When building a Minimum Viable Product (MVP), the faster you learn, the better the final product. But how do you learn quickly (i.e. “fail fast”) without turning off your customers?

At streaming music service Spotify, engineering teams sprint to develop new features that they hope their users will love. As they begin each new project, they use the framework of the “Minimum Loveable Product,” shown above, to ensure they’re learning as quickly as they can without frustrating their users.

The analogy is simple: if you need a means of transportation, a car provides the greatest benefit, but it takes the longest to make—and worse, it’s hard to learn as you go. A skateboard, meanwhile, lacks a lot of features, but it’s better than nothing. More importantly, it’s quick to build and delivers immediate feedback.

Spotify Agile coach Henrik Kniberg explains further:

…The squad needs to figure out the smallest possible thing they can build to fulfill the basic narrative and delight the users. We need it to be narrative-complete, not feature-complete. Perhaps a better term is Minimum Loveable Product. A bicycle is a loveable and useful product for somebody with no better means of transport, but is still very far from the motorcycle that it will evolve into.

Source