We realized that Wall Street was writing reports on stocks based on Amazon product reviews 2 weeks after product launches. The stocks would trade up 5%+ up or down on these reports. But everything in the reports could be automated, and thus done much faster . Also, we were exposed to internal data products at banks and hedge funds which didn't account for skew or absolute level of reviews . Fake reviews on Amazon (such as those seen on The 4 Hour Work-Week) also lowered the signal to noise ratio unacceptably in such products.

We fixed this.

Our target user is a hedge fund, but our business plan will involve proprietary vetting of what we think is a very strong signal that we managed to extract during the hackathon. Brand momentum would have allowed a trader to profit massively on a long Skechers, short Adidas position - which we don't feel is spurious and aligns with consensus core drivers for the stocks. By working directly with a prime broker, or a hedge fund - we'll generate track record for our signals which will help build our moat and keep our margins strong over time. Conversations with both are under way.

Our vision is more than just parsing Amazon - it's combining large numbers of stock-moving data sources into a single, unified signal generation system.

Built With

Share this project: