James Maskell

Minimum Viable Products

04 December 2012

Over the past year, since reading The Lean Startup, I've been increasingly focused on following a strategy of launching Minimum Viable Products (MVP). By building MVPs it is possible to validate ideas without sinking huge amounts of time and money in to more fully featured products that may never take off. Building MVPs also encourages you to interact with your customers and discover exactly what they want - rather than making assumptions that are often incorrect.

Recently I've seen a few posts criticising the idea of MVPs and their use. My feeling is that the authors of these posts are sometimes following the theory too rigidly and being too minimal with their products.

The following points detail some of my thoughts on the use of MVPs and the lessons that I have learned over the past year:

Start really small

Your first MVP should be very basic. Possibly as basic as paper wireframes that have been sketched out and put in front of someone. Or a simple web landing page with a newsletter signup form.

With your first MVP the key is to gauge interest in your general idea. Get the message across simply and quickly. See how people react.

Be iterative

Once you've validated the very basics, move on to the next stage. This could be building out your wireframes in to a functioning web page.

Make sure you keep building minimal products. If your email signup page works, don't assume that all of your ideas will. Be disciplined and get in to a smaller release cycle - build a habit.

Test one hypothesis at a time

Make sure you can test your ideas independently. Don't build MVPs testing several different ideas unless you can isolate each hypothesis and find out what works and what doesn't.

No need to release to everyone

If you have a more mature product, release/test with a small set of users first. Iron out the bugs and develop with feedback.

A/B testing is easy enough to set up but you can also learn a lot by sitting down with your customers and watching how they interact with your product.

Listen to the results and your customers

Pay attention to the results of your testing. If you can't validate an idea make some changes and try again. Sit down and talk with your customers - find out if they understand what you're doing and see how they react.

Don't be scared of failure - if your product isn't working (and you can't get it to after several attempts) kill it and move on. Focus your limited resources on the things that work.

Be prepared for lots of manual work

One of the recent criticisms of MVPs it that founders felt like they were pushing their customers off a cliff (e.g. they were misleading people and hadn't actually developed the site out yet).

There is a simple solution to this - capture the data required from your customers and process it manually. This might take a bit more work - but it's still relatively simple and you're not throwing away users. This is the viable part of MVP.

Adapt the theory to your own needs

The Lean Startup and MVP theory is not an exact science and not intended as a rigid framework. Study it and experiment - find out what works for you.

Just be careful not to bury your head in the sand and pretend to be minimal when you're not. Take yourself out of your comfort zone - keep releasing products, keep testing and keep failing.


Any comments? Let's discuss on Twitter - I'm @jmaskell.