Normalization Powers Software Defined Underwriting
Kavin Stewart, Partner at Tribe Capital, on Reddit's 10x opportunity
The real edge in data driven venture investing is not collecting startup metrics, it is forcing messy company data into one common schema so patterns are comparable across sectors and stages. That means turning each startup’s custom chart of accounts, KPI definitions, and reporting cadence into the same fields, so a firm can line up retention, payback, burn, and growth against a large internal benchmark set instead of judging each company on founder specific framing.
-
This is manual because startups rarely report the same way. One company books annual contracts upfront, another recognizes revenue monthly. One reports users, another reports active teams. Normalization means recasting those inputs into standard units before any ranking model is useful.
-
Tribe’s approach builds on its growth accounting work, which was designed to break growth into comparable drivers rather than headline vanity metrics. The same idea in underwriting is to separate real business quality from differences in how founders package the numbers.
-
The payoff is speed and consistency. Once normalized, the same dataset can feed benchmark reports, market maps, and automated memos. That is especially valuable in AI, where many companies look similar on the surface and product market fit is easier to judge through usage and retention patterns than narrative alone.
This heads toward software defined underwriting. As normalization gets automated, venture firms that still rely on partner intuition and bespoke spreadsheets will look slow, while firms with structured company level data will be able to screen more deals, update views faster, and spot outliers earlier.