...dueling anecdotes aren't likely to be helpful.

I guess I'd just ask what you mean when you say a plan "worked"?  Do you mean the employees did better work than they otherwise would have done?  More work?  The same work, but they got it done sooner?  All of those would be hard to measure--you don't have a control group of matched employees that didn't have an incentive plan.

It's easy to feel like an incentive plan was a success:  everybody worked really hard, they hit their deadline, and the product was a success.  But what if you look at things  a bit more broadly?  Did some of the key people who made that project a success take their bonus money and move on?  Was the next product harder to create, because developers trying to hit their metric took some shortcuts?  Did workers on some other project, who didn't meet their target (and didn't get a bonus) become so demotivated that they started spending their work hours updating their resumes?

Alfie Kohn's book looks at some of these issues--and does it with actual data, rather than just anecdotes.