Meeting with Mike Savino

Meeting with Mike Savino

Meeting Notes from chat with Mike Savino at Launch.co, a VC fund. He also ran an e-commerce shop in a past life.

Pre-Meeting Notes

• Stanford student studying CS — background in ML, have done stints doing more AI research, software engineering (recently at Splunk), also did venture at TCV
• Currently taking a class where we build software for corporate partners. Our corporate partner is a venture fund (Unusual Ventures, John Vrionis from Lightspeed, Jyoti Bansal from AppDynamics and Harness), so we’re building a “startup” for our class project
• We iterated a lot, but all our backgrounds are in ML and systems, and we’ve all kinda had this shared thought that search would be the breakthrough use case for ML.
• We also do online shopping for items, and we recently had struggles with it, where we realized that if we could give more of a ChatGPT type description, we could probably get better answers
• So that’s what we conceptualized — basically, this like, Perplexity/ChatGPT, but specifically for shopping. We built a quick demo that I can show you just for books rn, but we plan to make it almost this “wire cutter at scale”
• [SHOW DEMO] — only works on a small amount of books now, but hoping to scale in the future
• Think biggest thing up for debate is about what the right motion is. We could easily build a consumer app, but also feels like shopping is super social now, and way to monetize would probably be either being an affiliate or making people sign up to do more searches (a la ChatGPT/Claude or smth). If we could build really strong generalizable algorithm, could work
• On the other hand, think there’s an interesting B2B motion, where we instead become Algolia. In other words: we bring your data, we make it searchable in this semantic way by fine-tuning on your own custom model. This could be more consistent revenue — for instance, just hit an API — but we’re still in the phase of doing more customer discovery and trying to find design partners to build out a prototype
• So yeah. Biggest questions. In general I think there’s also a lot of noise in AI space, so want to build something that’s enduring. Biggest barrier in kinda going more full mode on the project.

Meeting Notes:

• Search e-commerce software development shop
• Built sites on Magneto
• Thinking about how ambitious you want to be
• Idea that you’ll build a company built on affiliate revenue
• B2B SaaS, solid revenue, could get plugged in
• Ambitious — We’re gonna be Google, Consumer reports, for Shopping — huge Unicorn
• Path of commitment
• Uncommon — underlying recommendation, not always about the algorithm, Consumer Reports, still use
• Certain group of people who want to know about certain products — finding products
• Search engine — Wirecutter, Consumer Reports, provide unbiased advice within a category
• People tend to pay more for
• Our co-founders, which way do you have to take it?
• If you choose one path, you’re never gonna go to the other path
• Shopify store
• How many people does Wirecutter have testing their products
• Wirecutter, Consumer Reports, paywall
• Focus groups, B2C, consumer discovery,
• How much do people really need this?
• Identifying early adopters, ideal customer profile, etc.
• Superhuman — had to schedule 30 minute class
• Small group — heavy research shooters
• Consumer reports subscription
• Founder University — 12-week program, take an idea, build an MVP, etc.
• 2-founder companies — good way to be around this same stage