0→1 vs 1→N
Building something new vs scaling something working — different jobs.
The short version
Building something that does not exist yet (0→1) and scaling something that already works (1→N) are different jobs that reward opposite instincts. 0→1 is about finding product-market fit under deep uncertainty — speed, learning, and killing bad bets. 1→N is about reliability, efficiency, and compounding what works. Knowing which one you are in tells you what "good" looks like.
Why it matters
Most PM advice is implicitly about one mode and silently wrong for the other. Apply 1→N rigor (process, dashboards, optimization) to a 0→1 problem and you will polish something nobody wants. Apply 0→1 scrappiness (move fast, skip the hardening) to a 1→N product at scale and you will break trust you spent years earning.
The senior move is diagnosing which job you are actually in — often the two coexist in the same org, and even the same product at different stages — and shifting your own behavior, metrics, and risk tolerance accordingly.
How the two differ
Common mistakes
Bringing heavy process and dashboards to a 0→1 search for fit — measuring a thing that has no signal yet.
Staying scrappy into 1→N — skipping the reliability, instrumentation, and ops that scale demands.
Assuming the skills (or the people) that won 0→1 automatically win 1→N — they often do not.
Declaring product-market fit prematurely and scaling a thing that never actually fit.
In a startup
You are almost always in 0→1: optimize for speed-to-learning and be ruthless about killing what is not working. The danger is scaling before you have real fit — growth amplifies a broken model, it does not fix it.
In an enterprise
You are usually in 1→N — efficiency, reliability, and compounding an existing P&L. The danger is treating a genuine 0→1 bet like a mature product: smother it in process and stage-gates and it never gets the room to find fit.
Quick check
Optional — test the lesson. Nothing is gated.
1. What is the primary risk in a 0→1 effort?
2. Why is applying 1→N rigor to a 0→1 problem a mistake?