Most first-time founders spend 6–12 months building their MVP. The successful ones ship in 90 days. The difference isn't talent or speed — it's scope discipline.
Here's the 90-day structure we walk founders through.
Days 1–7: write the one-page brief
Before any building, write one page that answers:
- Who's the user? (Specific. Not "small businesses" — "15-person Pakistani retailers managing inventory across 3 outlets".)
- What's the painful job-to-be-done? (Specific. Not "manage inventory" — "reconcile end-of-day stock without keying numbers into three systems".)
- What's the simplest possible solution? (Single-screen, single-action where possible.)
- How will we know if it's working? (One measurable signal. "Eight retailers using it weekly within 60 days of launch.")
If this page is more than 500 words, you haven't cut enough. Cut.
Days 8–14: validate with 10 conversations
Before any code, talk to 10 prospective users. Not survey them. Talk to them.
Three questions per call:
- "Walk me through the last time you faced [the painful job-to-be-done]."
- "What did you try? Why didn't it work?"
- "If a tool did [the simplest possible solution], how often would you use it?"
After 10 conversations, you'll have one of three outcomes:
- Strong signal. Build it.
- Weak signal. Iterate the brief, do 10 more conversations.
- No signal. Kill the idea. (This is the most valuable outcome — saves you 90 days.)
Most founders skip this step because it feels slow. It saves 6 months on every wrong-direction product.
Days 15–35: ship the ugly version
Three weeks. Build the one-screen, one-action version.
Constraints:
- No login if you can avoid it (reduces friction).
- No mobile app (PWA is enough).
- No payment integration (if monetisation is uncertain, get to usage signal first).
- No design polish. Tailwind defaults. Lucide icons. Done.
The version you ship at day 35 should make you embarrassed. If you're proud of it, you over-built.
Days 36–60: get to 10 users
Spend 25 days getting to 10 actively-using users — not 10 signups, 10 users.
This is harder than the build. Most founders treat this stage as "marketing." It's not. It's another round of conversations, with code in hand.
For each of the 10 users:
- Onboard personally (not "send them the link").
- Watch them use it the first time.
- Note every place they hesitate, click wrong, or ask "what does this do?"
- Fix one thing per user per week.
By day 60 you'll either have 10 users using it weekly (signal: keep going) or you'll have churn (signal: revisit the brief).
Days 61–90: harden, then launch
Final month:
- Days 61–75: Fix the top 5 friction points from user observation.
- Days 76–82: Set up basic monitoring (errors, key metric, customer feedback channel).
- Days 83–89: Soft launch — extend from 10 users to 30 in adjacent network.
- Day 90: Decide.
The day-90 decision is one of three:
- Keep going. 30+ users using weekly, you have signal worth investing in.
- Pivot. Some users love it for an unexpected reason. Reorient the product around that reason.
- Kill it. Less than 10 weekly users despite 90 days of work and onboarding. Cut losses, keep the lessons.
What kills the 90-day MVP
Three patterns we see kill 90-day timelines:
- Scope creep. Founder adds "while I'm at it" features in week 4. By week 8, MVP has 12 screens. Don't add. Cut.
- Premature scaling. Founder spends week 6 on "how will this scale to 100,000 users?" — before reaching 10. Wrong question. Get to 10 first.
- Solo without accountability. No external check-ins. Two months in, no one but the founder knows what's actually shipped. Find one mentor or peer who reviews your week-end progress every Friday. Just one.
The point of the 90 days
The 90 days isn't to build a finished product. It's to ship a thing real users use, learn from their behaviour, and decide what to do next with hard data instead of imagination.
Most successful products you've heard of had three to five 90-day cycles before they found their shape. The discipline isn't about getting it right the first time. It's about being able to run the cycle at all.
