Forced Performance Distribution: A Failure of Leadership

I am often asked for my thoughts on using a forced distribution (which typically involves requiring managers to identify a certain percentage of their reporting employees as performing poorly, a certain percentage who are performing exceptionally, etc.) for performance appraisal.  This practice gained particular fame and popularity based on the endorsement of Jack Welch, who used it extensively during his tenure at GE. 

I think it's a really bad idea, for a host of reasons, but I have never been able to articulate my objections as well as Ed Lawler, who shares the following thoughts on this approach in his book Treat People Right:

I believe that the forced distribution approach is a bureaucratic solution to a serious leadership failure.  It ignores the reality that in some work groups there are no poor performers and in others there are no good performers.  It causes managers to disown the appraisal event and to essentially say, "I was just following the rules."  Finally, it leads to a kind of unfair and unreasonable treatment of employees that moves the organization signifcantly away from a virtuous spiral environment toward one that fosters survival of the most political or luckiest.  It also can lead to lawsuits because it can be considered unfair to whomever ends up in the lowest-rated group.  After lawsuits were brought against Ford and Goodyear on this charge, both companies abandoned their forced distribution systems.

Given these problems, why do companies use the forced distribution approach?  The answer is simple but not particularly flattering to many managers.  It represents an easy answer to solving a classic problem:  rating inflation.  Just as in universities where professors tend to give high grades to everyone, many managers find it easier to be generous with high ratings, and as a result, many organizations suffer from top-heavy performance appraisal scores.

Because it is a leadership problem, the best solution rests in creating effective leadership rather than the top-down bureaucratic mandate of a forced distribution system.  Mandating a certain distribution is a second leadership failure that just compounds the problem.

Here, here!

Comments

[Michael Haberman, SPHR]

Ann, I too am a fan of Edward Lawler. You quoting him made me want to ask you this question. If I were a non-comp specialist who wanted to quickly get up to speed on the area, what top 5 books would you recommend? Basics and/or thought leaders, whomever you think we should study. Hope this makes for a post topic for you.

BTW, I have not forgotten about the 6 word bio, I just can't seem to capture one.. but I will get there.

[Ann Bares]

Michael:

Thanks for the question; it's a good one and a great suggestion for a future post. I don't have have a top five book list that I can recommend off the top of my head at this time, but it is something I should definitely give some thought to. With one exception: My top five would definitely include The Reward Plan Advantage by Jerry McAdams. I had the priviledge of working with Jerry during my time at Watson Wyatt and I think his book is an outstanding one, particularly for non-comp specialists. It is wise, funny and real. I have given away at least a dozen copies to clients.

No worries on the 6 word bio - I'll look forward to hearing what you come up with when you do get there.

[Dick Grote]

I’m an advocate of creating well-designed distribution guidelines for the distribution of performance appraisal ratings.

The most common reason that companies adopt some kind of forced distribution requirement for performance appraisal ratings (whether rigidly required or more loosely offered as “guidelines”) is because their compensation system employs a pay-for-performance philosophy and relies on performance appraisal ratings as the basis for distributing merit increases. When performance appraisal ratings determine the amount of merit increase an individual will get, managers are under pressure to rate everyone as high as possible in order to maximize the dollars flowing to every member of the team.

Top executives and compensation specialists want to see managers make significant differentiations among their subordinates reflecting the quality of their performance; managers however routinely discover that their lives are less stressful if they prevaricate about the quality of Suzie’s performance on the official form, award her as big an increase as they can, and then deal with her performance deficiencies on an off-the-record basis.

Ed Lawler says, “I believe that the forced distribution approach is a bureaucratic solution to a serious leadership failure. It ignores the reality that in some work groups there are no poor performers and in others there are no good performers. It causes managers to disown the appraisal event and to essentially say, ‘I was just following the rules.’”

Lawler’s right, of course. In some work groups there are no poor performers and in some there are no good ones. But these tend to be the exceptions. For the most part, within any given group of people there will be some who excel and some who lag, with most meeting the expectations that have been set. And tough-minded reviews of all performance appraisals and ratings conducted by the organization’s senior leadership, with apparently-inflated appraisals routinely kicked back for rework, will be the best solution to the ratings-inflation concern. But even that won’t stop spineless managers from telling underlings, “I put you in for a superior rating but the big boss (or HR) kicked it back.”

Forced distribution requirements can have an appropriate place in the design of performance appraisal systems, particularly when they are structured in such a way as to recognize that in any work group there are likely to be many more good performers than poor ones, and they provide a mechanism for managers of exceptional work teams to secure an exemption from the guidelines’ requirements through demonstrating genuine excellence in business results.

Dick Grote

[Frank Giancola]

Ann,

By way of background...

In reading Jack: Straight from the Gut (pg 158-161), Welch and GE were primarily interested in identifying the bottom 10% of performers in a group, and in Welch's words, they "generally had to go." The process of identifying the bottom group repeated itself each year, with a new group of bottom performers being identified each year. By the third year, it was "war" with managers in their employee development sessions.

In terms of good books on compensation, you'd think that Edward Lawler would have written some good ones (he averages about one every 2-3 years and they are very well-wriiten and have sound ideas), but I have yet to find any ideas that are uniquely his and are important. One of his main themes since the 1980s is skill-based pay and that has not taken hold. Jack Welch has probably had more effect on HR practices than he has.

Frank Giancola

[Ann Bares]

Dick:

Thanks for visiting and sharing your thoughts!

I like your distinction between forced distribution and distribution guidelines. I can appreciate that the latter, when appropriately positioned, can offer a helpful benchmark or reference against which managers can gauge their performance management efforts and outcomes.

