We've spent years giving speeches and writing articles about managing for results, explaining the fine points of establishing mission statements, goals and targets. We've always claimed to know how difficult this stuff is to implement. But only recently did we have the illuminating experience of discovering for ourselves just how real the obstacles are.

Here's the story. Our 13-year-old daughter plays soccer pretty much all the time she isn't eating, sleeping or doing homework, and we've become managers of a couple of the teams she plays on. The manager's job is an administrative one, while the coach worries about winning games. So we had a bright idea: Why not use managing for results principles to run a soccer team?

First step, we decided, was to establish goals for the team. We chortled a bit about the idea that goals were the goal and immediately fell on the idea that the primary one was to win games. We mentioned this to a couple of other parents. They disagreed. The real mission of a youth soccer team is to develop players, they argued. We decided to combine the two, indicating that the primary goal was to win games while developing players.

But wait. Sometimes developing players means using weaker girls in games and risking a loss. The two goals were inconsistent. And thus we confronted a common obstacle in managing for results. It wasn't dissimilar to the problem most states face in trying to accomplish two of the primary goals for their Medicaid programs: maintaining access and cutting costs. Doing both simultaneously is very difficult. Texas, for example, has successfully put cost control at the forefront but has lost ground on access as tens of thousands of people have been dropped from its Medicaid rolls.

All right, so maybe we would have to forget about finding a thoroughly consistent set of goals. We'd have to live with ambiguity. But that didn't mean we couldn't come up with some solid performance measures for the players. We decided to start with the forwards. Their performance could be measured by the number of goals they scored, right?

Not so fast. First of all, the number of goals they scored was not a measure of how much they were developing but of their contribution to winning games. A better measure of development would be shots on goal. Moreover, if the only way for the forwards to get credit was by scoring goals themselves, there'd be no benefit in passing the ball. And occasionally passing instead of always taking shots would unquestionably result in more points for the team. Such trade-offs are also a risk for governments when they create measures. Dallas County, for example, gives credit to judges who run their courtrooms efficiently: keeping costs down while moving cases at a steady pace. But judges have complained that the emphasis on efficiency might hamper the quality of judicial decisions, although county managers say this hasn't happened.

This one, though, we were able to figure out. We'd measure the success of our forwards by the total number of goals scored for the team as well as the number of shots taken. Those are probably good measures. But they also confront another landmine in the performance- measurement field: The forwards were going to inevitably complain that scoring goals wasn't entirely their responsibility. They can't score, for example, if the midfielders don't get the ball to them. This is just like police departments that resist being held accountable for crime rates. Their rationale is that there are so many other agencies involved in reducing crime, including the schools, that it's unfair to measure their performance on this one criteria.

Here the parallels to government gave us some courage. If you're going to measure the success of a forward, this was the way to do it, even if they didn't have complete control of the process. It turns out that cities such as New York that have resolutely tied success in dealing with crime to the police departments have tended to be more successful.

We slogged along with the other position players. Measures for the defensive players were reasonably easy to come by. They were somewhat harder to define for the midfielders. Somewhere along the line our daughter heard these discussions and indicated that if she thought she was being measured every time her foot touched a ball, she'd be too scared to play well altogether. That sounded familiar: How many government employees resist managing for results for fear that it will be used punitively? We took the same line that government managers take: This information will be used by the coach only to help run the team better--not to make life difficult for the players. We don't think she believed us.

The next trick, of course, is getting the coach to buy in. That should be just about as easy as getting elected officials to run their governments with results measures. Wish us luck.