SaaS MVP in 90 days. Real playbook.
The 90-day SaaS MVP is not a mythology. Scope discipline, stack picks that compound, a 3-to-5 person team, honest billing from day one, and three cuts that keep the timeline realistic.
Small scope, boring stack.
Ninety-day SaaS MVPs work when three conditions are in place. Scope discipline: one core workflow, Stripe billing, basic auth, one integration maximum; everything else gets cut via the three-cut rule (delete 50 percent, then 25 percent, then another 10 to 15 percent). Boring stack: Next.js plus TypeScript plus Tailwind on the frontend, PostgreSQL on Supabase or Neon for data, Clerk or Auth.js for auth, Stripe for billing, Vercel for deploy. Team: minimum 1 full-stack engineer plus 1 product-owner founder; comfortable 2 engineers plus 1 designer plus 1 founder-product-lead. Budget: founder-led near zero out-of-pocket plus infrastructure; agency-led 80K to 350K depending on depth. Most failed 90-day attempts fail from scope creep or novel stack decisions that eat time; the successful ones ship narrow, learn, and expand from week 13 onward.
Cut 50, cut 25, cut 15.
First cut: remove 50 percent of the initial feature list. The initial list always looks reasonable because it was built without a calendar constraint; with 90 days as the ceiling, half of it was always going to be a year-two deliverable. Cut everything that does not directly demonstrate the core value prop. If the product is a team-collaboration tool, the core is team workspace plus real-time editing plus invite flow; cut threaded comments, notification center, activity feed, third-party integrations, and mobile. All of those will wait.
Second cut: remove another 25 percent of what is left. The first cut trimmed obvious non-essentials; the second cut is harder because what remains all feels important. Challenge every feature: what happens if we ship without it? If the answer is "users can still complete the core workflow," cut it. Common cuts on the second pass: advanced permissions (ship with one admin role, defer multi-role to month 4), custom themes (ship with one theme), advanced search (ship with basic filter), granular notifications (ship with daily digest only).
Third cut: remove another 10 to 15 percent. This is where discipline hurts. What remains feels essential; cut the least-essential anyway. Trust that it can be added in weeks 14 to 20 if users actually want it. The remaining 10 to 15 percent is the true 90-day MVP. Week 13, you ship to 10 to 50 design-partner customers, watch what they actually use, and add back features based on real usage data. Almost always, 30 to 40 percent of the features cut during the three-cut pass never get re-added because users did not want them.
Boring compounds. Novel does not.
Frontend: Next.js (React) with TypeScript and Tailwind CSS. Next.js handles routing, server rendering, API routes for small backends, and deploys to Vercel with zero configuration. The alternative of SvelteKit or Remix can work but the hiring pool and ecosystem around Next.js is 3 to 5x larger, which matters when you hire your third engineer.
Database: PostgreSQL on Supabase or Neon. Both give managed Postgres with migrations, connection pooling, automatic backups, and pricing under 50 dollars per month until you scale past 100GB. Skip NoSQL; Postgres handles 95 percent of SaaS data needs cleanly and migrations stay simpler when everything is in one database.
Auth: Clerk or Auth.js (formerly NextAuth). Rolling your own auth in 90 days steals 4 to 8 weeks for something that does not differentiate the product. Clerk handles email/password, social login, SSO, and user management UI; 0 dollars up to 5,000 monthly active users then usage-based. Auth.js is the open-source alternative that is free forever but requires more integration work. Billing: Stripe with Stripe Checkout and Customer Portal - the combination handles subscriptions, upgrades, cancellations, and invoicing without building custom UI. Deployment: Vercel for frontend, Railway or Render for any Node.js or Python backend services.
Thirteen weeks, four phases.
Weeks 1 to 2: setup and auth. Stack provisioning, repo structure, Clerk integration, basic UI shell with navigation and account pages, CI and deployment pipeline. By end of week 2 a logged-in user should land on an empty dashboard.
Weeks 3 to 6: core workflow. The one feature that defines the product. If the product is a team-collaboration tool, the real-time editing and sharing flow. If it is a CRM, the contact-and-deal-management flow. This is the bulk of engineering time because getting one workflow right - with good defaults, clear empty states, and a complete user-journey loop - takes longer than shipping five half-built workflows.
Weeks 7 to 9: billing plus second feature. Stripe integration (Checkout + Customer Portal + webhook handling) takes a full week done properly. In parallel, the secondary workflow or the one integration users have asked for. By end of week 9, a user can sign up, pay, use the core workflow, and have enough surrounding features to justify the subscription.
Weeks 10 to 12: polish, onboarding, and beta launch. Onboarding flow that walks users from first-login to first-value-delivered in under 5 minutes. Empty-state screens with helpful CTAs. Email notifications via Resend or Postmark. Basic analytics (page views, feature usage) via PostHog or Mixpanel. Week 13: design-partner launch to 10 to 50 customers from the founder's network.
Three to five, high trust.
The minimum viable team: one full-stack engineer who can build both frontend and backend, and one product-oriented founder who makes scope and UX decisions in hours rather than weeks. Below this shape, someone becomes the bottleneck. A team of one rarely ships a SaaS MVP in 90 days because context-switching between strategy, coding, customer conversations, and billing setup adds up to more hours than 90 days contains.
The comfortable team: two engineers, one designer who can also write clean product copy, one founder-product-lead. Four people, high trust, all co-located or in the same time zone with tight async cadence. Each person makes decisions in their own domain without multi-layer approval; the founder-product-lead arbitrates when cross-domain decisions need a call.
