If you have asked three developers what your SaaS idea costs, you have probably received three numbers that are not remotely close to each other. One said ₹3 lakh. One said ₹15 lakh. One quoted in dollars and made you do maths.
That spread is not because two of them are trying to overcharge you. It is because "build me a SaaS product" is not a specification, and each of them silently filled the gaps with different assumptions.
This is an honest breakdown of where the money actually goes, written for the person paying rather than the person coding.
What "MVP" Should Actually Mean
The phrase has been stretched until it means almost nothing. Two useful definitions, and they cost very different amounts:
- A prototype — something you show investors or test on ten friendly users. It may not handle payments, may break under load, and may have a hardcoded admin login. Perfectly valid. Cheap.
- A first commercial version — real users pay real money and expect it to work on a Tuesday afternoon without you watching. This needs authentication done properly, billing, an admin panel, error handling, backups and monitoring.
Most founders say "MVP" and mean the second, then compare quotes from developers who priced the first. That single ambiguity explains most of the variance you have seen.
The Cost Breakdown
For a first commercial version, here is roughly where a budget goes. Percentages are more stable than absolute numbers across project sizes.
- Authentication and user management — 10%. Sign-up, login, password reset, email verification, roles and permissions. Unglamorous, and one of the most expensive things to retrofit badly.
- The core workflow — 35%. The thing your product actually does. The only part most founders picture when estimating.
- Billing and subscriptions — 15%. Razorpay or Stripe integration, plans, upgrades, downgrades, failed payments, invoices. Deceptively large: the happy path takes a day, the other paths take two weeks.
- Admin panel — 15%. How you run the business — view users, fix data, issue refunds, see what is happening. Founders cut this first and regret it within a month, because the alternative is messaging a developer every time a customer has a problem.
- Infrastructure and deployment — 10%. Hosting, domains, SSL, CI/CD, backups, monitoring, error tracking.
- Design and UI polish — 15%. Ranges from "use a component library" to a full custom design system. Highly compressible if budget is tight.
Realistic 2026 Numbers
Indian market, working with an experienced individual developer rather than an agency:
- ₹4–7 lakh — a focused single-workflow SaaS. One user type, one core feature done well, standard billing, basic admin. 8–10 weeks.
- ₹8–15 lakh — multiple user roles, a couple of third-party integrations, a real admin panel, custom design. 12–16 weeks.
- ₹15–30 lakh — multi-tenant B2B with organisations and team permissions, an API for customers, meaningful compliance requirements. 4–6 months.
- ₹30 lakh+ — you are no longer describing an MVP. That is a product, and it should be funded and staffed as one.
Agencies typically run 1.5× to 2.5× these figures. Sometimes that is worth it — more capacity, continuity if someone leaves, dedicated QA. Often you are funding a sales team.
The Five Things That Quietly Multiply Cost
When a project comes in far over the initial estimate, it is nearly always one of these:
- Scope that grows weekly. Each addition sounds small. Twelve small additions are a second project. This is the number one cause, by a wide margin.
- Unclear requirements. If it is not written down before building starts, it gets built twice — once to your developer's interpretation, once to yours.
- Integrations with systems you do not control. A poorly documented third-party API can absorb two weeks with nothing visible to show.
- "Just make it look like Stripe." Custom design work at that standard is a specialist discipline and prices accordingly.
- Slow feedback. If review takes a week, momentum dies and context gets rebuilt each cycle. Founders underestimate how much their own responsiveness affects the invoice.
What You Should Never Cut
Cutting features is healthy — that is what "minimum" means. Cutting these is not:
- Authentication done properly. A security incident with paying customers is existential, and "we will fix auth later" never survives contact with a real user base.
- Your data model. Restructuring a database after launch means migrations, downtime and touching every part of the codebase. This is the single most expensive thing to change later.
- Backups you have actually tested. An untested backup is a belief, not a backup.
- Basic monitoring. Without it you find out about outages from customers, which costs more than the monitoring ever would.
How to Get Comparable Quotes
If you want numbers you can actually compare, give every developer the same four things:
- A written list of features, split into "launch" and "later"
- Who the user types are and what each can do
- Any integrations that are non-negotiable
- Your realistic budget range
That last one causes discomfort, but withholding it wastes everyone's time. A developer who knows you have ₹6 lakh will tell you what is achievable for ₹6 lakh. A developer guessing will either overspecify and lose the work, or underspecify and disappoint you later.
The Honest Summary
A serious SaaS MVP in 2026 costs somewhere between ₹4 lakh and ₹15 lakh depending on how much it genuinely needs to do. Anything dramatically below that is a prototype. Anything dramatically above is scope that has not been challenged hard enough.
The largest lever on your final cost is not who you hire. It is how clearly you define what you are building before anyone starts.
If you want a straight answer on what your specific idea would take, here is how I approach SaaS builds — or just describe what you have in mind and I will tell you honestly whether I am the right fit.