TL;DR
- Cut to the one workflow that creates value. Everything else is v2.
- Build for the first 100 users, not 100,000. Optimisations come later.
- Use boring tech (Next.js + Postgres + Stripe). Novel stacks add weeks for nothing.
- Ship in 8 weeks max. Anything longer means scope is wrong, not engineering.
The 5-question scope filter
For every feature you're tempted to add, run it through these. If it can't survive all five, it's v2.
- Does this feature directly create the value users pay for? If not — admin convenience, "polish", "nice-to-have" — drop it.
- Will the first 100 users notice if it's missing? If they won't even ask, it's not MVP.
- Can you fake it manually for the first 50 users? If yes — Excel, WhatsApp, manual admin — fake it. Build it when the manual cost crosses the dev cost.
- Is there a free SaaS that does this? Use Calendly, not a custom booking widget. Use Mailchimp, not a custom email engine. Save weeks.
- Will this feature take more than 5 days to build? If yes, defer to v2. MVPs ship by stacking small wins.
What every MVP needs (the actual minimum)
| Layer | Minimum | What you DON'T need yet |
|---|---|---|
| Auth | Email + Google login | SSO, SAML, custom OTP, social-of-the-month |
| Core workflow | One happy path, end-to-end | Edge cases, bulk actions, advanced filters |
| Payments | Razorpay one-time + UPI | Subscriptions, refunds, multi-currency |
| Admin | List + CRUD on key entities | Reports, dashboards, role hierarchy |
| Notifications | 1 transactional email per action | Preference centre, digest, push, SMS |
| Analytics | PostHog free tier | Custom dashboards, attribution modelling |
If your MVP spec is >2 pages, you have a v1 product, not an MVP. Cut until it fits on 2 pages. The cuts hurt — that's the point. See our 8-week MVP playbook →
The "manual first" rule
Every operations feature can be faked manually for the first 50 users. Examples:
- Onboarding: personally WhatsApp each new user. They love it; you learn what's broken.
- Matching / routing: sit in a Google Sheet, match users yourself overnight. Build the algo once you know what "good match" means.
- Reports: SQL query against the prod DB once a week, paste into a Slack channel. Productionise when someone asks for it twice.
- Refunds: manual via Razorpay dashboard. Build automated flow when refunds >5/week.
You'll feel like a fraud doing manual work. You're not — you're learning what the system actually needs to do before paying to automate the wrong thing.
FAQ
What if my investor wants "all the features"?
Investors want traction, not features. Show 100 paying users on the lean MVP and they'll fund the v2. Show a feature-complete product with 0 users and you'll close no round.
What if a competitor has feature X and we don't?
You're not competing on feature parity at MVP stage. You're competing on whether anyone uses it. The competitor with 50 features and 200 users is in worse shape than you with 5 features and 200 users.
Last reviewed: 21 April 2026.
Want this built for you?
Talk to Kashvi — 30-min call, honest assessment, no pitch deck.