Anti-Fragile Universities

I recently had a conversation with a retired IT administrator from Boston University. He commented that people in universities, just like in most other organizations, usually operate to preserve the existing order. This reminded me of how difficult it is to alter the curriculum, whether by introducing new content, modifying its architecture, or eliminating failed programs. People who work in universities don’t like to replace something that isn’t working with something else that might work a lot better.

Yet we tell students something else entirely: that they should have the courage to fail, and to learn from it. And any system — whether a business or a living organism — grows weak if its constituent parts are not allowed to fail. It’s much better to fail often and fail fast, because learning what doesn’t work early on and responding to this information will minimize harm and generate the largest benefits.

In other words, to paraphrase a regional director from Google whom I’ve heard speak on innovation:

  • Move fast
  • Experiment, but make sure to measure what you’re doing
  • Accept that you might be wrong
  • Focus on results and use a data-driven approach to make difficult decisions
  • Cut your losses without remorse

Evolution functions according to these principles. Universities generally do not. And this idea led me to revisit Nassim Taleb, whose work, at least on the non-technical side, I was already somewhat familiar with. In the changing landscape of higher education, many universities now fit his description of a fragile system, which is represented by the graph below.

The vertical axis is relative gain or loss in whatever the system is dependent on. For businesses, and that includes universities, this is revenue. The horizontal axis is delta, or change in some input. For a standard business, that means sales. For a university, it’s basically students.   The graph shows a system that is at high risk of catastrophic failure if there is a relatively small negative change in inputs.

Fragility Graph

Let’s say a university enrolls five percent more students — move a certain distance to the right on the horizontal axis — and there’s a corresponding increase in net revenue. The green line goes up. For the purposes of example, let’s say revenue increases by $20 million. If we make delta negative — and we decrease enrollment by five percent — the revenue change is not -$20 million, it’s much greater. The red line drops farther and faster than the green line goes up. There’s a second order effect in which the downward rate of change accelerates. So the university, or whatever system you might be looking at, is at extreme risk if there is a relatively small unfavorable perturbation in the environment.

So what’s the solution? Universities should do what we tell our students to do. They should be constantly experimenting to figure out how they can become less dependent on their traditional products and services — the markets in which competition is the most aggressive  — and enter market niches that lack dominant players. This might mean continuously and ruthlessly testing new courses, delivery formats, degree programs, or even educational services that aren’t oriented around completion of a degree.

5 thoughts on “Anti-Fragile Universities

  1. Chad,

    The list your Google exec gave you is most emphatically not how evolution works. Evolution is inherently conservative and builds slowly (very) around incremental changes that prove to enhance adaptation. Further, most of the innovations in genetic material evolve from a random process and are either recessive or prove to be part of the large segment of every genome that doesn’t appear to have any actual purpose at all.

    Or, to put it another way, evolution works in pretty much the same way that post-secondary institutions have done for most of their history. This is one reason why so many of them have lasted so very, very long. To extend the analogy, “moving fast” and cutting losses can deprive the schools in question of the variation in courses and faculty necessary to meet new changes in the environment. An example = data science. This is a hybrid discipline of the social sciences, computer science, and statistics. It didn’t exist 7 years ago. What if a school cut out its faculty resources in such a way as to make it impossible for them to exploit such an opportunity?

    I’m not saying that schools should ignore the bottom line; they can’t. What I am saying is that any changes need to be both incremental and aimed at preserving institutional variation that can support new trends in the environment.

  2. Law of large numbers: given enough time and a fixed set of possible permutations, or even fairly well predictable variation, outcomes coalesce around the probabilistic mean. Feet and legs are fairly inefficient at propelling an animal through water, so fish have fins instead of legs. The climate was quite comfortable for reptiles for over a hundred million years, and dinosaurs did very well. Ignoring all the mutations that appeared along the way, didn’t contribute enough of a reproductive value to a species, and disappeared relatively quickly is to engage in survivorship bias. It’s quite possible for an organism, or an organization, to become highly adapted to an environment that suddenly ceases to exist. Giant meteor hits the earth, temperature drops, and the small mammals that did several things only passably well survived; the dinosaurs that were so specialized went extinct. So you are correct, it is not good for an organism to overspecialize if the environment in which it exists is subject to unpredictable variations. Yet I don’t see the vast majority of colleges and universities eagerly experimenting with small incremental changes to find out what will be of benefit and what won’t so that they can better withstand change. And the meteor is about to hit.

Leave a Reply