When people say it won't work, build enough to find out

The short version of this line travels well on posters. In a set of advice Crowley gave in late 2010, it continued: if an idea is stuck in your head, find a way to build it, so that you can prove to yourself whether it works. That ending changes the advice from a call to ignore critics into a call to run a test.

Crowley had reason to think that way. As a graduate student at New York University around 2000 he built Dodgeball, a service that let friends share their location by text message. Many people thought the idea was pointless. Google bought it in 2005, gave it little attention and shut it down in 2009. Crowley later said the team inside Google could not get engineers or attention for it, which is a quieter way an idea gets told it will not work: nobody argues, and nobody helps. By then Crowley had left and, with Naveen Selvadurai, built Foursquare, which launched at South by Southwest in March 2009 and had millions of users within two years.

The advice holds best for new behavior. People are poor at predicting whether they will do something they have never done, such as tell friends where they are eating. Critics answer from their current habits, and the only way to learn what people will do is to put something in front of them.

It holds worst when the critics are right, which they often are. Most new ideas do fail. The useful move is to separate who is saying no: friends guessing, experts with real knowledge of the market, or potential users who have tried an early version. Only the last group’s opinion counts for much, and building something is the way to reach them.

Building to find out does not have to be expensive. Dodgeball began as a student project running on text messages, with no app and no funding. Foursquare’s first version was built in a few weeks for one conference. A test sized to the question, rather than to the ambition, lets a founder learn whether the critics were right before spending years proving them wrong.