Thursday, June 10, 2010

Planning is Guessing?

So says 37Signals.

37Signals has a lot of good ideas about software design that I like. In fact, I absolutely love Writeboard and use it often. Unfortunately, their generic statement about planning really only works when the cost of development and other opportunity costs are near zero. Otherwise, when one faces committing expensive capital to the task of generating a return, especially with other people's money, good planning is not only prudent but a moral obligation.

When most people plan, what they are generally doing is laying down a series of tasks with desired durations and costs, essentially a deterministic script, leading to a poorly defined goal. Then they express surprise and frustration when life doesn't follow their script. Then, possibly after several failures, people replace one type of pathology (over analyzing with deterministic assumptions) for another (shooting from the hip).

The big problem is not with planning but with a failure to plan properly. When people come up with one idea about "how to get there", they fail to consider that there is often more than one pathway to their goals (if they even have a clear understanding of what their goals should be), maybe a more valuable one. Furthermore, they don't assess the impact of uncertainties on reaching their goals, both within and between alternate pathways to the same goal. As a result of both failures, they don't develop a proper understanding of comparative capital efficiency of choosing one pathway over another, nor do they develop proper contingency plans when things don't go as anticipated. Good planning is mostly about clearly framing an opportunity, comparing alternative approaches, developing a reasonable plan forward, and developing contingency plans for when deviations occur. But I cannot overemphasize proper framing and comparing alternatives.

I understand 37signals point. Analysis paralysis leads to slow death by value attrition. Unfortunately, when people tire of analysis paralysis, they frequently resort to a "just get it done, already" approach that can actually lead to bigger problems, namely, costly rework or disaster control. I've seen that outcome in about half of the decision failures I've encountered over the last 15 years of decision support and risk consulting. Right now, we're all too familiar with an ongoing disaster in the Gulf of Mexico. A broader body of published evidence shows that "just getting it done" is probably more destructive to shareholder value than analysis paralysis. The solution, though, is not to replace one pathology with another.

Planning is guessing. Yes, but the real question is: what kind of guessing is best? The right approach is one that avoids both pitfalls of analysis paralysis or flying by the seat of one's pants.