The case at a glance
- Annual revenue: crossed $1M ~3 years ago, still $1M+
- Profit margin: 40%
- Channel split: 95% Amazon, 5% Shopify
- Core products: 3-4. Total SKUs: ~1,500
- Capital burned before the first durable winner: $10-15K
- Team: 3 employees in the Philippines plus freelancers
- Paid traffic: almost none. Organic Amazon search and storefront
Started late 2017. Ate a string of failed launches, then landed one product in 2018 that's still running and was profitable from its first shipment. Everything expensive about a physical products business — ad spend, a US team, heroic capital bets — he removed one piece at a time.
Stack: Alibaba for sourcing and samples, Amazon FBA, 3PLs in the US and Canada, Shopify on the side. Products are self-developed and explicitly non-commodity: if it's a thing anyone can source and resell, he walks.
What he's actually selling: Amazon long-tail keyword coverage
The product isn't toys. The product is 1,500 doors into Amazon search.
Every colorway, size, and bundle is a fresh shot at a long-tail query somebody is typing right now. He isn't bidding against ten sellers for one fat head term — he's occupying 1,500 thin ones nobody bothered to write a listing for. That's why the margin is 40%. It's not better merchandising. It's an absent cost bucket. In a category where competitors hand Amazon Ads a slice of every sale, the gap is the margin.
The mechanism holds up because of the ratio underneath it. Three to four cores producing ~1,500 SKUs is a 375:1 return on creative work. He designs once and permutes forever, amortizing tooling, photography, and listing copy across hundreds of children. The operator launching a brand-new product every quarter starts from zero every time: new samples, new photos, new reviews.
And the reason nobody with a marketing budget copies him is boring. Routing 1,500 SKUs through third-party warehouses into FBA is operationally miserable. Misery is a filter.
Templatize the variation, not the product. Whatever you build once, ask what the permutation axis is before you ask what's next.