Every product I've taken from idea to market at Xcellerant Ventures has run on the same constraint: a working MVP, live, in under 90 days. Said out loud in a planning meeting, that sounds like a scheduling exercise. In practice, it's a willingness exercise. Are you willing to launch something that's missing things you know a customer will ask about, on purpose, because shipping without them teaches you more than guessing with them ever would?
The list that's harder to write than the roadmap
Every 90-day MVP I've run has had the same quiet ritual: a short, explicit list of things we are choosing not to build yet, written down and agreed to by everyone in the room, not things we forgot, things we decided against. For FastRead.io, that list included a proper admin dashboard, most account settings, and any kind of collaboration features. All reasonable things a writing tool "should" have. None of them were the reason anyone would use it in month one.
The roadmap tells you what you're building. The cut list tells you what you actually believe matters, and it's a much more honest document.
What's never allowed on the cut list
Ninety days doesn't mean cutting until it's easy. Some things stay non-negotiable regardless of deadline pressure:
- The single core promise has to work end to end, with real quality (for FastRead that meant a finished, readable chapter, not a demo of a feature that produces one).
- Whatever creates the trust the category needs. For Ovela, that meant the concierge intake and a vetted specialist. Cutting that to hit a date would have shipped a faster version of the wrong product.
- A way to actually learn from real usage (basic analytics, a channel for user feedback). Shipping fast without a way to learn fast is just shipping slowly with extra steps, later.
Across FastRead.io, Myautobio, Clintwin/Medsim AI and Formulai, every one went from idea to a live, paying-customer-facing product in under 90 days, and every one launched visibly missing things a longer timeline would have included.
The conversation that actually determines the deadline
The 90 days was never really a technical estimate. It was a negotiation about embarrassment. How much of "not done yet" could the team and the stakeholders above them tolerate being visible to a real customer? The teams that blow past 90 days aren't usually behind on engineering. They're behind on that conversation, still quietly hoping they can ship something complete enough that nobody notices what's missing. Nobody ever stops noticing. The only choice is whether you decided what's missing, or let the deadline decide it for you at the last minute, badly.