The Power of Defining The Problem (Article Analysis)

My latest article analysis comes today from the Harvard Business Review. The article is called "The Power of Defining the Problem" and it was written by Dwayne Spradlin, here's the link (http://blogs.hbr.org/cs/2012/09/the_power_of_defining_the_prob.html).

So What's the Problem Man?
In a previous post I defined entrepreneurship as problem solving at its most fundamental level. With this as my thesis it seems obvious why today's article was relevant. As a clinical researcher and statistician this topic also hits very close to home for me outside of the classroom. Defining the problem before you start to work at a solution is so straight forward it is almost taken for granted, and yet I see poorly defined problems being tackled by poorly designed methodologies on a somewhat regular basis. This could be a major reason why start ups falter, to borrow from fellow ETR500 blogger John Levin, you can not start with a solution and back into a problem. There are plenty of examples of this behavior; developing a product then trying to "create" its market, presupposing that a clinical intervention works before you start doing an actual ROI analysis, or inventing an entire new footwear paradigm that nobody was looking for (Timberland case). These are examples of solutions that either failed to address a real problem, or facilitated poor problem solving with questionable methodologies. How can someone arrive at a destination without first ensuring that the destination is a real place and then determining the logistics for getting there?

Your Problem is Finding Your Problem!
Now that we appreciate the value of a well-defined problem, how do you define it. The article highlights 3 examples of simply asking the right question. Rather than say "we need a way to handle this frozen oil" Exxon said "we need a way to move work with extremely viscous liquids". I do this at work quite a bit, whether it is converting a business question into a data question, or finding additional insights within the data for management (the mythical unasked question). For example, I recently conducted research on racial disparity in healthcare; the question from management was a very open, very vague "do we have any disparities" however, that is a poorly defined problem. To make the problem more tangible I isolated several key metrics that would be reflective of racial disparities and then investigated those and reported results. So not only does management get a simple "yes, there are disparities", which should have been expected given the research by Johns Hopkins University, but now they know what specific disparities exist, how to mitigate them, and where to focus their efforts. That is like the Exxon problem; we knew something needed to change, however, through refinement we were able to pinpoint not only what needs to change but also how to change it.

So Do I Still Have a Problem After I Define My Problem?
One of the biggest pitfalls in research is bad experiment design. While this topic is probably worthy of its own blog post, I will touch on it briefly. Experiment design is so important my master's program at Northeastern dedicated an entire course to it, to put it in perspective. A good experiment, a good investigation, has to be freed from bias. For example, when you conduct an ROI analysis motivated by the question "so how big was our ROI" instead of "did we produce an ROI at all" the mentality of the analyst is going to be different. This is a very easy trap to fall into and there are several logical fallacies that address and attempt to combat this. For brevity I will distill experiment design to a single nugget; a good control group is the absolute most important thing for a good experiment. It does not matter how sophisticated your statistical tools, or how many best practices you want to leverage, if the control group is bad or even worse, non-existent, then your problem-solving methodology has a very serious problem.

Bottom Line For Entrepreneurs:
Without a clearly defined problem, you can not solve it. If you are not solving problems as an entrepreneur then what are you doing? The nature of entrepreneurship is to solve problems, sloppily defined problems get sloppily designed solutions, but how many people are in the market for sloppy? If you want to succeed your mission, your motivation, your raison d'etre has to be clearly defined. You can not win at step 10 if you botched step 0, step 0 is clearly defining your problem. Once the problem is defined, the methodology can emerge, and the solution can reach fruition. Problem-solving does not have to be problematic.