There is a general feeling that performance appraisals are subjective and that when managers rate their team members, a certain bit of bias seems to creep in.  This thought process leads to un-necessary friction between the managers and employees.

Can we eliminate the subjectivity in performance appraisals by automating the rating?

Sometimes, people ask us if we can automate the rating completely using our software.  So, they expect that the rating should be automatically given based on employee’s performance and pre-defined rating for such performance and managers would not be responsible for rating the performance.

So, can rating automation be a solution to remove the subjectivity in appraisals? In reality, will it be possible to remove the subjectivity element and automate the rating completely, without a manager’s involvement?  This is definitely a point to ponder on.

First, let us understand the parameters or the dimensions which are rated, as part of the performance appraisals.  Generally, there are set of performance dimensions or goals that an employee is required to achieve during the course of a year.  Also, there are a set of behavioural dimensions which an employee needs to exhibit to fulfil these goals.  So, generally, rating is given to both these dimensions, but the weightage may vary between both.

Problems in automating the rating

The following are some practical problems that we will face, when we try to automate the rating during a performance appraisal:

1)      Performance dimensions or goals are generally quantifiable.  But, for employees working in some departments, setting SMART goals may be quite a challenging task.  For instance, defining quantifiable goals for a sales department is much more easier, than defining goals for people working in a functional team.

2)      Even when goals are quantifiable, setting proper performance measures might be difficult.  This will lead to difficulties in finding out the actual achieved numbers. So the input of the actual achieved numbers (against a target) would be done by a manager or an employee manually, and there is bound to be some subjectivity in entering those numbers.

3)      Behavioural dimensions, like communication skills, cannot be quantified. So, automation of rating cannot be considered for rating behavioural dimensions, and so the bias element cannot be totally eliminated.

4)      Even if we consider that automation of rating is possible, the inability to meet the performance target by an employee might be due to various external reasons outside of the control of the employee.  Penalizing a person for this by automating a rating might be too harsh.  Some subjectivity of manager involvement in such cases would be better.