Look Before You Leap
The loop, planned backwards
Most people plan forwards: build this, then that, then launch. Winners plan backwards — starting from the one thing they need to learn.
Progress is laps of learning, not features shipped. You act Build → Measure → Learn, but you PLAN in reverse: decide what you must learn, then what you'd measure to know it, then the smallest thing to build. Then make every lap faster.
- Name what you must LEARN — e.g. 'will buyers pay 200 for this?'
- Decide what you'd MEASURE to know it — paid preorders, not likes or 'maybes.'
- Build the smallest thing that produces that measurement.
- Run the lap, read the data, and feed it into the next idea.
- Then attack the lap TIME — more at-bats per dollar beats one big polished swing.
One owner spends six months building, measures once, then learns he was wrong — a single slow lap. Another runs six one-month laps and course-corrects five times. Same six months; only the second now knows the market cold.
Michelangelo was asked how he carved David. The idea — 'remove everything that isn't David' — is the whole point: iteration is not a detour on the way to the work, it IS the work. Owners who treat every revision as proof they failed quit right before the statue appears.
Confusing motion with a loop. Building for six months without measuring is not one long lap — it's zero laps. You only learn when the loop closes on real data.
Don't measure yourself by what you've built; measure yourself by how many complete learning laps you've run — and spend your cleverness on making the next lap shorter.
Take your current plan and write the loop backwards on one line: 'To learn ___, I'll measure ___, by building ___.' Then ask how to run that loop in weeks, not months.
Look Before You Leap