SaaS development · MVP to scale
Building a SaaS product is not just a matter of clean design and a database. What separates a successful SaaS from an expensive failure are choices made in week one that only become affordable to fix in year two: how you isolate tenants, where you store state, how you wire up payments and how you ship features without breaking existing customers.
MG Software has been building SaaS products in the Netherlands since 2018. We have launched MVPs within 8 to 12 weeks, grown them into platforms with thousands of active users, and rescued the occasional project that derailed elsewhere. What we can share with founders just starting: the difference is in the invisible details.
This page gives an honest look at what SaaS development costs, how long it takes and which decisions belong in which phase. No marketing fluff about disruption and scalability; just how it actually works in practice.
Building a SaaS is not a single project with an end date, but three different projects in sequence. What you need in phase one differs from what you need in phase three, and money spent badly in phase one costs you ten times as much to repair in phase three.
The goal is not a finished product. The goal is proof that customers will pay for it. We build only the core functionality that lets the first ten customers actually solve the problem they have. No branding flourishes, no ten user roles, no self-service onboarding tour.
You now have paying customers and are discovering what they actually use and where they complain. We build what we learned from real customer conversations, not what a founder thinks the market needs. At the same time we start repairing technical debt you deliberately took on in the MVP phase.
Implement multi-tenancy properly, automate billing, set up a team-account system and build in analytics. This is where many SaaS projects stall because the codebase was not designed for hundreds of customers. Early investment in a scalable architecture pays off here.
A SaaS is never finished. We offer retainers with a fixed monthly budget for security patches, feature work and dependency upkeep. The alternative is a platform that becomes unmaintainable after three years and needs rebuilding.
An MVP for a B2B SaaS realistically falls between 25,000 and 60,000 euros. Below that range it is rarely serious enough to attract paying customers; above it you are spending money on things that are not relevant at this stage. The variables that determine the price within that range are the number of screen flows, whether real-time functionality is needed, whether multi-tenancy is required from day one and how complex the business logic is.
A PMF phase, the 3 to 6 months after launch, typically costs an additional 30,000 to 80,000 euros depending on how many iterations are needed and how much technical debt from the MVP phase needs paying down. A typical engagement runs on a quarterly budget and scales with the customer base.
A scalable version with multi-tenancy, billing through Stripe or Mollie, team accounts, audit logs and analytics dashboards typically lands between 60,000 and 150,000 euros on top of what is already invested in MVP and PMF phases. Summed together, an average SaaS from idea to scalable platform sits between 120,000 and 290,000 euros of investment, spread over one to two years. That sounds like a lot, but it is conservative compared to what a team of two to three senior developers in salaried positions costs over the same period.
What these numbers do not include are marketing, sales, support and infrastructure costs. Budget 200 to 1,500 euros per month for cloud hosting on an average B2B SaaS, scaling with customer volume. External services like Sentry, a transactional database, email delivery via Resend or Postmark, and analytics add up to 200 to 600 euros per month. This is recurring, not one-off.
These decisions look small when you make them and expensive to undo later. We help you make them deliberately rather than by accident.
Do you store all customers in a shared database with tenant_id columns, or give each customer their own schema or database? Both work, both have costs and risks. For most B2B SaaS we recommend a shared database with row-level security; for enterprise customers with strict data isolation requirements you can switch later.
NextAuth, Clerk, WorkOS or self-hosted on Supabase: each choice has consequences for enterprise features, price at scale and how much work you spend on auth maintenance in year two. We work with all options and help you choose based on your growth plan.
Stripe as the single source of truth for subscriptions works until you want to add Mollie for SEPA-only customers or a Belgian partner. An own subscription table with billing providers as plug-ins costs more upfront but gives flexibility later.
Emails, exports, AI calls: handle them directly in an API call or via a job queue? For low-volume MVPs synchronous is fine; the moment a customer complains about a 30-second wait you need a queue. We build this pragmatically: not too early, not too late.
Without feature flags you have to roll every new feature out to everyone at once. With feature flags you can give beta features to ten customers, A/B test on pricing and disable problematic features without doing a rollback. Setting up takes a day, not having them costs you months over the lifetime of the platform.
Who changed which customer data, when and from where? For enterprise customers a requirement, for you a lifesaver during security incidents and bug investigations. Adding it after the fact usually requires rewriting your mutation logic; building it in from day one adds a few hours per feature.
We do not build consumer apps with social-feed mechanics. Other agencies are better at that, and the monetisation models are fundamentally different from B2B SaaS. We say that honestly in a first conversation if it turns out that is what you are looking for.
