When Success is Bad – The Math Behind Why Failure Is Essential
Posted by Plish on April 30, 2013
We all know Murphy’s Law: “Whatever can go wrong, will.”
But, what happens when the right things happen for what we think are the right reasons?
Or, restating in a slightly different way:
In any system there are ways of achieving the correct result through a combination of known and unknown means.
As any product developer will tell you, there are times when you test a product (or code) and it works. You get excited. You decide to show your results to others. You do everything the exact same way you did last time, only this time, it doesn’t work.
Not. At. All.
How does this happen?
Let’s imagine a product as a mathematical expression: A+B+C=X (eq.1). A, B, and C are things that we know, things that we do to bring about X which is the result we want to have happen. Let’s call this the “Success Equation”.
The equation for some undesired outcome could be depicted as: A+B+C+W=Z (eq.2). W is some known wrong step or condition that causes Z, which is an undesired result. We can call this the “Devil We Know” equation.
Now, when working with a product prototypes we actually only know what we know. Sounds obvious right? Another way to say this is: We don’t know what we don’t know.
What this means is that the REAL equation for our product is very often: A+B+C+D+E+F=X. (eq. 3) This is one of many versions of what I’ll call the ”Devil We Don’t Know” equations.
A, B and C are known and are BOLD in the equation. D, E and F, are grey because we don’t even know these variables exist. Nevertheless, they are a part of the equation and if they all come into play, X occurs, so we’re happy.
And that’s a problem.
What happens if D, or E, or F, or some combination of these disappear? We could get the formula A+B+C+E+F=Z, (eq.4) . What gives? We’re doing everything right, just like we were before, and getting the wrong result!!!
And it gets worse…
If n=”The number of variables we don’t know, but when all are present result in success”, then there are 2(n)-2 possible permutations of possible failure modes. In other words, if there are two unknown variables, then there are two possible failure mode combinations; 3 variables translates to 6; 4 to 14; and 5 unknown variables could lead to 30 possible failure modes!
Which brings us to the main point of this post.
Failure in the product development process is a necessity!
We ultimately want to get to the Success Equation (eq.1). We want to be able to know that every time we do A and B and C, we get X.
The best way to get there is to convert every “Devil We Don’t Know” equation (eq.3) to a ”Devil We Know” equation (eq. 2). And that only happens through testing, experimenting, failure, and learning from those failures.
So the lesson from this Law is this: Next time you’re testing a product and it works as expected, don’t fall into the trap of thinking that your product is working perfectly. Test. Fail. Test again! Avail yourself of digital tools and novel testing techniques (and people that like to break things) to create failures and learn from them. Find out what you don’t know.
Fail early, fail often, fail to learn, fail to succeed.
Oh, and this Law needs a name. Any suggestions?