When I first went to work for Applied Strategies, I didn’t really understand what they did. Applied Strategies (at that time) specialized in doing demand forecasting using decision analysis, which meant that we constructed mathematical models to estimate the size of a market for a drug or a vaccine. Our analysts used complicated decision trees, taking into account the probability of getting to clinical trials, succeeding in each phase of clinical trial, getting FDA approval, what other competitors were entering the same space, how much it would cost to build a factory and scale up production, etc.
At the end of all of these calculations, a number was delivered as to the size of the market. And I was horrified to think that this number was a deliverable. There were an absurd number of assumptions made in the calculation of that number, and changing any of those assumptions would alter the number significantly. So I couldn’t figure out what the value of that number was.
I eventually figured out (or had it explained to me) that the real deliverable wasn’t the number; it was the model. By constructing this model that had all of these assumptions and possible inputs, our customers could then see for themselves how altering certain assumptions would affect the result. The model also helped our customers do a sensitivity analysis and figure out which inputs had the greatest effect on the final result, which determined the inputs for which they needed better data. In other words, the deliverable wasn’t even the model – it was using the model to help the customer determine the relevant factors in their analysis, changing how they viewed the world.
In class this evening, we were discussing various frameworks for strategy analysis, from Michael Porter’s Five Forces Analysis, to a standard SWOT analysis (Strengths, Weaknesses, Opportunities, Threats), to a PEST framework (Political, Economic, Sociological, Technological). One of my classmates at the break was commenting that he wasn’t sure that he saw the value in such analyses, because things change so fast these days that the result of the analysis is stale as soon as you’re done.
His comment reminded me of my Applied Strategies experience, and I observed that the point of the analysis was not in the specific results, but in the change in view that results from doing the analysis. For instance, us technologists find it very easy to get sucked into the everyday grind of specific deliverables. But if we do a five forces analysis, even if the results of that analysis are immediately obsolete, it starts to get us thinking about the big picture of the overall industry and where it’s heading. Doing such an analysis changes how we see the world, and that is the important result.
I am also reminded of scenario planning, as described by Peter Schwartz in The Art of the Long View and at a Long Now talk. The point of doing a scenario analysis isn’t to predict the future – it is to prepare oneself for the future. By examining the larger trends that can shape the future and how one’s organization would react to those trends, scenario planning allows the organization to identify key pointers that indicate which way trends are heading and therefore be able to take proactive rather than reactive action.
This idea that the results of a process are a change in view rather than a tangible result resonates with me. Many organizations claim that people are their most important asset, but few organizations follow through in realizing that changing the viewpoints of their people is among the most effective investments they can make. If everybody in the organization is thinking strategically (shades of Semler), then the organization will have achieved greater results than they ever could have by investing in technology or tangible assets.
Several people have asked me why I’m pursuing this degree in technology management, when it appears to them that I have all the skills for management already. This viewpoint change is a significant component. I know how to think like a scientist, an engineer, a technologist, and an end-user. I even have some inkling of how to think like a manager. But this program is giving me the tools to think like an executive, and to put issues into the terms they will understand.
As an aspiring generalist, I believe in the power of bringing multiple perspectives to an issue, of having many mental models in my toolkit. I want to be able to communicate effectively with every level of the organization, from the peons doing the front line work, all the way up to the executive board who are concerned with ROI and shareholder value. I want to be able to re-frame issues and ideas in such a way that they make sense to anybody I talk to (part of the point of this blog is for me to practice explaining ideas within the constraints of a few paragraphs). I’m still not quite sure what precise career path this is preparing me for, but I’m investing in the most valuable resource I have, my mind, and I look forward to the eventual returns on that investment.
“The value in forming a plan is not the plan, but in having done the planning,” or some such. I’m sure that’s isomorphic to some famous quote, but I can’t find it.
I took a seminar on decision analysis, and I thought one of the more interesting things is that it can show that certain things that seem important may not matter at all, since they won’t change your decision in the end.
Even when you’re a technologist down in the mucky details of writing code, it’s valuable to occasionally stop, step back, and ask “wait, what am I ultimately trying to accomplish here, and how does this fit in?”
As you probably know all too well, I am an unrepentant builder of these complicated, mucky models, and yet I fully agree with your point that the value is not the output, but the model itself.
I would actually go as far as saying that the model itself, as a finite product, is of little value; what really matters is its inception. The iterative discussions in building the model itself is what matters, in that it progressively elicits what question is really being asked, and how the user represents the world.
In that sense, decision analysis is very analogous to the socratic approach – discuss and criticize, and help your partner in discussion until his/her ideas are well-formed into a coherent entity. The rest – running the model – is essentially a tautology.
As an aside, it’s clear why technology-minded humans would lose sight of the question and get excited with nitty-gritty technical aspects; but too often, users themselves do forget that the model is nothing more than what they said it was – and yet, they invest it with quasi-mythical powers of divination, when really, they have done all the work.