Using decision analysis to optimise R&D investment: the case of SmithKline Beecham

I want to talk about decision analysis in organizations, and I want to walk you through this problem that SmithKline Beecham had, and this was in the late '80s. It was one of many mergers in the pharmaceutical business - SmithKline on one hand and Beecham on the other. They came into a single company, and their problem, really, as in most pharmaceutical companies, is how do you select the R&D projects, the research and development projects to fund? You want to have new drugs in the pipeline all along, but how do you know where to funnel the money? Well, typically, what was done was the so-called 'project champion' adversarial process is how I think about it. The team leader on any given R&D project would develop his or her pitch, try to sell it to the manager. They gradually work their way up the hierarchy, and have to go back every time and fix something, change something to satisfy that next level up. This decision-making process was taking an enormous amount of time, so they were actually looking for an organizational solution, an organizational decision process.

They came up with an organizational approach which is relatively novel. Now, it had been used in some other organizations, but this was the first time it had been applied in an R&D organization, and pharmaceuticals in particular. So I've got this snaking graph here. It represents time moving from left to right. At the bottom left, we've got project teams. So the first thing that happens is project teams get together, and they actually create alternatives. In this approach, they were given some templates to work with. What if we fund you at current level? What if we cut you back by 50 percent? What if we double your funding? They had to look at a number of these standard scenarios and think about what would happen in those cases. Well, on creating alternatives, and this means thinking through how to modify their various research projects to fit the different funding levels, they would generate specific alternatives like, "Okay, we'll abandon this particular aspect of the study, we'll go this way, we'll think about only an injection form, not an oral form," and so on. Well, after alternatives are created, then they go to the decision-makers, and they have - think of it as - a decision board meeting. At that first meeting, they get guidance, and ultimately they'll get approval from the decision-makers. They get buy-in at that point from the decision-makers on the alternatives that they're going to analyze. Then let's evaluate the alternatives. Now rather than a project team evaluating, they develop an independent analysis team that would evaluate those alternatives, and they had specific kinds of analyses that they would do for all projects. So it was such a problem previously when different analyses were done, and you know that the project leader would choose an analysis that showed his project in the best light. Well, with standard analyses, that made the comparison of the projects much easier. So the second meeting took the evaluated alternatives to the decision-makers, and at the second decision board meeting, we would have critical review and approval. Again, buy-in. Now, what's nice is, at this stage, the decision-makers don't get to say, "Oh, you should have evaluated such and such an alternative, " because they already got to say that before. The third step, and again this is the analysis team, they gather all the different evaluations of all the different projects because now the question is, "Okay, how do they compare and how should we allocate resources?" And this is an optimization step, where they're essentially building an optimal R&D portfolio. They make their recommendations, and they take that to this third meeting where decisions are made and implementation plans are drawn up for exactly how to move forward now that we know how to allocate funds. Now, I like this because you can imagine a project leader whose funds were cut coming to the managers, the decision-makers, and saying, "By cutting my project, you will be losing out on, say, $10 million in value," and the decision-makers can look at him and say, "Well, yeah, we lose that $10 million but, in place of it, we get this $15 million over here." Now, it's hard for the project manager to argue with that.