YC OS · Chapter 04
Find the 90/10 Solution
Solve 90% of the customer's problem with 10% of the effort a 'complete' solution would take — then ship it.
After this lesson
- Separate the 10% of a problem that's actually hard from the 90% that's just work
- Ship a narrow version that removes most of the pain, not all of it
- Avoid building the fully general version before anyone needs it
The 90/10 solution
Most painful problems have a version that removes 90% of the pain for 10% of the engineering cost — usually by narrowing scope, accepting a manual step, or serving one persona instead of everyone. Startups that survive long enough to build the complete version are almost always the ones that shipped the 90/10 version first.
How to find your 10%
- Ask which single step, if fixed, would make the customer stop complaining
- Look for the version that works for one persona instead of every persona
- Accept a manual or semi-automated step behind the scenes at first
- Cut edge cases that affect few users but would double the build time
- Prefer 'good enough and shipped' over 'correct and theoretical'
Case
Case: Zappos sold shoes it didn't own
Before building any inventory, warehouse, or supplier relationships, Zappos founder Nick Swinmurn photographed shoes at a local store, posted them on a bare-bones website, and told the store he'd buy at full retail price if anyone ordered. When an order came in, he drove over, bought the shoes, and shipped them himself — losing money on every sale.
Takeaway: The 90/10 version didn't need to be profitable, or even efficient. It only needed to answer one question: will people buy shoes online without trying them on? Once the answer was yes, building real inventory and warehouses stopped being speculative and started being obvious.
Where founders overbuild instead
- Building for scale before anyone is using the product at all
- Solving hypothetical enterprise requirements before the first real customer asks
- Generalizing a feature for every persona before one persona is delighted
- Treating 'technically correct' as more important than 'used every week'
Is this feature the 10% or the 90%?
Before you build it, classify it honestly.
Ship the 90/10 version
- Write the full 'ideal' solution in one paragraph
- Cross out everything that isn't needed by your first real customer
- Replace automation with a manual step wherever that's faster to ship
- Ship what's left this week and watch what breaks or gets requested next
Read the original
YC's essential startup advice'Find the 90/10 solution' sits right next to 'do things that don't scale' in YC's own pocket guide — they're the same instinct applied to product and to hustle.
This chapter is a working summary. When the idea matters, read the source once — then come back and do the practice.