Above five people, coordination cost starts to eat into velocity. Standups get longer, pull requests queue up waiting for reviews, design handoffs introduce friction. The 90-day timeline is built around a team small enough to have dinner together on a Friday. Agency engagements can run larger teams by segmenting work across parallel streams (one team on frontend, one on backend, one on integrations), but that pattern requires mature agency management and typically runs 120 to 180 days rather than 90.
Weeks 13 to 52, real growth.
Weeks 13 to 26: design-partner phase. Launch to 10 to 50 customers from the founder's network, typically at a discounted price (free, or 50 percent off first year). Track every week: which features they use, which they ignore, where they hit friction, what they say they want next. Make product decisions weekly based on this data. Week 20 to 24, you should have enough signal to know whether the product has product-market-fit pulse or whether the thesis needs adjusting.
Weeks 27 to 52: scale phase. If PMF signals are positive (retention above 40 percent at day 30, willingness to pay full price, organic referral starting), move to paid acquisition. Google Ads for search-intent B2B SaaS, LinkedIn Ads for high-ACV B2B, content marketing and SEO for longer-term compounding. Add the second integration that customers most asked for. Begin SOC 2 compliance work if enterprise customers are in the pipeline; it takes 4 to 8 months and is a gating requirement for most companies above 200 employees.
Year 2 and beyond: the features cut in the three-cut pass get added back if users requested them. Most of the original cut list never returns because nobody missed it. The product that ships in year 2 looks very different from what the founder imagined in week 0, which is how it should be. Related reading: SaaS pricing page design, SaaS onboarding metrics, and React vs Next.js for SaaS.
Six answers.
Is 90 days actually enough to ship a SaaS MVP?
Yes, with scope discipline. A 90-day MVP covers: one core workflow, authentication, basic billing (via Stripe Checkout), one integration if needed, and a minimal onboarding experience. It does not cover: multi-tenancy edge cases, SSO, SOC 2, native mobile apps, advanced analytics, or features that serve 10 percent of potential users. Most failed 90-day attempts fail because scope slipped past these boundaries. The teams that succeed keep the product cut narrow and iterate after launch based on real usage data. A 120-to-180-day build is often better than a 90-day build with compromised quality; the choice is about whether the market window demands speed or allows for depth.
What does a 90-day SaaS MVP cost?
Three budget ranges. Solo founder with technical cofounder: 0 dollars out-of-pocket for the build plus 200 to 500 dollars per month in infrastructure and SaaS subscriptions. Small agency team (3 to 5 engineers): 80K to 180K depending on scope and hourly rates, typically 3 engineers plus a product designer for 12 to 13 weeks. Full-service agency engagement with discovery, design, development, and launch support: 150K to 350K for enterprise-ready output. The best founder-led builds cost near zero; the best agency-led builds cost around 120K. The gap reflects the tradeoff between founder-time-invested and capital-invested.
What stack should I use for a 90-day SaaS MVP?
Boring, well-supported, compound-friendly picks. Frontend: Next.js (React) for the web app, TypeScript, Tailwind CSS. Backend: Next.js API routes for simple APIs; Node.js (Fastify or Express) or Python (FastAPI) for heavier backend work. Database: PostgreSQL hosted on Supabase or Neon for managed Postgres with migrations. Auth: Clerk or Auth.js for authentication because rolling your own auth in 90 days is expensive. Billing: Stripe with Stripe Checkout and the Customer Portal for subscription management. Deployment: Vercel for the frontend, Railway or Render for any heavier backend services. This stack has about 80 percent overlap with what most successful 2024-2026 SaaS launches used; deviations should have specific reasons.
What team do I need to ship in 90 days?
Minimum viable team: 1 full-stack engineer and 1 product-oriented founder who can do design decisions and product scoping. Comfortable team: 2 engineers, 1 designer who can also write copy, 1 founder-product-lead. Above that, coordination cost starts to eat into velocity. Fewer than that and one person becomes a bottleneck. The team shape matters more than team size: co-located (or fully remote with tight async cadence), all willing to switch between code / design / copy / customer calls, and empowered to make decisions without multi-layer approval. The 90-day timeline compresses to impossible if decisions sit in review queues.
What is the three-cut rule for MVP scope?
First cut: delete 50 percent of the initial feature list. The initial list is always too ambitious because it mixes what must ship with what would be nice. Cut everything that is not required for the core value to be demonstrated. Second cut: delete another 25 percent. On the second pass, cut polish features that do not change whether the product works: advanced search, custom themes, role-based permissions, integrations beyond the first one. Third cut: delete another 10 to 15 percent. This is the hardest cut because what remains all feels essential; push yourself to cut the least-critical thing, trust that it can be added in weeks 14 to 20. The remaining 10 to 15 percent is the true 90-day MVP.
What launches after the 90-day MVP?
Weeks 13 to 26: the growth phase. Launch with 10 to 50 design-partner customers from the founder's network, iterate based on their usage data weekly, add the features they actually use. Weeks 27 to 52: the scale phase. If the MVP shows product-market-fit signals (retention above 40 percent at day 30, willingness to pay, organic referral), move to paid acquisition (Google Ads, LinkedIn Ads for B2B SaaS), add a second integration, begin SOC 2 compliance for enterprise-ready customers. Year 2 and beyond: the features cut during the 90-day pass are added back if users request them. Most features cut in the three-cut rule never get re-added because users did not actually want them.
MVP is scope discipline.
Our SaaS engagements deliver the 90-day MVP from scope through launch: product definition, stack selection, 13-week build, design-partner launch. Scoped quote in 48 hours.