Skip to content
Faizan Ali home
Menu

How to build an MVP for your startup


Updated , 8 min read

Short answer

To build an MVP, write down the one problem you're solving and who has it, cut your idea down to the single main task that solves it, build only that with login and payments, then get it in front of real customers quickly. An MVP is a way to find out whether people will pay, not a smaller copy of your final product.

Most software projects don’t fail because of bad code. They fail because they’re built before anyone really understands the problem. I’ve seen this more than any other kind of failure, and an MVP is meant to protect you from it.

Here’s how I’d go about it, after building more than eight products with founders.

Step 1: Write down the problem

Before thinking about features, finish this sentence: “[This type of person] struggles with [this specific problem], and today they deal with it by [what they do now].”

What they do now is your real competition. If people aren’t doing anything about the problem today, not even something clumsy in a spreadsheet, it may not bother them enough to pay you.

Step 2: Talk to people before you build

Ask potential customers how they handle the problem now, what it costs them and what they’ve already tried. Don’t sell. Just listen for the same complaints coming up again and again, and for signs that they already spend money or time on it.

Step 3: Cut it down to one main task

List every feature you want. Then ask: what’s the least someone needs to do to get the main benefit once? That’s your main task, and that’s your MVP.

For Grade’it, a verified review platform I built, the whole point was proving that a reviewer was a real customer. That check had to work properly. Almost everything else could be simpler.

What to leave out

  • Team accounts and complicated permissions
  • Phone apps, if a website that works well on phones will do
  • Detailed reports and dashboards
  • Most integrations with other tools
  • Any feature that starts with “it would be nice if”

What to keep

  • Secure login. Use a well-tested login service rather than building your own.
  • Payments, if you’re testing whether people will pay. One real payment tells you more than a hundred people saying they’d buy.
  • Backups and error monitoring. They take little time to set up, and they save you when something goes wrong.
  • Simple analytics. You’re building to learn, so make sure you can see what people actually do.

Step 4: Choose boring tools

Use well-known tools that other developers can work with later. I usually pick Python with Django or FastAPI, React or Next.js, and PostgreSQL. None of it is exciting, and that’s the point. Your product should be the interesting part.

Step 5: Decide how to build it

Option Good when Watch out for
AI builder (Lovable, Bolt) Testing an idea cheaply, even if you’re not technical Security and data problems once real users arrive
No-code (Bubble, Webflow) Simple products you’ll manage yourself Limits and rising costs as you grow
One senior engineer A real product with payments or tricky logic Finding someone you trust, and relying on one person
An agency You need lots of specialists at once Higher cost, and less contact with the people building it

You can also mix them. Build a quick version with AI to see if people want it, then have an engineer build properly, or make safe, the parts that work.

Step 6: Launch before you feel ready

Pick a launch date when you start. If you fall behind, cut features rather than moving the date. Getting a working version you can click through every week makes it obvious early when something is going off track.

Step 7: Watch the right numbers

Sign-ups don’t tell you much. Watch whether people finish the main task, whether they come back and whether they pay. Talk to the people who stopped using it, not just the ones who stayed.

Should your MVP include AI?

Only if AI is what makes the product valuable, or if it clearly makes the main task better. If you do include it, plan from day one for wrong answers and for what each use costs you. When I built SAP Assistant, charging by credits was part of the plan from the start, because every question costs real money to answer.

Where people go wrong

  • Building in private for six months. You learn nothing until someone else uses it.
  • Adding features instead of finding customers. New features feel like progress, and they’re easier than sales calls.
  • Treating the MVP as throwaway. “Minimum” is about how much you build, not how well. Build the parts you keep properly, because you’ll probably use them for a long time.

Common questions

What should an MVP include?

One main task done well, sign-up and login, a way to pay if you're testing whether people will pay, and simple admin tools so you can support your first customers. Everything else can wait until customers ask for it.

How long should it take to build an MVP?

A focused SaaS MVP usually takes 3 to 5 weeks with an experienced engineer. If your plan takes six months, it probably isn't small enough yet.

Can I build an MVP with AI?

Yes, and for testing an idea it's often the right place to start. AI tools are great for prototypes. Before real users and real payments arrive, get someone experienced to check permissions, how your data is stored and how payments are handled.

Should I hire a developer or find a technical co-founder?

A technical co-founder makes sense if software is the heart of the company for years to come and you've found someone you trust. Hiring an engineer is usually faster if you need to test the idea first. Plenty of founders hire for the MVP and look for a co-founder once they have paying customers.

Got a question this guide doesn’t answer? Email me.

hi@faizanali.dev

Get new guides by email

New guides on building software and using AI, about twice a month. No spam, and you can unsubscribe anytime.

Tell me what you’re trying to solve.

Twenty free minutes. You’ll leave with a clearer scope, an honest opinion on what to build first, and no pressure to work with me.