Friday, February 16, 2007

Accountability, systems, and loop gain

For some time, I've been thinking about this subject; it's time I mention it here.

For organizations that want results, I see a real problem with a focus on accountability.

I have no problem with taking responsibility for one's work. I have no problem with organizations wanting results.

I do see problems with management uping the ante and, under the flag of accountability, raising the rewards for good work and the penalties for bad.

That sounds vaguely problematic, as if I'm condoning poor work. I'm not; I'm actually encouraging better results. Let me see if I can explain.

Some years ago, I did a system dynamics exploration of a problem with a particular organization's ability to manage their expenses. A system dynamics exploration is a simulation-based investigation to find the reason for a problem and offers a platform on which potential solutions can be tested.

I described the project in an article in National Productivity Review ("Pipeline Inventory: The Missing Factor in Organizational Expense Management"), and I later published a bit more technical explanation of the computer model online. The project worked well; we reduced the organization's problem by about 95%.

There was one parameter in the model that represented how much managers would change their behaviors based on data they saw in real life. For a while, I couldn't figure out what to call it except perhaps a fudge factor. I set it to make the simulation results representative of real life.

Then it dawned on me: that variable represented management pressure, which I think might be another word for accountability as it's sometimes used today. If management pressure is low and you overspent your budget last month a bit, you might have a discussion with your boss, and you will try to adjust your spending next month to compensate. If management pressure is high, your boss may call you in and yell at you a bit, threatening your next raise, your current salary, or perhaps your job. Unless you are ready to be hunting for a job, you'll underspend by a lot next month to ensure you don't face that tirade or those threats again.

What's the matter with that, you say? Well, if you looked at the model output, you'd see that the manager who underspent more drastically because of management pressure accumulated an unsatisfied backlog of purchasing needs—all the computers, software, and other things needed to achieve the organization's goals. When the pressure was finally off because the budget was again balanced, that manager tended to buy all those needed things, putting the budget back in the red. You can see the results on page 6 of the article.

The model shows a most interesting lesson. In a poorly designed system, high management pressure (high externally-imposed accountability) made things worse, while low accountability actually made things better. In a well designed system, management pressure really didn't matter so much; one could view the level of management pressure to be applied as a matter of personal preference. As I suspect management won't always have a perfect system in place, lower management pressure may still be in order just for those cases when the system has unrecognized or unacknowledged problems.

Stated this way, it sounds positively Demingesque: the focus should be on the process (the system) and not on the individuals, if you want to achieve success. Management's primary job is to create systems that work well, not to push people to do well.

I still haven't mentioned loop gain, but this essay is getting long. I'll cover loop gain, how these ideas may generalize to other situations, and perhaps how one might go about creating such systems at a later date.

Thanks to Michael Patton for kicking off a conversation on evaltalk about David Weinberger's The Folly of Accountabalism that prompted me to write up these ideas.

Labels: ,


Post a Comment

<< Home