Be Embarrassed or Look Up Cerulean Blue

This isn’t a post about how it’s OK if you suck.

Not even how it’s OK if you suck just at first at whatever you’re doing, or how it’s to be expected because you’re a newbie, or how you’re encouraged to suck because that’s how you improve.


This is a post about how if you don’t share a creation that sucks, then you’re doing it wrong.

How ’bout them apples! I feel a sense of relief just thinking about it. Screw pride perfectionism! If it stinks to high heaven, then air it out!

This thought derives from a quote by Alberto Savoia from his GTAC 2011 Opening Keynote: Test is Dead.

If you are not embarrassed by your initial launch, you’ve not launched early enough.

The premise of his talk is on how there are two types of bugs: code bugs and idea bugs.

Bugs in code are what tests are supposed to find. If you press a button, the thing should turn cerulean blue. If you test the thing by pressing the button and the thing turns robin’s egg blue, then it was a good test ’cause it shed light on how the thing doesn’t meet the cerulean blue requirement. So then you Google ‘cerulean blue‘ ’cause, honestly, who could possibly know such mysterious corners of the rainbow except my wife, fix this bug, test it again, and the test passes.

Bugs in idea are what users end up finding. If you press the button, and the thing turns cerulean blue, and the intended user can no longer find the thing among the malachite green things because the target market is blue-green color blind, then it was a good exercise ’cause it shed light on how the cerulean blue requirement is a bad idea. So then you Google ‘tritanopia‘ ’cause, honestly, who could possibly know such blurred reductions of the rainbow except the internet (and optometrists, and opthamologists, and tritanopiacs…), fix this bug, give it to the user again, and this type of test passes with flying reds, cyans, blues, and byzantiums.

This colorful example highlights the Law of Failure:

Most product ideas fail in the market… even if they are very well implemented.

The first example highlights a bug in implementation, which, when fixed, represents progress, albeit in the wrong and ultimately embarrassing direction. The second example is preferred, enduring early embarrassment, ensuring the right and ultimately market-relevant direction. The idea bugs are more valuable than the code bugs and thus are the ones we want to catch early. Testing has historically focused on verification, “build the product right,” whereas focus is shifting towards validation, “build the right product.”

In the software world, this means distributing, early and often, a build to a small portion of the market for feedback.

In the personal world, this means sharing, early and often, a creation to a small portion of your fans for feedback.

Your first shared version should be embarrassing. Hmmm. OK.