Why Forced Ranking Doesn’t Make Sense
- March 31st, 2010
- Posted in Systems Thinking Concepts . Systems Thinking and Measures
- Write comment
Like there isn’t enough politics in the workplace . . . command and control managers love to rank employees. There needs to be forced ranking by assessment of performance to be a good manager and have a well-run company.
Some rank to give bonuses or incentives and others rank to RIF employees (have to get rid of the bad ones). I never have found good reason to rank and always advise against it. It is a bad practice that leads to waste and sub-optimization.
I won’t dispute that there is always someone at the top and the bottom in performance of any entity. But the waste of performance appraisals, competition, back-stabbing and manipulation far outweighs any conceivable benefit. Money and morale is lost with these activities.
The distribution of people and there performance is typically bell-shaped:
The diagram above indicates that the winners get the bonuses and the losers get cut. But what if we take a different approach and instead of focusing on forced distribution, we focus on improving the design and management of the work.
Instead of a select few we make the work better for everyone and all benefit including the company. A small shift in a system improvement of performance always outweighs moving a few of the “high performers.”
In working with a client recently that “ranked and rated” employees, I was told that this ranking is what made the company so great. The first thing is that when we looked at actual performance it was atrocious and customers we spoke with told us so. They didn’t know how bad their performance was until we started studying customer demand.
But this organization swore by the ranking and rating that “made their company great.” In fact, they were successful in spite of the ranking and rating, not because of it.
You see there are three scenarios that typically separate employee’s performance:
- They really have a better method or are a better match for the job. In this case, why would you want to have forced rating where competition for incentives and bonuses keeps employees from sharing with each other. The end result is individual wins and company loses. If the person is a better fit than others organizations need to learn this and feed it back to the hiring process. Personality (Myers-Briggs Assessment) might be helpful in assessing difference
- There is manipulation at play. Employees learn to game the system by cutting corners and cheating. When rewards or your job is at stake all is fair.
- There is no difference in performance. Over the past 20 years I have found this is usually the case. Very few times do I find better method. And when the incentive to manipulate is taken away, I usually find very little difference in performance.
So how do we know if we have a difference? We chart there performance if data is collected. If no data, then your ranking is subjective and political and a huge waste of resources. There is no objective performance appraisal system despite all the consultants that sell software to do so.
In my most read post Service Metrics: What You Need to Understand I outline how to analyze data. You can do the same thing with employees performance when data is present. It may look something like this:
Any data within the limits (44 and 155), and performance is attributable to the system. This means that working on the system is your biggest opportunity for improvement. The system is comprised of the work design, structure, technology, management thinking, etc.
Our greatest opportunity to improve is to design better work and learn better ways to manage. The old thinking of command and control, ranking and rating is flawed thinking and too many organizations have succumbed to this thinking. If you are looking for a competitive advantage, this may be the one opportunity not to miss.
Leave me a comment. . . share your opinion! Click on comments below.
Tripp Babbitt is a speaker, blogger and consultant to service industry (private and public). His organization helps executives find a better way to make the work work. Download free from www.newsystemsthinking.com “Understanding Your Organization as a System” and gain knowledge of systems thinking or contact us about our intervention services at info@newsystemsthinking.com. Reach him on Twitter at www.twitter.com/TriBabbitt or LinkedIn at www.linkedin.com/in/trippbabbitt.
It’s certain the organisational system within which people work (team and job structures, business process and workflow design) is a vital factor in business performance which can enable or constrain people from delivering performance.
I would agree that managers need to understand how their organisations function as systems …. but that doesn’t mean they shouldn’t, also, understand how employees work as people … and help guide and shape their work within the system.
Appraisal does not, however, have to be about forced ranking and bonuses … though there are many organisations that work this way. Performance management does not have to be about “assigning blame”.
Using appropriate tools and techniques to help people understand method, develop skills and knowledge, identify the need and opportunities for learning and improvement can be very positive. Appraisal does not, in itself, trigger “command and control” thinking.
I would also argue thera are many situations where individual knowledge and skill are as important as systemic factors. This is particularly true where “method” remains unprescribed and the freedom to make judgements, apply experience, explore and experiment with approach are vital to success.
While systems thinking is immensely powerful, no paradigm can, alone, be applied to every situation – it cannot cover everything.
Ackoff says business schools continue management techniques that have “an ability to withstand any amount of disconfirming evidence”. Systems thinkers mausn’t fall into that trap.
Maslow’s quote, “If the only tool you have is a hammer, you tend to see every problem as a nail” applies to us all. I think it’s vital to learn to see the organisation as holistically as possible and through “many lenses”.
Systems thinking should be more widely applied, but it’s important, while making the case, people take care not to reject other valuable knowledge out of hand … just because it doesn’t fit the new paradigm.
Vince-
Thank you for the comment.
