Hi Srinivas,
I think what you are referring to is a forced ranking system of appraisal. Typically, as per this mechanism, you would get a bell-shaped distribution as follows:
15% - Top performers
20% - Average Top
30% - Average
20% - Average Low
15% - Low performers
Or there could be other distributions as well similar to the one above. For this to work, you would have to consolidate performance across all dimensions into a single rating value to obtain one curve for the entire workforce. As always, the appraisal process should be unbiased, with fair comparisons across work groups and teams.
Regards,
Joshua