Condemned to repeat it
On Honda, the Olympic Games, and the difficulty in making use of the three branches of institutional memory
Variations on a well-worn saying go something like, “Those who fail to learn their history are condemned to repeat it.” Now, thanks to a paper by three scholars at Oxford, we have empirical evidence: They looked at the history of cost overruns at the Olympic Games between 1960 and 2024 and found that the same old problems happen over and over again -- even though the Olympics happen on a regular schedule, are at the center of enormous attention, and have extremely well-documented track records.
■ Nobody seems to learn, the scholars conclude, because moving the Games from site to site keeps anyone from really accumulating useful institutional knowledge. This is no particular shock to anyone who knows about the problems of institutional memory, but it nevertheless seems odd that despite an obvious (and costly) problem and an abundance of opportunity to improve, nobody appears to know quite how to do it. The Games go on, they end up blasting through budgets and cobbling things together in the 11th hour to pull off the event.
■ Institutional memory comes in three forms: Event memories, decision memories, and process memories. Event memories get documented fairly well -- it’s not hard to find photos, videos, and news stories about the last several Olympics. It’s the other two that are much harder. Decision memories explain why one path was chosen over another. Unfortunately, though, even though decision trees aren’t hard to build, the motivation to record them is hard to find (especially if the people making today’s decision don’t expect to have any part in the event two or four years from now).
■ And process memories are similarly hard to get down on paper. A recipe book is really just a collection of many different process memories: How to start with nothing and end up with the creation you wanted. The case may be especially hard with a sui generis kind of institution like the Olympics, but it’s rare even among successful organizations to find well-documented process memories. A few firms do it exceptionally well (Honda, for instance, reputedly sets the gold standard), but in a huge share of cases, people either actively or passively decline to record their process memories because keeping the knowledge locked up between their ears is a form of job security. Fire me, lose my insider knowledge.
■ The authors think the “myopia of learning” that plagues the Olympics is only rectifiable through sweeping reforms (which are unlikely, given the structural incentives involved). But the problem is far more widespread than just with the IOC and host city committees, and it’s fascinating to consider just how obvious the answers are and yet just how few institutions seem equipped to implement them.