Four models, ±15-25% accuracy, zero public revenue data to start from. Gumroad, Systeme.io, and Whop never publish what a product earns. So how can anyone estimate it? This is the honest answer: no black box, no hand-waving, just the signals we collect, the models we run, and the limits we accept.
The benchmark we copied: BrandSearch
Before building InsightRaider, we studied how others estimate revenue for private businesses. The gold standard is BrandSearch, a company valued at $110M that estimates revenue for Shopify stores and Amazon sellers. They proved you can hit +/-20% accuracy by combining public signals with proprietary algorithms. We adapted that playbook for infoproducts.
Pillar 1: scraping the public signals
Digital product platforms expose more than you would expect. We systematically collect ranking data (bestseller position, overall platform rank, trending flags), social proof (review count, average rating, rating distribution, review velocity), and product metadata (price, launch date, creator follower count, products per creator).
