Mark Pollmann's blog

Notes on How to Plan your MVP by Michael Seibel

· Mark Pollmann

This is a short talk Michael Seibel gave at Startup School on his approach to building an MVP (Minimum Viable Product) and getting your first users before launching your startup.

** An MVP is the thing with which you can ask your first potential users if you provide any value at all. **

“How to find your first customers?”
-> Wrong question because if you found the problem you want to solve the customers are apparent

Launch something bad quickly

Get anyone to use your MVP. Many founders have never shown anything to potential users before the startup fails.

Get feedback!

“Feedback on this little thing we built is useless”
-> No, The fully fleshed out thing you would like to build might not be the one that customers want.

Iterate until product actually solves the problem.

Build it in weeks not months. doesn’t have to be software, could be landing page and spreadsheet or click dummy.

Focus on small set of users and their most pressing problems.

Learning from users while having a MVP is easier than without one.

Examples of Lean MVPs

Most of the time a lean MVP is all you need.


Airbnbs first website

When they started they had no payments, no map view and only a part-time CTO.


twitchs first website

Twitch started with just one channel, a low-res video blog with no games.


stripes first website

Stripe started out without any bank deals and very few features. The founders actually came to your office to integrate the software for you.

Examples of Heavy MVPs

Insurance and banking startups need to comply with regulation

Hardtech like building rockets can’t be done in a short time, same as biotech.

In that case a website might be a good idea to explain potential customers what you’re about.


Launches are not special. Nobody remembers when exactly twitter or facebook launched. Make your launch the day somebody actually buys your product.

Hacks for building an MVP quickly

Timebox your spec. Let’s say 3 weeks.

Write your spec. It’s easy to lose track of what it was you wanted to build exactly and add features after features.

Cut your spec. If, after a week, you notice you won’t finish your MVP in this timebox cut features.

Above all: don’t fall in love with your MVP


Never ask users for features. That’s not their job. Their job is to give you problems. If they make suggestions for features ask why they would like this feature and drill down what their problem is.

You know when you have product-market fit when you don’t know how to serve all the customers running in your door.