« The CMDB as a Knowledge Base | Main | Who's Your Daddy Now? »

December 6, 2006

Performance Recap

Recently I ran across the phrase "You can't manage what you can't specify" - which is a very nice iteration of the central role of definitions in the sphere of assessing performance.

If you'd like to run across it too, see Paul Allen's book, "Service Orientation: Winning Strategies and Best Practices" from Cambridge University Press (www.cambridge.org).

To celebrate the phrase further, here is a skeletal recap of ideas posted throughout Archestra on the subject of performance management:

1- Performance management presumes performance measurement, but measurement is worthless unless it measures the right things.

2- Definitions of performance are conventionally all about identifying how well execution towards a target went. Picking good targets in the first place is of course more important than any followup form of execution.

3- Targets are meaningful when they mark the demonstrable generation of significant differences (or "values" in Archestra parlance); the significance must have an even better definition than does the targets.

4- For a difference to be significant, there must be some context, usually a model, that invests particular observed conditions (and changes of them) with meaning. Therefore, the model is the most fundamental thing in the assessment of "performance".

Example:

In a "re-building" year, a team may lose many more games than it wins -- but the pertinent model is not wins-&-losses; instead it is "increases in competency".

But in a year expected to be a "contending" year, the pertinent model is "wins and losses attributable to the game plan" -- an entirely different model.

So a rebuilding team may show excellent performance against its targets; but it might be nowhere near performing like a contender.

Finally (or in fact, right from the beginning), trying to practice "performance management" without practicing "change management" is a joke.

Posted by Malcolm Ryder at December 6, 2006 4:47 PM

Trackback Pings

TrackBack URL for this entry:
http://www.malcolmryder.com/cgi-bin/mt-tb.cgi/297

Comments

I'd like to take a bit of issue with what I think is the conventional wisdom here--the idea that management requires measurement.

It ain't necessarily true. It presumes, I think, that all successful management techniques must have a component of measurement. It also presumes, I think, that you can't move forward without some fairly explicit set of measurements.

Let's take the first. Management by walking around, I would argue, is a form of management without measurement that can be quite effective; you could measure it to prove its effectiveness, but it is in many cases commonsensically obvious. Fear and intimidation are very effective means of maangement. So is role-modelling. So is story-telling. So are strategic intent and BHAG (big hairy audaciuos goals).

The cult of "management requires measurement" is a lot like "prediction requires explanation" in the sciences. Any guy whose bunion hurts before it rains doesn't need an explanation to predict. Nor did cavemen need to understand the solar system to know which way the sun rose. It is a conceit of our intellectual times to believe otherwise.

Now take the other notion, that you can't move forward without explicit measurements. A great case example is Toyota's development of the Lexus. The first thing they did was decide who the competition was: they decided Mercedes. Next they identified key attributes--speed, power, style, etc. and ranked them. Some measurement going on there, to be sure--basic market research stuff.

Then they did the critical thing: they determined that the car-to-be had to exceed the Mercedes on every single one of five key dimensions--speed, power, acceleration, weight, and price (lower). They then set about managing to get there.

What it took was innovation and stretch goals. It doesn't seem to me that it had much to do with measurement, other than beating the obvious performance and price metrics.

I'm not quibbling with the usefulness of measurements, and certainly not their ubiquity, just the often-almost-sacred assumption that they "must" be present for management to occur.

It seems to me that, in some very simple but profound, daily, commonsensical examples--it just ain't true. Measurement is valuable in context; but it's not a necessary, or even a sufficient condition, for good management.

Posted by: Charles H. Green at December 12, 2006 9:03 AM

Post a comment