The SaaS MVP playbook.
Ninety days from blank repo to paying customers. Scope, stack, pricing page, onboarding, framework choice. The decisions that compound across the next five years.
Ninety days. One core workflow. Paying customers.
A SaaS MVP ships in 90 days when scope is one core workflow plus auth, billing, and the minimal admin to support paying customers — not the team's ten-feature wishlist. Stack: Next.js plus Postgres plus Stripe plus Vercel for 80 percent of B2B SaaS in 2026. Pricing: three tiers, anchor the middle. Onboarding: instrument activation rate, time-to-value, day-30 retention. Cost: USD 30K to 150K depending on scope and team. Brands that compress aggressively to 90 days outship 9-month efforts because they validate market demand before fully scaling. Published by Prasun Anand.
What ships in the right 90-day MVP?
One core workflow that solves a specific problem for a specific user. Authentication. Billing. The minimal admin to support paying customers. That is it. The MVP is not feature-light; it is decision-narrow. The team picks one user persona, one workflow that creates value for that persona, and ships only what those users need to convert and stay.
The right scope decisions. One workflow: what does the user actually do in your product? Not "managing customers" — what is the specific sequence: import contacts, set a goal, get a recommendation, take action. Pick one. One user persona: not "small businesses" — describe one human: she runs a 5-person remote team, lives in spreadsheets, has been frustrated with X for 18 months, has budget authority for tools under USD 200 per month. The first ten: who specifically signs up? Have ten people already committed before you write the first line of code; if not, the demand is hypothetical and the build is research.
The wrong scope decisions. Building "a platform for X" rather than "a tool that does Y for Z user." Trying to serve multiple personas — small business AND enterprise AND solopreneur — with V1. Building features the team thinks users will need rather than features the first ten users have explicitly asked for. Adding integrations, exports, dashboards, and admin tooling before the core workflow has been validated with real users.
The brutal compression rule. If you cannot describe what the MVP does in one sentence to a stranger and have them understand it, the scope is still too broad. Brands that ship the right MVP in 90 days do so because the core workflow is so narrow it forces aggressive prioritization. The deeper 90-day plan with the week-by-week build sequence is in SaaS MVP in 90 days.
Next.js plus Postgres plus Stripe plus Vercel.
For 80 percent of B2B SaaS MVPs in 2026, this combination ships fast, scales further than the MVP ever needs, and hires from the largest available talent pool. Each piece chosen for time-to-ship, not theoretical purity.
Next.js for the application. Server-side rendering for the marketing pages (which need SEO from day one). Server actions for backend logic without standing up a separate API service. Mature TypeScript ecosystem. The largest hiring pool of any modern web framework. The free Next.js documentation and the open-source ecosystem mean a single engineer can ship a production-quality SaaS app without enterprise scaffolding.
Postgres for the database. Battle-tested. Scales beyond where any MVP will go. Excellent integration with the modern hosted-Postgres providers (Neon, Supabase, Vercel Postgres). JSON columns when you need flexibility, structured columns when you need rigor.
Stripe for billing. Subscriptions, usage-based billing, taxes, dunning, customer portal — all handled. Stripe Billing is the standard for a reason; trying to build billing primitives yourself is a quarter of engineering time you do not have. Stripe's tax handling alone (Stripe Tax) saves months of compliance work for international SaaS.
Vercel for hosting. Zero-config deploys from Git. Edge functions when you need them. Generous free tier for early-stage. The integration with Next.js is tight enough that you spend zero engineering time on deploy infrastructure during the MVP phase. The deeper React vs Next.js comparison for SaaS is in SaaS React vs Next.js.
Three tiers. Anchor the middle.
Three pricing tiers with the middle tier designed to be the obvious choice for the target customer. The lowest tier is a real entry-level option, not a fake decoy. The highest tier is expensive enough to anchor the middle as the value pick. The pricing page is the single highest-leverage marketing surface in SaaS.
