Scope the Smallest Thing That Works

Why it matters

You have spent the last two weeks in Spark listening. Now you switch to building, and the biggest risk in the Make stage is not that you build the wrong thing — it is that you try to build too much of it. Ten screens, five features, a login system, a settings page. None of that is the reason someone would use your thing. One core action is.

The gate for this stage is simple: a working thing, live at a public URL, that a stranger can actually use. You cannot hit that gate by disappearing for four weeks to build something impressive. You hit it by picking the smallest version that still solves the problem you heard about in your ten conversations, and shipping that.

This lesson is where you draw the line. What is the one thing someone does when they use your product? Not the fifth thing, the first thing — the one action that, if it worked, would prove your idea is real. Everything else gets written down on a “Not in v1” list, not deleted, just parked for later.

You are also choosing your tool today: an AI app builder for something people click through, or a simple page for a digital download or a service you deliver yourself. Get this spec tight, and next lesson you will build the whole thing in one sitting.

What to watch for

How to Build An MVP | Startup School

• Watch for how Michael Seibel defines an MVP: the smallest thing that lets you test your core hypothesis, not a smaller version of your full vision.
• Notice the examples of founders who launched with something almost embarrassingly simple, and why that worked in their favor.
• Skip trying to take notes on every framework mentioned; focus on the one question he keeps returning to — what are you actually trying to learn?

Start with Claude

Bring your Problem Inventory notes and your one-liner into a new conversation, and let Claude help you cut your idea down instead of writing the spec for you.

I'm scoping the smallest working version of my idea for the Make stage of Make First. My one-liner is: [paste it]. Here's what I learned from my 10 conversations: [paste a short summary].

Don't write my MVP spec for me. Ask me what the single core action is — the one thing a user does that proves the idea works. Push back if I list more than one action or more than 3 screens. Help me sort everything else into a "Not in v1" list instead of just cutting it. At the end, ask me whether a stranger could finish the core action in under 2 minutes, and if I'm not sure, help me figure out why.