You may have over-read my post. I didn’t say that the individual isn’t relevant. In fact, the individual becomes more relevant when decision-making is put back with the work. You may wish to read an earlier post http://blog.newsystemsthinking.com/the-humanator-rise-of-the-front-line-worker/.
True that appraisal does not have to be about forced ranking and blame, but here in the US it is a rare occasion that I find they are not tied.
To be clear, appraisal where people are being given better jobs, sales leads,etc. based on arbitrary assumptions by management when no evidence exists is this issue. The better workers become more of a self-fulfilling prophecy than a statement of fact.
The individual skill and experience does matter (although I would argue as Deming did that experience by itself teaches nothing) to me this is still method. Skill is method from learning and others can learn better method. I use the analogy of “touch” in golf as back-swing and timing.
Systems in the US are full of assumptions that wielding control over the worker explicitly or implicitly will improve the system. This IS command and control thinking.
I attempt not to be sanctimonious about systems thinking, but it is a way of thinking that challenges many core assumptions and paradigms. Whether or not they apply or not . . . to me depends on each system as I study them and see the effect.
Regards, Tripp
Hi Tripp
I’m not sure if I interpret your example control chart correctly. If it shows the performance of a single person over time, I agree that UCL and LCL serve to predict performance and discover special cause events.
What if your x-axis is not time but represents different people? So the graph would plot the (average) performance of different people next to each other. Do you still think that UCL and LCL make sense in such an example? Would it still be right to say, in your opinion, that differences in average (!) performance are due to systems conditions, as long as the variation between the data points is within the control limits? Or would it be wrong to use a control chart to compare average performance of different people? After all, it seems intuitive that if you compare people’s (long-term) AVERAGE performance, the differences are really related to the individual, not the system, even if the variation is within the control limits. That’s how call-centers usually justify all these AHT-comparisons. They say, as long as we compare the agent’s long-term AVERAGE AHT, we will have objective performance comparisons.
Would be great to have a clarification from your side, I thought about this before when reading some of your previous blogs on SPC.
Regards, Frank
Hi Frank-
This is a good question. Thanks for bringing it up. I want to say that I want to keep answers high-level because if we get to the Six Sigma world we wind up fighting over statistics and not find ways to improve systems. This is a dead end.
Let me start by saying that SPC is helpful in discerning how a system performs. The bottom-line is that we gain insight to asking better questions about using data. I usually start with an aggregate number in service industries to assess overall performance knowing that differences exist as I dig deeper. Homogenous data helps separate (by operational definitions) that we are measuring like things. Peeling the onion back we find more.
The question in service to me has to begin with the variety of demand (something that John Seddon showed me, even if he wasn’t schooled in statistics). This variety plays an important role when people try to assess performance individually. The individual SPC charts with different limits and averages many times are caused by this variety of customer demand and not differences in performance. This is why you have to listen to calls to be sure of what is happening. If you are sorting calls by type then we may have less variety by type, but more by customer (the way they present demands are always different). The large variety is unique to service.
I would not use an x-axis of people. I may run charts in aggregate and individually (individual by calls and time) but I usually only do this to prove a point. That said we are talking about different charts meaning the UCL and LCL are relevant. I do not know what causes the differences until I go to the work and see, but usually they can be traced back to system conditions (rate and rank, CYA, etc.).
With regards to AHT, we can not ignore variety of demand and besides AHT works on the wrong problem. How fast we handle a call may not help the customer and call centers should be set-up to help the customer. The right problem to work on is demand. This is where failure demand plays a role. Failure demand is caused by the failure to do something or do something right for a customer (Seddon). When this represents between 25 and 75% of all demands and realize it is systemic, we then work on the right problem. Demand offers us our greatest opportunity to improve.
If we did look at (which I already suggested not to do) AHT averages over time there may be differences . . . but so what. We are still left with a design of work that people are (usually ) performing within limits. Improve the design, get rid of the failure demand (or as much as possible) and change thinking to better customer measures (like end-to-end time for service) and you are on your way to a better system.
I hope this helps.
Tripp
Tripp,
Here in MN the govenor is criticizing the teachers union because they were unwilling to agree to merit pay based on student performance. Polictics aside, do you have any suggestions for ways to ‘educate’ members of the school system about the implications of your column above; i.e that education is a system, a service, and student performance is a function of the system, not the teacher?
Thanks,
Joel Gingery
Joel-
John Seddon spoke to Calfac (California Faculty Association) about a week ago and outlined the problems in education. He spoke more against Deliverology ( a bad UK method), but he also outlined how to approach education. I should have this on my website soon (week or so).
Many of these Republicans including Tony Bennett in Indiana (Republican) see the teacher’s union as the problem. This only sets up a fight and works on the wrong problem. I would encourage you to read http://blog.newsystemsthinking.com/the-waste-of-targets-in-no-child-left-behind/ I need to write more on this subject W. Edwards Deming was against merit pay for good reason . . . it drives the wrong behavior. Yet political hack after political hack seems to dig up the same old theories that don’t work (like shared services).
Thanks, Tripp