Pitfalls in Performance Based Bonuses

Performance based bonuses (PBB) is an effective management tool but quite often it causes unintentional effects. PBB are supposed to improve productivity by compensating employees according to their contribution to productivity. However organizations are complex systems and it is impossible to measure an employee’s contribution accurately. Companies are forced to use performance models that are a simplification of reality. As a result, employees are focused on maximizing performance under the model which not completely aligned with the benefit of the company. In banking, for example, the bonus models underemphasized risk, causing a worldwide catastrophe.
Leaving the financial crisis aside, there are other common pitfalls. It is simple to construct performance-based contribution for sales representatives. The more they sell the more they earn – simple. It is much harder to construct a model for back-end functions, say an IS developer. This is where it gets messy.

Measurement by schedule

The two most common schedule bonuses are a bonus for timely completion and a bonus for accurate planning. Combined together these two bonuses can cause a funny outcome. Well funny for a bystander, not nearly as funny for the company. This is especially true when the manager that plans the task and dictates a timeline is measured for performance according to her own predictions and plans. More often than not, that is the case. Obviously, if a manager is measured by meeting her own timeline she has a strong incentive to insert large buffers to the plan. This by itself is not a serious problem. In the worst case, the manager’s team will finish ahead of time and will still be congratulated for a job well done. However, what happens if the manager is also measured for accuracy of plan? In that case, finishing before time is not desirable. The team gets the feeling that they are expected to utilize all the allocated time and will slow down the intensity of their work and even be “quietly” idle at times. At the end, everyone is happy all the tasks are “green” (finishing in time), the predictions self-fulfill, workers don’t have to work and the manager gets a hefty bonus.
Corporate has managed to create an inhibiting bonus system. When executives look at their colorful dashboards (that show goals), everything looks fine. I call it the self-fulfilling paradox of over planning.

Measurement by defects

Imagine a defect found in UAT (user acceptance tests) that comes down to a discrepancy between the code, the functional specification and the data model. Who should be assigned the defect? Many times it is easier to fix the defect than to answer this question, especially when the answer affects income. The focus shifts from solving the problem to assigning blame. The defect is tossed around from one person to the other. Such a defect is likely to have more than one “parent”: the data modeler for not synchronizing with the functional specification, the analyst that has an overall responsibility for the design, and the QA who did not spot it. Should you assign it to multiple people? That would add a serious complexity to the defect handling process. Wouldn’t it be easier to just assign the defect to the person who is most likely to solve it?
What about when an employee detects a defect that he is responsible for, in which case the resolution involves formally opening a defect on oneself? Say, a developer who found a problem with a code piece that is already in production. Nobody but him noticed it so far. If he gets lucky maybe nobody will until such a time when he has the chance to slide the fix into the next patch. When it affects his bonus, the developer has a strong incentive to let it slide.
Rewarding for opening defects can cause a profusion of pettiness. In matters of form and documentation there are countless ways of doing things and therefore there are many opportunities to open defects. I have seen this turn quick informal feedback loops into a red-tape burdening communication.

So what to do

‘How to do it right’ is a hard question. One thing is for sure, trying to develop a measurement system that takes everything into account is not the right approach. Additional complexity to measurement systems increases maintenance cost. Time spent on maintaining complex metrics can out-value the increase in accuracy.
Metrics should be used as a reference but the ultimate bonus decisions should be left to direct managers. Especially in parts of the organization that are not profit centers. There is no substitute for good and honest management trusted by subordinates to make decisions based on insight and integrity.