We do not build blockchain solutions, Web3 wallets or NFT marketplaces. In our experience the vast majority of those projects look for technical solutions to problems that do not exist. If you have a legitimate use case where blockchain genuinely is the right answer, we refer you on.
We do not build without a discovery phase. A founder who says the scope is already clear and we should just start, we know from experience will come back within three weeks on decisions that were made awkwardly in week one. A proper discovery of one to two weeks is not delay but an investment that always pays back.
We do not work on a fixed price for the entire engagement. We do not know in week one everything we will discover by week ten; neither do you. A fixed price tempts both parties to scope creep on one side and cutting corners on quality on the other. We work on a time-and-materials basis with clear sprint budgets and an agreed-upon ceiling.
Not the only correct choices, but the ones that work well for 90 percent of our SaaS projects without unpleasant surprises.
Next.js 16 / React 19
Frontend and server actions
PostgreSQL / Supabase
Database with RLS, auth and storage
Stripe
Subscription billing and invoicing
Mollie
iDEAL and SEPA for NL-only customers
Resend / Postmark
Transactional email
Sentry
Error monitoring and performance
PostHog
Product analytics and feature flags
Vercel / Railway
Deployment, with fallbacks per use case
Related services and pSEO content covering specific decisions in detail.
For a B2B SaaS MVP we plan 8 to 12 weeks from the moment scope and design are locked. The variables are the number of screen flows, whether real-time functionality is needed and how complex the business logic is. Faster timelines are possible but typically mean leaving features out that have to come back later. Slower than 12 weeks is often a sign that the scope is too broad for an MVP.
A serious B2B SaaS MVP in our experience costs 25,000 to 60,000 euros. Below that we make too many compromises on quality, security or architecture that you pay for expensively in year two. Above it you spend money on things that are not relevant in this phase, like self-service onboarding tours or advanced analytics. Consumer SaaS has different ranges; we are less often the right partner there.
Three phases: MVP phase of 8 to 12 weeks to prove customers will pay, PMF validation of 3 to 6 months in which you learn what they actually use, and scale phase from month 9 onwards in which multi-tenancy, billing automation and analytics are built in. Total investment from idea to scalable platform typically falls between 120,000 and 290,000 euros, spread over one to two years.
Not always. For B2B SaaS where you expect ten larger customers in year one, a simpler per-customer architecture is fine and you migrate to multi-tenancy in phase three. For SaaS where you expect dozens or hundreds of self-service signups from day one, multi-tenancy needs to be right immediately, because retrofitting means redesigning the entire data layer. We advise based on your real growth plan, not best-case scenarios.
Stripe is stronger for international customers, subscription management and developer experience. Mollie is stronger for Dutch customers with iDEAL and SEPA, and has a simpler dashboard for non-technical founders. For SaaS with international ambitions we typically recommend Stripe, with Mollie as an optional addition for specific payment methods. For SaaS serving primarily Dutch customers, Mollie can cover the whole need.
Occasionally, but reluctantly. Consumer SaaS has fundamentally different monetisation mechanics, growth loops and design requirements than B2B SaaS. We are better at B2B and say so honestly in a first conversation. For consumer projects we refer you to agencies that specialise in that.
Then we often play that role in practice. We act as a technical sounding board for architecture choices, security considerations and long-term implications of decisions. For non-technical founders it is important that you have the business side and customer contact firmly in hand, because we support you technically but cannot validate your market for you. We are a development team, not an incubator.
In the MVP phase we deliberately take on technical debt to gain speed. We document what we postponed and why. In the PMF and scale phases we pay that debt down, tied to moments when you are touching the code anyway for new features. What we do not do are shortcuts that fundamentally undermine the architecture, like skipping multi-tenancy support where it will clearly be needed later. That is not debt but a rebuild promise.
You do. We set up all infrastructure on your own accounts (Vercel, Supabase, Stripe, AWS where needed), with you as owner and MG Software as developer with scoped access. Repositories live on your GitHub organisation. Domains you register yourself. No vendor lock-in, no hidden dependencies on our infrastructure. This is also practically important: if we are not responsive enough at some point, you can work on the infrastructure without our involvement.
Through retainers starting at 2,500 euros per month for small maintenance, security patches and dependency updates, scaling with how much feature work per month. An average growing SaaS has a retainer of 5,000 to 15,000 euros per month to scale alongside customer requests, bug fixes and gradual expansion. For enterprise customers with SLA requirements we plan proactive monitoring and shared on-call.
An hour is enough to determine whether your idea is technically feasible within your budget and timeline. We share an honest estimate at the end of the conversation, even when it differs from what you hoped to hear.