Amazon Market-Research Web Crawler
From October 2017 to December 2019 I founded and ran Everyman's Products LLC, an Amazon e-commerce company that generated $119,000 in product sales in its first year. The riskiest decision the company faced was market selection: designing a product, negotiating with an overseas manufacturer, and ordering inventory all commit money before the first sale, so picking a weak market is expensive. I built and deployed a web crawler that scraped product sales data from Amazon.com so those market-entry decisions ran on data instead of intuition. The code predates this portfolio and isn't public.
What it did
- Extracted product sales data from Amazon.com across candidate product markets.
- Fed that data directly into the company's core commercial decision: which market to enter next.
The stakes
- $119,000 in first-year product sales on Amazon.
- Three products I designed generated $20,000+ of those sales.
- Every market pick triggered negotiations with overseas manufacturers and real inventory commitments, so a wrong pick tied up capital in products that wouldn't move.
The crawler was one input among many into those decisions — but it was the one that put actual Amazon sales numbers on the table before any money was committed. That's the pattern I still care about: software built to answer a specific commercial question, judged by whether the decisions it informed made money.