The tier structure. Starter / Free / Trial: real entry option for genuine first-time use. Capped feature surface, capped usage, but a working product. The job here is conversion to a paid tier within 14 to 30 days. Pro / Growth / Standard: the middle tier. Designed to be the obvious choice for your typical target customer. Priced at what your most-engaged early adopters are willing to pay — usually USD 49 to 299 per month for solo B2B, USD 99 to 999 per month for team B2B. Business / Enterprise / Premium: the high tier. Designed to anchor the middle by appearing visibly more expensive while offering enterprise-class features (SSO, audit logs, custom contracts, SLA). Few customers buy this tier; that is fine — its job is anchoring, not volume.
Annual versus monthly. Offer both, with annual at 15 to 20 percent discount. The annual price funds growth (cash up front) and reduces churn (lower-friction renewal moment). Most SaaS brands shift to annual-default pricing display once they reach product-market fit.
Test pricing on real customers, not internal debate. The right pricing for your product is what your best target customers are willing to pay; that is empirical. Test-mode the pricing on the first 10 to 20 customers, watch conversion and churn, adjust. The full pricing-page design playbook covering layout, anchoring, and trial conversion is in SaaS pricing page design.
Three onboarding metrics predict long-term retention.
Activation rate (share of signups reaching the aha moment in first session, healthy 40 to 60 percent). Time-to-value (median minutes between signup and aha moment, healthy under 10 minutes). Day-30 retention (share of activated users still active 30 days post-signup, healthy 40 to 60 percent). Together these predict long-term retention more reliably than any vanity metric.
Defining the aha moment. The aha moment is the specific in-product event where a user experiences the value proposition. It must be a concrete tracked event, not a feeling. For a project management tool: "user added a second task." For an analytics tool: "user viewed first chart with their data." For a writing tool: "user saved first document." If you cannot point to a specific event, you have not defined your aha moment yet — and you cannot optimize what you have not defined.
Instrumentation. Track every step from signup to aha moment as a discrete funnel stage in product analytics (Amplitude, PostHog, Mixpanel). The funnel exposes drop-off points where users abandon. Each drop-off is a specific UX problem to solve — empty state confusing, form too long, value not yet clear. Optimize the funnel iteratively; activation rate moves from 25 percent to 50 percent over three months of focused work in most cases.
Retention beats acquisition. A SaaS with 30 percent activation and 40 percent day-30 retention spends every marketing dollar twice — once to acquire, once to replace the customers who churned. A SaaS with 60 percent activation and 60 percent day-30 retention compounds. The difference is engineering and product work, not marketing spend. The full onboarding-metrics depth including the funnel-optimization patterns is in SaaS onboarding metrics.
USD 30K to 150K. What changes in the range?
USD 30K is one full-stack engineer building a focused single-workflow product on a standard stack in 8 to 12 weeks. USD 60K to 90K is a small team (two to four engineers) shipping a polished V1 with marketing site, billing, basic analytics, and admin tools in 12 to 16 weeks. USD 100K-plus adds custom integrations, advanced analytics, multi-tenant architecture, longer build window.
What drives the range. Team size and seniority — a senior solo engineer at USD 250 per hour for 10 weeks lands USD 100K just on labor. Custom integrations — every external system (Stripe is included; Salesforce, HubSpot, custom enterprise APIs are not) adds two to four weeks. Multi-tenancy — a true multi-tenant SaaS architecture with tenant isolation, custom domains, and per-tenant configuration adds three to six weeks. Advanced UI — a SaaS app that is itself a creative tool (whiteboard, design tool, complex visualization) takes longer than a CRUD app with forms.
What does NOT belong in MVP cost. A custom design system. Performance optimization beyond reasonable defaults. SOC 2 compliance work (defer to post-PMF). Enterprise SSO if you have no enterprise customers (defer). Custom analytics dashboards beyond what off-the-shelf tools provide. Each of these adds two to eight weeks of work that contributes nothing to the validation question MVPs are designed to answer.
