Data-driven algorithmic VC investing fails at the seed stage, says Zelby
Elaine Zelby · Brainstorming Million Dollar Ideas with Elaine Zelby | My First Million Ep. #175 · Apr 26, 2021 · at 24:28
SignalFire VC Elaine Zelby discusses the limitations of automated data models and algorithmic underwriting in early-stage venture capital.
0:00 / 0:10exact quote · 10.0s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“In my opinion, it only works at certain stages and in certain categories. This does not work at all at the seed stage. You're betting on people for the most part, and you're betting on the growth of a category. There's no data.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Elaine Zelby
Long funnels work for consumer products, but B2B requires shorter onboarding
“One nuance here that is a hundred percent accurate for consumer and a hundred percent inaccurate for B to B for anything enterprise shorter is better.”
SignalFire uses proprietary algorithms for deal sourcing and talent intelligence
“So we buy, scrape, or, you know, aggregate every data signal you could possibly imagine. And our engineers have proprietary algorithms that use that data to do one of two things. One, we have different systems that are alerting systems for us on the investment…”
SignalFire completes a minority of investments purely from automated alerts
“On the sourcing side, we've definitely done deals that were completely from our alerting system. In addition, we just do a ton more outbound than the average firm because we'll go in these meetings and we'll triage based off of, oh, this looks interesting, and…”
Enterprise systems of record achieve defensibility because data migration is painful
“Ripping and replacing software is so painful, and it's incredibly hard to get people to do it. You've already trained your team on it. It has all of your information and data sitting in that system. It's just really hard to rip something out when it's there, w…”
Shoe subscription startups must build B2B infrastructure for major legacy brands
“The only way I think you do this as an independent company is you provide the picks and shovels infrastructure for every brand to do it. So you go and sell to the Brooks, to the Nike, to the Adidas, to the Reebok. I think that's doable.”