Ninety days is a useful constraint precisely because it is uncomfortable. It forces the question every early product must answer: what is the smallest thing that would teach us whether this idea deserves to exist?
We have run this playbook with dozens of founders. The timeline holds when three disciplines hold.
Weeks 1–2: Decide what you are not building
The discovery sprint exists to shrink scope, not grow it. We map the riskiest assumption in the business model and design the MVP around testing exactly that. Every feature that does not serve the core learning goal goes to the someday list — publicly, so the decision is visible and reversible.
“An MVP is not a smaller version of your product. It is the fastest honest test of your idea.”
Weeks 3–10: Build in public, internally
Working software from the second sprint onward, demoed weekly to the founding team. The demo cadence is the quality mechanism: nothing hides for a month and emerges wrong. Boring technology, managed services, and one well-understood stack keep the velocity honest.
Weeks 11–13: Launch to dozens, not thousands
The first release goes to a hand-picked cohort you can actually talk to. Instrumentation, feedback channels, and a weekly learning review matter more than press. The MVP succeeds if it changes what you build next — that is its entire job.
The founders who win with this playbook treat day ninety-one as the real beginning: they arrive with a live product, actual usage data, and a roadmap written by users instead of assumptions.
Emperor Brains Engineering
Emperor Brains LLP
