Incentives are often used to mitigate what is known in economics as the principal-agent problem, which is just a fancy way of saying “managers want employees to work really hard, but employees would rather sit on the couch and eat cheetos.”
There are two ways of overcoming this misalignment of incentives. One is for the manager to sit and watch the employee and make sure that he is working hard. Not surprisingly, this is pretty inefficient, since a lot of the point of the manager hiring an employee is so that he can run off to the golf course or whatever rather than being stuck in the office. Also, who then watches the manager to make sure that he is vigilant enough in watching the employees?
This is where incentives (or pay-for-performance, in management speak) becomes relevant. Most incentive systems implicitly acknowledge the fact that it’s difficult and/or irrelevant to see the employee’s inputs in terms of effort, ability, etc. and instead give a payout based on output. This gives the employee a reason to not shirk (a fancy term for sitting on the couch and eating cheetos) and better aligns the incentives of the employee with those of the manager.
In practice, it is the case that pay-for-performance is a bigger part of an employee’s compensation when the underlying level of effort and competence is difficult to observe. For example, salespeople who are on the road talking to clients all day are paid almost exclusively on commission, since what else (at least in the short term) is stopping them from going and hanging out at a sports bar instead and then reporting back that people just weren’t interested in the product? In other words, with lack of proper incentives you run the risk of getting something like this:
Don’t say I didn’t warn you.