Nicole - Great Topic!
First, decide what you are trying to measure. As you develop metrics, be sure you are actually measuring what you think you are. And, be careful of unintended consequences. Measurement will drive behavior, but not necessarily the behavior you expect. Also - besides individual metrics, I've found that TEAM metrics can drive behavior by having the team members put pressure on each other to drive the overall score up.
That said, a couple of metrics come to mind, but it all depends on the work that your teams do.
For work order driven teams, turnaround time (entry to closure) measured against the Service Level Expectation or Agreement (SLA) is a great measure.
Another metric could be the difference between entry and actual start date.
On time delivery, as already mentioned, can be a great metric, but needs to be tempered with reality. I still think it has a place, and overall the average time to deliver will likely find its place even if outside forces interfere in one or two tasks. If there are these external forces, perhaps requiring an issue logged if a task is going to be late, then you can ignore (and audit) those that have an issue logged to the task.
So, again, much of this depends on how you define "Performance in WF", and really goes back to change management. Set the expectations. Measure against them. and POST the measurements for all to see. Spotlight effect can be HIGHLY motivating, but be careful about unintended behaviors.