TITLE: Fail Early, Fail Often... But Win in the End AUTHOR: Eugene Wallingford DATE: August 09, 2010 3:43 PM DESC: ----- BODY: I seem to be running across the fail early, fail often meme a lot lately. First, in an interview on being wrong, Peter Norvig was asked how Google builds tolerance for the inevitable occasional public failures of its innovations "into a public corporation that's accountable to its bottom line". He responded:
We do it by trying to fail faster and smaller.
One of the ways they do this is by keeping iterations short and teams small. Then this passage from Seth Godin's recent e-book, Insubordinate, jumped out as another great example:
As a result of David [Seuss]'s bias for shipping, we failed a lot. Products got returned. Commodore 64 computers would groan in pain as they tried to run software that was a little too advanced for their puny brains. It didn't matter, because we were running so fast that the successes supported us far more than the failures slowed us down.
In a rapidly changing environment, not to change is often a bigger risk than to change. In an environment most people don't understand well, in which information is unavailable and unevenly distributed, not to change is often a bigger risk than to change. However, it's important not to fetish-ize failure, as some people seem to do. Dave Winer reminds us, embracing failure is a good way to fail. Sometimes, you have to look at what failure will mean and muster a level of determination that denies failure in order to succeed. This all seems so contradictory... but it's not. As we humans often do, we create rules for behavior that are underspecified in terms of context and the problem being solved. There are a lot of trade-offs in the mix when we talk about success and failure. For example, we need to distinguish between failure in the large and failure in small. When an agile developer is taking small steps, she can afford to fail on a few -- especially if the failure teaches her something about how to succeed more reliably in the future. The new information gained is worth the cost of the loss. In the passage from Godin, successes happened, too, not only losses, and the wins more than offset the losses. In that context, it seems that the advice is not about failure so much as getting over fear of failure. When we fear failure so much that we do not act, we deprive ourselves not only of losses but also of wins. Not failing gets in way of succeeding, of not learning and growing. Winer was talking about something different, what I'm calling in my mind "ultimate failure": sending the employees home, shutting the doors, and turning the lights off for good. That is different than the "transitory failures" Godin was talking about, the sort of failures we experience when we learn in a dynamic, poorly understood environment. Still, Winer might not take any comfort in that idea. His company was at the brink, and only making the product work and sell was good enough to pull it back. At that moment, he probably wasn't interested in thinking about what he could learn from his next failure. Sometimes, even the small failures can close doors, at least for a while. That's why so many entrepreneurs and commentators on start-up companies encourage people to fail early, before too many resources have been sunk into the venture, before too many people have been drawn into the realm affected by success or failure -- when a failure means that the entrepreneur simply must start over with her primary assets: her energy, determination, and effort. When I was decorating my first college dorm room, I hung three small quotes on the wall over the desk. One of them comes to mind now. It was from Don Shula, the head coach of my favorite pro football team, the Miami Dolphins:
Failure isn't fatal, and success isn't final.
This seemed like a good mantra to keep in mind as I embarked on a journey into the unknown. It has served me well for many years now, including my time as a programmer and a teacher. -----