This post is part of the series on 9 Realities of Modern Workplace.
In this post, we talk about Reality #5: “There is always a stack ranking and a bell curve of performance rating, even in companies that claim they don’t have these“.
Let’s define the terms first.
Stack ranking is ordering employees in decreasing order of their ‘value’ to the organization. Typically, this ‘value’ is measured by performance during the review period, but it can incorporate other factors like potential, yrs. of experience so far, etc. (see why this is a bad or a good thing)
Bell Curve is force-fitting the rating of employees to a certain distribution: X% of employees have 4 or above, Y% have 3 or above, Z% have 2 or above, etc. (see below for math, or Forbes for why this might be wrong way of looking at employees).
Both of these are applied to an identified group of employees, mostly by grouping 2-3 nearby job levels into one group.
Let’s take an example of a group of 10 people (by the way, 10 is too small a sample size to apply these concepts, this is just for illustration).
||Performance Rating (out of 5)
Note couple of things:
- It is not necessary that highest performance rating corresponds to highest stack ranking. Performance rating reflects performance against defined goals for the period. Stack Rank reflects ‘value’, which may be more than just performance in the given year (a recent hire who got 4 because he had only 4 months to show his performance may be the best bet for the company and hence can be ranked ahead of a senior guy who also got 4, who in turn may be ranked ahead of an average guy who somehow delivered exceptional performance this year and got a 5).
- In this sample, 20% got 5, 40% got 4, 30% got 3, and 10% got 2. Put another way, assuming 4 and above is excellent performance, 60% of the sample produced excellent results. Statistically speaking, this number is lop-sided – excellent performances are usually not so abundant.
As discussed in Reality #3, performance rating systems in most organizations are broken, which means #1 will be similarly broken.
According to the mathematics behind bell curve, the distribution in #2 above (if it was a large enough sample size) is an anomaly, usually created because managers are not strict enough in their evaluation. Hence managers are asked to change their rating to ‘fit the curve’, usually like this: 20% top performers, 70% average performers, 10% bottom performers, which makes a broken rating process more messed up and subjective.
Why would an organization want to have a stack ranking?
Looking from organization’s perspective, reason is simple to understand. As discussed in Reality #4, an organization will pay the minimum it needs to pay to keep you and will want to get the maximum out of you. Stack ranking is a good way to identify the employees who provide best ROI to the organization. It makes sense to identify them, reward them, and make best use of them.
Whether the above exercise happens in a formal way or not, organizations need this information and so managers will always keep such a sheet handy which tells them about performance rating and stack ranking of employees so that they can do their investments in employees accordingly. This is natural.
Implications of Stack Ranking and Bell Curve
Here are some of the implications that should be obvious now:
|You have done great work, expect a 5 out of 5 performance rating, but end up with a 4.
||Company decided they can’t have so many people with 5 rating, and your rating was downgraded in the name of ‘curve fitting’.
|Your manager says, “I wanted to give you 5, but management decided otherwise”
||In the meeting where the rating rationalization is done, your manager (or his manager who was representing you) was unable to convince others about your rating.
|You get a 4 but the person who got 3 gets more rewards than you
||Your company rewards favor the stack rank more than performance rating
Dealing with the reality
Good thing is that there is enough buzz in the air that these are bad things – Microsoft recently got rid of both rating and stack ranking, Adobe got rid of its annual performance review system. However, you still need to understand the rationale and deal with them, because as I mentioned above, even when these are formally gone, they exist informally, and so they impact you. Best way to deal with these is to work closely with your manager (or manager’s manager) to make sure you understand where you are in that ‘valuable employees’ list. Also, it is more important to observe the behavior, just listening to your manager may not help much. Watch for these:
- When a new/interesting project comes, are you asked to work on it?
- When you ask for a reassignment to an interesting project, how hard it is to convince the decision-maker?
- How much do you have to haggle with your manager for bonuses/raises?
- How many of your ideas/thoughts are entertained or accepted by your manager?
If you think you are not high enough in that ‘valuable employee’ list, you need to change something.
In the next post, we will discuss the Reality #6: “The new hire can replace you any day if your only strength is technology“. Stay tuned.
Maths behind bell curve
A theory called Central Limit Theorem suggests that if you take random observations from a large enough sample size, their average values, when plotted on a graph, will resemble a bell curve (normal distribution). Applied in performance rating context, it means that with a large enough number of employees in a group (large sample size), their performance rating will be a normal distribution and produce a bell curve.