The 50th birthday of Decision Analysis led to an interview (podcast and transcript here) of Ron Howard in the HBR IdeaCast series. Judging from social media it’s been attracting some attention in the DA community, and one hopes more broadly. This is the sort of thing I’d like to see more of, DA mavens interacting with members of the media who are business oriented but not part of the DA or even OR/MS choir.
Justin Fox, the interviewer, focuses on the meat-and-potatoes issues of a management methodology: what is it and where is it used? Howard gives answers that sound familiar, probably because I’ve heard him and his many former students give similar answers in other fora.
People with limited exposure to DA often have the impression that it lives in some sort of specialized niche. They have good reasons to think that: DA is not very common or well known, but its practitioners are enthusiastic, so one naturally assumes it’s extremely useful for some narrow class of problems. You can see this impression in Fox’s line of questioning:
JUSTIN FOX: What kinds of decisions does decision analysis best lend itself to? What are the best ones to use it on?
RONALD HOWARD: Any decisions involving your money, your health. I once had a student say, well, I can see using this for your financial decisions. But what about the health of your family? And I said, well, if I had to choose one area in which I could use it, that’s the only one. I’d say the health of my family, that’s most important thing for me. And I’d get some financial advisor to manage my money. But in fact, it applies to any decision. So any decision where you’re allocating a resource, your body being a resource. It only doesn’t apply to where there is no resource. So deciding who to love somebody …
[… some banter about love and money omitted …]
JUSTIN FOX: What are the areas in– I mean, you started doing this project with a nuclear power group.
RONALD HOWARD: Yes.
JUSTIN FOX: And what are the other areas in which this approach has had the most success, had the most uptake?
RONALD HOWARD: Well, the two biggest areas, I think, of use are, first of all, oil and gas companies. And the second biggest, I think, would be pharmaceutical companies. And if you think about why those areas, they both have this characteristic of, first of all, billions of dollars often to get a successful drug or for you oil field. Secondly, a long time between making the decision and seeing how it worked out. And thirdly, a lot of uncertainty. So if you put those three together, that’s prime hunting ground for decision analysis.
Decision Analysis is about decisions, and there are so many decisions out there that when people see DA their first response is often “it can’t be that general, can it?”. Let’s reflect on Howard’s three characteristics of the places DA has had the most uptake, both descriptively and proscriptively.
1. High stakes
DA done well (or even badly) takes some time, often many person-months, and it’s self evident that you should spend more time thinking about big decisions than small ones. At the same time, small to medium decisions made frequently can add up.
2. Long time horizons
In the realm of common sense, it’s hard to separate time and uncertainty. Forecasting the prospects of a business 5, 10 or 20 years from now is hard, because so many things can change in that length of time. Just the same, time is an imperfect proxy for risk. There’s more uncertainty about Uber’s 2016 revenue than about Duke Energy’s 2026 revenue.
3. A lot of uncertainty
Most DA people say this, but is it true? Oil & gas and pharma are certainly big DA industries, and they have plenty of uncertainty, but there are many sectors with ample uncertainty and very little DA. The entertainment industry, for example, involves spending large sums (sometimes nine figures) on films and games with very uncertain prospects. I’m not aware of anyone taking a decision analytic approach to managing an entertainment pipeline (by all means tell me if you are).
I would characterize the industries with the most DA uptake as being those with officially acknowledged uncertainty. Nobody knows whether a drug in phase I clinical trials will be successful, and everyone knows that nobody knows. In many industries, uncertainty is still seen as a sign of poor management (“A good A&R man can tell if a song will be a hit”). It’s tough to develop a culture that embraces uncertainty when you’re still looking for white knights who can banish it. Fox alludes to the question of why there isn’t widespread use of DA in startup Silicon Valley (as distinguished from capital intensive tech like semiconductors), and white knight culture may be part of the reason.
Press is good. Read the interview if you haven’t already.