Four weeks to productize.

I built QuoteLogic. It ran inside a public enterprise for years. When I left, I had a product. What I didn’t have: auth, billing, a real database, production infra. The stuff enterprise companies usually let you borrow for free, but not in my case. Read the origin story

One day to build. Four weeks to learn.

Convex for the database and realtime. WorkOS for auth. Stripe for billing. Sentry for errors. PostHog for analytics. I picked all of them because they’re widely used and proven in production.

Most of those four weeks was one-time learning cost. Figuring out Convex, WorkOS, and Stripe for the first time. Staging vs prod environments. Making the app recognize both dynamically. Wiring Google and Azure OAuth through WorkOS. Iterating on security.

Today, with everything learned, that same backend takes a day or two. That’s the number I quote to new clients. The four weeks was the cost of learning, not the cost of building.

The thesis

This is what I tell founders now: the unique thing you’re building is the product. The foundational pieces most products need are the stack. Compose the stack, build the product. I proved it with QuoteLogic before I started recommending it.

Many teams prioritize infrastructure first, building custom auth, billing, queues, and realtime early on. That work is valuable, but it can pull focus away from the product itself and slow down delivery. The risk is spending more time on infrastructure than on what actually differentiates the product.

The opposite move is to defer to platforms that already solve those foundational pieces, then spend your engineering on the part of the product platforms cannot ship for you.