[loewenskind]

It is good at lock in ("sheesh, I've been here for 8 months, I've only got 4 to go. If I leave I could be leaving $20k on the table"), especially since bonus payouts are often months after hearing the number so you've already began earning the next one before you get your current one paid.

The problem is, when you get that "big" bonus and it's not big then you can't help but be disappointed. I recently saw a company lose about 5 key people at one time because of the disappointment caused by this. Before bonus time those people were happy with their job and even their pay. But they all got nice bonus the year before.

In my estimation, that's 5 people they lost by a non-existent problem they created themselves.

[doki_pen]

I call that doing it wrong. When you are hiring, you need to promote the bonus as part of the salary, and make sure you pay most of it, even if tough times. You can skip the raise, but don't skip the bonus. That is penny-wise, pound foolish.

[cynicalkane]

At my company, bonuses are based on a standardized bonus earnings target. The target is unpublished, but the company keeps people updated on the approximate chance of the bonus meeting certain percentage benchmarks. Keeps people from being unpleasantly (or pleasantly) surprised by too great a degree.

When bonuses are paid out this way, I see them more as a way to sort of vary compensation based on market health, while still guaranteeing a fixed salary. They're performance incentives inasmuch as employees are compensated based partly on performance and partly on how well the industry is doing, making compensation market-driven in a loose sort of way.