#TBT: Designed to Fail

Happy #TBT. To celebrate, we’re sharing an oldie-but-goodie post from Jim: Designed to Fail.

I was talking to a new employee about failure and risk taking, and the point I was making was that while it’s true that failure proceeds success, there’s a certain kind of failure that I can live with and a certain kind I cannot.

And I used a crazy metaphor to explain it. So crazy, I thought it just might be worth writing about …

"Short Ladder - Big Wall," © 2012 oatsy40, used under a Creative Commons Attribution license.

“Short Ladder – Big Wall,” © 2012 oatsy40, used under a Creative Commons Attribution license.

Suppose your goal were to climb onto the roof of the two-story apartment building next to the Goldstar office in stately Pasadena, Calif., and a convenient, obvious solution (like a staircase) didn’t already exist. Suppose further that you came up with these two solutions:

1. Stale bread sticks and shellac. The bread sticks are pretty thick, and if we put them together with shellac, we can form a ladder that could hold one of our lighter team members, and up we go. It sounds crazy, and we don’t know enough to say whether or not it WILL work, but we know that it COULD work.

2. A six-foot ladder. Well, since the roof is two stories, that’s about 20 feet, so the six-foot ladder simply CANNOT work. It’s a solution that is designed to fail.

In other words, I can live with failures where what we try COULD HAVE WORKED, but ultimately didn’t because of something we didn’t understand fully. I can’t really abide by failures where the solution never stood a chance because it was designed to fail, like our six-foot ladder.

A good test is one where we don’t know if it WILL work, but we know that it could, if some of our guesses hold up. A bad test is one that is doomed from the start. If your online banner ad campaign needs a 2% click-through rate to get the ROI you want, it CANNOT work. If, like the concert industry, your business model starts with a fixed expense (in their case, legacy artist guarantees) that make the business fundamentally unprofitable, in the long run, it CANNOT work. If your server has a maximum capacity of 50 concurrent users and you’re putting it in front of a demo that may have 100 people using it at once, it CANNOT work. You could have done your ‘test’ on these things on your whiteboard, for free, in five minutes to see that they can’t get where they need to get.

So, yes, fail, but do so in service of learning if something that COULD work actually DOES work. You don’t need to climb it to know that a six-foot ladder will never get you 20 feet up.

Got a comment or question? Join the conversation on Twitter or Facebook.

Sign Up for Emails