How to Plan an MVP

Seibel defines an MVP as the simplest thing that can deliver value to the first users you want to reach. Talk to potential users before building, and know who has the problem. If you cannot identify those people, question your assumptions about the customer.

Seibel says a pre-launch startup should launch quickly, get initial customers, and ask them for feedback. Keep the problem and customer clear while staying flexible about the solution. If the product does not help the intended user, improve the product instead of searching for a different problem it might solve.

A lean MVP should take weeks rather than months and offer only a limited set of features for a small group of initial users. Seibel points to early Airbnb, which had no payments or map view, Justin.tv, which started with one channel, and Stripe, whose founders helped customers integrate it. He says regulated or technically ambitious projects may need a simpler starting point, such as a website explaining the product.

Seibel distinguishes getting any customers from a press launch that brings attention and buzz. He argues that customers learn more by trying a product than by discussing an idea without one. He recommends spending time building something customers can use instead of focusing on a pitch deck.

Seibel recommends time-boxing and writing down the spec, such as limiting a first release to what can be built in three weeks. A written plan makes scope changes visible; if the deadline is at risk, cut features, including features that seemed important. Treat the MVP as an initial step that may change, rather than something to become attached to.