Below USD 30K, scope is too narrow to be a real SaaS. Above USD 150K, you are no longer building an MVP but a fully-featured V2. The MVP question is "will users pay for this workflow?" — once that is answered yes, the next question becomes "how do we scale?" and the budget shape changes.
Six answers.
Can a SaaS MVP really ship in 90 days?
Yes for the right scope, no for the wrong one. The right scope: one core workflow that solves a specific problem for a specific user, plus authentication, billing, and the minimal admin to support paying customers. The wrong scope: an internal team's ten-feature wishlist that mistakes feature breadth for product. Brands shipping in 90 days typically have a clear single workflow, a defined first user persona, the first 10 paying customers already identified, and a team of two to four engineers focused exclusively on the build. Brands shipping in 9 months are usually still scoping in week 12.
What stack should a SaaS MVP use?
Next.js plus Postgres plus Stripe plus Vercel for 80 percent of B2B SaaS MVPs in 2026. Why Next.js: server-side rendering for SEO on the marketing side, server actions for backend logic, mature TypeScript ecosystem, large hiring pool. Why Postgres: battle-tested, scales further than most MVPs ever need, excellent Vercel + Neon + Supabase integration. Why Stripe: every commercial concern (billing, subscriptions, taxes, dunning) handled. Why Vercel: zero-config deploys from Git, edge functions, generous free tier. The deviation cases: heavy real-time (consider WebSockets-first stack), AI-native (consider Python plus FastAPI for ML), or platform extensions (Shopify apps need Remix or Express).
Should I use React or Next.js for a SaaS app?
Next.js for new SaaS builds. Plain React (Create React App, Vite) is appropriate for prototypes and internal tools, but lacks production-quality SSR, routing, server actions, and image optimization out of the box. Next.js gives you all of those plus middleware, edge functions, and a deploy target that scales. The marketing site (which an SaaS app needs for SEO) is dramatically easier on Next.js than on plain React. Vue plus Nuxt is a defensible alternative; SvelteKit too. The thing to avoid: shipping a plain-React SaaS where the marketing pages need to rank but cannot SSR.
How should I price a new SaaS product?
Three pricing tiers, anchor on the middle. Make the middle tier the obvious choice for the target customer; make the lowest tier a real option (not a fake decoy) for genuine entry-level use cases; make the highest tier expensive enough to anchor the middle tier as the value pick. Price the middle tier at what your most-engaged early adopters are willing to pay — typical range USD 49 to 299 per month for B2B SaaS, USD 99 to 999 per month for B2B with team seats, USD 9 to 39 per month for prosumer. Annual pricing at 15 to 20 percent discount funds growth and reduces churn. Test pricing on real customers, not on internal debate.
What onboarding metrics matter for a SaaS MVP?
Three core metrics. Activation rate: the share of signups who reach the first meaningful value moment ('aha moment') within their first session. Healthy: 40 to 60 percent. Time-to-value: median minutes between signup and aha moment. Healthy: under 10 minutes. Day-30 retention: share of activated users still active 30 days post-signup. Healthy: 40 to 60 percent. Together these three predict long-term retention more reliably than any vanity metric. The instrumentation: define the aha moment explicitly (it must be a specific event, not a feeling), track it via product analytics from day one, optimize the funnel relentlessly.
How much does a SaaS MVP cost to build?
USD 30K to 150K depending on scope, team, and stack choices. The lowest end is a single full-stack engineer building a focused single-workflow product on a standard stack in 8 to 12 weeks. The mid range is a team of two to four engineers shipping a more polished V1 with marketing site, billing, basic analytics, and admin tools in 12 to 16 weeks. The high end adds custom integrations, advanced analytics, multi-tenant architecture for enterprise, and a longer build window. Below USD 30K, scope is too small to be a real SaaS; above USD 150K, you are no longer building an MVP but a fully-featured V2.
Ninety days. Paying customers.
More from Priya.
Senior PR & Comms Manager · New-york · 15 posts on the site
Published .