 or 10. You are NOT looking for a good listing to sell on. You are looking for a listing with a LOT of resellers on it. The more sellers, the better for this step. Those sellers are your goldmine. STEP 2: EXTRACT THE SELLER IDs Once you find a crowded listing, scroll down to the Keepa chart. Click the Data tab, then the Offers tab. You will see all the sellers listed there. Copy their seller IDs one by one. Yes it takes a bit of time. That is the point. You are building an asset. STEP 3: VALIDATE EACH SELLER Paste each seller ID into the Keepa Product Finder seller ID section. Check how many products they carry. If they have 15 to 20 or more products listed, they are probably a reseller just like you. Add them to your spreadsheet. If they only have 2 or 3 products, skip them. They are probably just clearing out their garage. STEP 4: TREAT THIS LIKE A SOURCING SESSION Here is the mindset shift that most people miss. Do not try to build your database AND source leads at the same time. Separate the two. Spend a dedicated hour JUST collecting seller IDs. No analyzing products. No checking margins. Just building the list. Then later, you use the list. STEP 5: USE THE DATABASE WITH KEEPA PRODUCT FINDER Once you have a solid list, take 20 seller IDs at a time. Format them as a comma separated list (I just paste my spreadsheet into ChatGPT and ask it to format them). Drop those 20 IDs into Keepa Product Finder and it will pull up every product those 20 sellers carry. That is usually around 1000 listings. Then you apply your filters to narrow it down to a realistic batch you can actually analyze in one sitting.