Haunted

In the late 1990s, Rosina Bierbaum of the U.S. Office of Science and Technology Policy (OSTP) asked me to lead a team to examine deficiencies in U.S. climate modeling and supercomputing efforts. The U.S. had launched several initiatives, spent a lot of money doing it, and was still unable to deliver products – computational simulations to provide climate projections – for the first National Climate Assessment.

A key finding of the report showed U.S. modeling activities were riddled with fragmentation, which was the self-evident reason I was asked to write the assessment. But it was more than just the existence of fragmentation the team identified. The report captured the pervasive nature of fragmentation and how it is sustained as part of our culture.

This experience would define (haunt?) a significant portion of my career.

The sum of its parts

I had been a manager at NASA and, more than most, was known as a student of organizations. I encountered exceptional individuals in every organization I’ve managed. But their collective output never seemed to rise to the level of what should have been their combined excellence in creating products, i.e., state-of-the art forecast models.

This reality stood in contrast to a widely held value that if we did “excellent science,” the products to address broader societal issues would occur organically. The pervasive culture dictated that directing scientific organizations toward a desired outcome ultimately killed innovation and resulted in deficient products.

It was true that our scientists ranked among the best in the world. But it was a naïve fantasy to believe one could synthesize the work of a fragmented collection of excellent scientists into equally excellent integrated outcomes. The products were not excellent. They were mediocre – fragmented mediocrity.

Focus, focus, focus

Creating a quality product requires focus on every detail from the beginning through the outcome. Someone has to manage to that outcome. There needs to be a balanced investment in every aspect that contributes to the product’s excellence.

Most of the products I managed required leading-edge computational resources. If adequate computing systems and software to use those systems effectively were not available, the promising research of these brilliant individuals could not be realized. Meanwhile, the complexity of climate modeling requires research produced by many diverse people and groups. A deficient component (or person or team) in one part of the system creates a ripple effect of deficiencies in other parts of the system and reduces the quality of the entire system. Fragmentation permeates the product.

Think of it this way: Putting a high-powered carburetor and racing tires on my 1980 Chevette would not make it a high-performance car.

The natural state

We have created a system of scientific organizations and research institutions in which fragmentation is their natural state. Individuals in such an organization often prefer fragmentation. They may lament the inability to achieve the collective goals that are institutional priorities, but they characterize the fragmented organization as the immovable nature of excellent research. They view organizational excellence as coming at the expense of individual accomplishments, for which they are rewarded in multiple ways.

Researchers are expected to generate funds from sponsors like the National Science Foundation or the National Institutes of Health. Take, as an example, a professor at U-M. The professor’s primary allegiance will likely lie with the research program managers at a federal science agency. A primary metric of the professor’s performance within their institution will be their success at getting funding.

A successful research proposal will be new, innovative, and exploratory. The grants process is not designed to build and sustain the excellent integrated science and infrastructure required for excellent use-driven products. Competitive acquisition of grants is yet another fragmenting process that leads to a treasure chest of glimmering research accomplishments.

Some of these research gems might get combined and turned into products. This is more likely to occur if the product has the potential to generate short-term revenue – be “commercialized.” But what of those essential products that society needs, such as environmental predictions, that are messy, uncertain, and not so easily commoditized?

Sabotage

Individuals learn to thrive in the fragmented system. Even in a simple case I managed, I caught my staff actively sabotaging efforts to improve the computing environment because they had so effectively figured out how to beat their colleagues to access the computational resources. In the fragmented organization, power structures develop. Managers figure out how to protect and grow their part of the organization. They, therefore, preserve the fragmentation.

The consequences of this fragmented culture are widespread, as people in these organizations behave to sustain that fragmentation. Some 20 years after that 2001 OSTP report, people still wring their hands over the state of U.S. weather and climate modeling. In a place like the University of Michigan, it slows efforts on such vital goals as how to teach ways to address climate change and carbon neutrality.

And so it goes

The structural elements of fragmented organizations described above are only the beginning. The distractions caused by excessive multitasking, increasingly fragmented funding sources, and the imperatives of multiple forms of instant communication anchor individuals in fragmented behavior that impairs the ability to achieve excellence. Indeed, poorly posed institutional efforts to coordinate and integrate activities only exacerbate fragmentation and usually validate the narrative that organizational interference leads to slow and deficient research outcomes.

Strategies and tactics exist for managing toward integrated institutional goals in fragmented organizations. They require the concrete definition of goals, a shared understanding of them, and the development of a managed project to achieve them. They require that a person or entity be responsible for those goals and that rewards and incentive structures are aligned with those goals. Finally, they require strategic alignment of resources. That is, one carves out a project to achieve a goal.

Even with such a level of definition, it isn’t easy to achieve the successful delivery of integrated outcomes. Such projects remain embedded in fragmented organizations, and organizational goals are constantly eroded. A rigor of management is required, which is alien to most researchers and research organizations. It is difficult to maintain continuity as leaders and managers come and go, each with an imperative to leave their personal, fragmented imprint on the organization.

I believe it is possible for an organization to produce products and outcomes, with intent, that are better than the serendipitous collective of their individually excellent contributions. Indeed, some organizations do this, and they provide use cases. This is not achieved by shifting the culture of entire research institutions, which are, in many ways, very successful. It requires thoughtful, adaptive leadership and management. It demands the persistent pursuit of values that we profess are important to the identity of the institution and an actual understanding of how complex problems are solved.