Steel threads: the single most important product market fit technique that I use (that no has heard of)

I learned about the engineering concept of ‘steel threads’ about ten years ago from a colleague who had come from Microsoft. I assumed it was fairly well known, but I haven’t heard of it since (and its wikipedia page is still a stub!)

But this concept has probably had the single largest influence on how I think about building new products and companies.

In essence, create the bare minimum end-to-end thing that works first. Maybe for a web app that is a single button that connects to an API backed by a database - instead of building out the entire UI first. For a startup, it might mean getting to a first sale as quickly as possible, even if you take every shortcut to get there.

It’s so easy for builders to get stuck in fleshing out the details. You’ll want to describe the whole database schema first. Or finagle the unit economics until that first sale is profitable. But none of that really matters.

What starting with a steel thread does is really shorten time to feedback. And you learn a lot. What breaks? What do you need to fix? What is truly the smallest thing you can do? What actually mattered to the customer?

The concept comes from building bridges, where you would use a thread to connect the two sides of land before you began building with steel. But I found this analogy a bit more useful:

Imagine you are building a vehicle to get people from one place to another. Do you begin by building a piece of an airplane wing, then the other wing, then an engine, and so forth, until you have a whole airplane - or do you begin with a bicycle, upgrade it to a scooter, then upgrade it to a car, then upgrade to an airplane.... you see the difference.

The steel thread concept is similar to building a minimal viable product (MVP), but I find it a lot more practically useful because you can kind of always argue about what should be in the MVP or not. And in truth - you don’t always know upfront.

But the steel thread is very simple: does the core action work end-to-end? Start there.