Stack ranking was the primary performance review process used at Microsoft engineering organizations before the recent cut. Employees are compared with each other to determine the rank/score, the distribution of the review score of a certain organisation falls into a bell curve, a few exceptionally good, a few exceptionally bad, most of us get middle score. And the employee gets the best score gets most rewards, whereas those get the worst score get little or no rewards. The purpose of stack ranking system is to encourage competition among employees so that the whole team can achieve higher level of performance. 

When does stack ranking system become unfair?

Stack ranking system can become unfair under a few circumstances. The first one is when the actual performance of the employees do not fall into the bell curve. Although statistically, at a large organisation, the performance of all the employees are likely fall into a bell curve. However, at a particular team, at a specific time, there is no guarantee that the performance of them fall into a the bell curve. They could all be performing well, or they could all be performing bad. When all the team members are performing well, they will feel they are under paid, thus this creates unfairness. When all the team members are performing poorly, they still get rewards like high performers at other teams, other high performing team will feel unfair. 

The second situation is when the employees are not competing against each other on productivity. When people feel pressure, depending on their culture, education background, they can adopt dramatically different behaviors, some of them can be counter productive. The feeling that some people on the team are not pulling their weight for the business also makes the rest of the team feeling unfair.