Friday, November 10, 2006

Execution and Audibles

It's a little thing, in itself, but it speaks volumes.

Last Spring, in consultation with the faculty senate, we agreed to shift a semesterly deadline. The discussion started within the administration, but when we floated trial balloons, we found widespread support for it. We consulted with faculty, proposed it formally, and got it passed after healthy discussion. Hearty congratulations all around.

This Fall, I noticed that the deadline didn't change.

I checked the notes from the Spring meeting, to make sure my memory was correct. It was.

I started asking around. What happened? Why didn't it move?

After the initial “yeah, that sounds familiar..” and “did we really do that?” and “hmm,” it became clear that the problem was that it crossed jurisdictions, and no one person was in charge of getting it done. So nobody did, and the separate silos of the organization just kept doing what they've always done.

Now, of course, it's too late to get it moved for this semester. In fact, we plan so far ahead that by the time it gets included in the next plan, we probably will have lost a year.

Sometimes planning is actually the enemy of execution.

In football, sometimes the quarterback notices that the defense is lining up in a way that would likely thwart the play originally announced in the huddle. When that happens, he “calls an audible,” meaning he yells out a new play at the line of scrimmage to adjust to the defense. (Peyton Manning goes so far as to do a 'chicken dance,' adding to the entertainment value for the fans.) Audibles don't always work, of course, but the idea is that sometimes you run into an unanticipated hurdle and need to adjust on the fly.

When you plan everything a year in advance, and delegate execution to fifteen different silos, and insist on consulting anybody and everybody, it's impossible to call audibles. If the plan falls flat, so be it.

It's frustrating on a number of levels. Any call for empowering somebody to call audibles would immediately be taken as an assault on shared governance, as if there's no discretion in execution. And the ideals that stand in the way are, in themselves, good: wide involvement, respect for jurisdictional boundaries, transparency of procedure. In the abstract, they've hard to argue against. The counterargument to the 'audibles' model is easy to conjure: 'discretion' is a blank check, this is a naked power grab, you people should have done it right the first time, etc. There's enough truth in each of those, separately and together, that they can't just be dismissed.

But it's silly to assume that a plan either has to be perfect and complete at the outset, or killed. The kind of continuous improvement that serious organizations strive for is based on a recursive pattern, in which quick feedback brings adjustments, which brings new feedback, which brings new adjustments, and so on. Our procedures are designed almost perfectly to defeat that, even if for all the right reasons. If the initial plan turns out to be flawed, there's always next year – if we're quick – or the year after that. Assuming we remember.

There's a copious literature on making change in large organizations, but it's almost all based on the private sector, where criteria for success are clearer, a competitive market is a given, and it's at least permissible to tie pay to performance. Academia is a different ballgame. If you called a quarterback sneak and the defense is stacked at the line, well, tough. Plow forward anyway, and try not to drop the ball. It's your fault for not calling it right in the first place.