This is the first of a probably series of a generally unedited “stream of consciousness” posts I plan on writing while reading Donald Reinertsen’s The Principles of Product Development Flow: Second Generation Lean Product Development. I haven’t been this affected by a book in a while, and I wanted to share my excitement with the world and test some of the ideas it is provoking. As such, this is intended to be an invitation to discussion, and I hope you will add your comments and disagreements below. If you wish to engage me in a detailed conversation, please feel free to contact me.
The Asymmetric Payoff Amount
In the section on economics there is something that struck me as so profoundly true I had to stop for a while. Okay, it isn’t the first thing that has hit me like that in this book, but it was the first that triggered an extended series of thoughts. He is discussing the asymmetry between the investment cost and the payoff amount for development projects. He talks about a conference that he was at recently where a speaker pointed out the 96% failure rate in new product development and that poor management is to blame. We tend to spout the same sort of statistics and give the same reasons. Here’s what Reinertsen has to say:
“In fact, economics would suggest that the 96 percent failure rate tells us more about the asymmetry of payoffs from product development than it does about our management competence.”
This struck me as profoundly and unequivocally true. If our failure rate is so abysmal, we would have ceased investment in software projects ages and ages ago. Why would we have continued to create IT departments at a prodigious rate?
What happens in 10 years? In 20 years? I immediately began to think about early manufacturing and such. In the early times, inefficiencies were ok. It took 50 years before it was really necessary to be efficient. There is a parallel there. The returns on software projects in the past could support incredibly high failure rates because the return on the few successes was so amazingly high. The economics could be ignored because it was all going to come out in the wash. Companies that are successful right now with crap processes are okay with high failure rates because of the asymmetric returns.
This can’t continue. Just like the western model of car manufacturing was crushed by Toyota after the market became saturated with cars, our market is becoming saturated with software. No longer are new software projects replacing some archaic pen and paper version of functionality. As that happens, the asymmetric returns are going to rapidly diminish. No longer will your 1 success return 1000x on your investment, masking your 4 previous failures.
I think this means a few things become super important in the new software development landscape. First, extraneous features need to die a quick death. Modern video games have realized this, releasing a core product then only adding new functionality when enough people scream for it. Case and point, Diablo 3 was released without any player versus player capabilities (a core component of the genre) and still sold record copies. If the game had tanked (which was a distinct possibility with expectations so high) Blizzard could have walked away and saved the cash.
Second, decisions can no longer be made in an economic sink-hole. Software development got away with this for a long time, but margins will(are? No empirical data at hand but I think I can safely assume this to be the case) shrink, and shrink rapidly. Think convincing people to buy Microsoft Office 2013 when they currently use Microsoft Office 2003. Now compare that with the jump from DOS to Windows 3.5.