I also agree - if the issue at hand is management accountability, and I think it is - that the best solution is the one you mention: tough minded oversight by the organization's senior leadership. In organizations where I see this practiced, the benefits are twofold: Not only are performance conversations more rooted in truth and reality, but there is also a clear message that the organization takes performance management seriously.

For sure, tying salary increases - or any kind of reward - to performance ratings results in a whole new kind of pressure on the process. And I appreciate the challenge this presents to many managers. I also believe, however, that managers who cave in to the pressure and give everyone high ratings - deserved or not - are not only abdicating their responsibilities as stewards of the organization's resources, but they are letting down their top performers. And this is indeed a leadership failure.

[Ann Bares]

Frank,

Thanks for the additional background on GE. I'd heard that before as well. It doesn't help make me a fan of the Welch way. But that is another topic.

I am a Lawler fan, and have read many of his articles, but only one of his books: Treat People Right (the one I quoted from). I probably find myself in your camp - while he is obviously one of the best thinkers out there on topics related to people in organizations, I also struggle with some of his compensation notions - including the strong emphasis on people-based and skill/competency-based pay.

[Frank Giancola]

Ann,

Final thought on good compensation books. It's odd that in the field of compensation management, people don't generally speak of a "bible" in the field. Do we?

There are good books on the subject and Lawler's books and McAdams' book are amomg the best. The compensation handbooks by Berger are good also. WorldatWork recently published a handbook that I haven't seen.

If I were to recommend one book to someone new in the field, covering all topics in depth, it would be Richard Henderson's textbook, Compensation Management in a Knowledge-Based World, which may go out of publication since the author died recently.

Frank

[Ann Bares]

Frank:

I don't think we do have a bible, per se, in this field. The Henderson textbook may come close - I'll have to take a fresh look at it. When I was teaching college-level compensation courses, we used Milkovich & Newman's textbook, Compensation, which is a great text, but a little heavy (literally and figuratively). Appreciate your additional thoughts, will take time to ponder Michael's challenge and the best response!

[Wally Bock]

Grading on a curve, whether at school or at work, is an abdication of responsibility or proof of incompetence or both.

[Ann Bares]

Wally:

We are in agreement! Thanks for reading and sharing the thought!

[Frank Giancola]

the "latest" from Edward Lawler:

(March 24, 2008) Bureaucratic inertia has slowed the adoption of skill-based or "people-based" pay in America's companies, but Ed Lawler thinks it's the compensation system of the future. Lawler, professor and director of the Center for Effective Organizations at the University of Southern California's Marshall School of Business, highlights this gradually-emerging trend in his new book, Talent: Making People Your Competitive Advantage. (The book will be available in April from Jossey-Bass.)

It's not easy to make the transition from job-based pay to a structure in which employees are compensated on the basis of the value of their talents. The long habit of paying staff on the basis of defined job descriptions is a hard one to break. Skill-based pay is most often instituted in start-ups and in corporations that are making top-to-bottom transformations. But Lawler looks to the examples of well-known companies that have made the shift, including P & G, Pepsico, and Frito-Lay, as evidence that skill-based pay will prevail over time.

Lawler observes that paying people on the basis of their suite of talents found its most obvious application in manufacturing. The shrinking of America's rust belt is one of the reasons that skill-based pay is not ubiquitous. An irony in Lawler's analysis is that manufacturing industries counted their assets in "tangibles" - dollars and equipment. But Lawler says that as much as 80% of the assets of today's American companies take the form of the knowledge of their workforces. For him, this makes skill-based pay all the more compelling.

"My new book is about the step beyond talent management. It's about the total design of organizations. You can't just insert talent management. Skill-based pay is part of a larger organizational design," he says. "It's a more personal, sensitive system of compensation."

Frank Giancola

[Ann Bares]

So it's only bureaucratic inertia that has slowed the progress of skill-based pay? Hmmm...

I wonder how pervasive this kind of pay approach (people based pay) really is at P&G, Pepsico and Frito Lay. It would be interesting to find out.

Thanks for sharing, Frank!

[Bernie]

A little help, please:
A client just asked about the range of distributions among ratings for large companies like GE, Sun, etc. Does anyone know of a repository for this information?
As for my own position, I agree with Ed Lawler's position. Quite simply forced rankings and ratings depends upon the business situation. Consider the distribution that is appropriate for a company that just added 30% to its workforce versus one that surpassed last year's performance by 30% in terms of per person productivity, profitability, etc., versus an organization where productivity, profitability declined by 10%. In large multi-unit businesses all 3 scenarios are likely to be true in any given period!
A distribution that is somehow arrived at or sustained regardless of where the business is, is ipso facto a leadership problem - Although it obviously makes compensation administration very much easier. Moreover, it is one thing to plot a contribution distribution curve when you are dealing with well measured and tightly coupled performance metrics, it is very much another thing entirely when contribution metrics are poorly defined or loosely linked to actual contributions.
The staffing model used by professional sports teams is probably a good one to consider. Better professional sports organizations will spend an inordinate amount of time examining the talent they have and the talent that is out there. Not surprisingly, the tone and tenor of the discussion will be very different at the Miami Dolphins and the New York Giants.
There is a huge difference between being wedded to a data framing method like forced distribution and a process that ensures sufficient time, energy and care is invested in people decisions. I think people genuinely confuse the two and, in general, are unprepared for doing this the right way. Therefore, the requirement to shed the bottom 10% or 5% or 20%, I believe, is more for mind share and to signal the seriousness of the task than for administrative purposes. We can't blame the compensation folks for pushing for something simple to implement, but we should really question when business unit "leaders" allow it to happen.

[Ann Bares]

Bernie:

Thanks for posting the question.

I am not aware of any repository of information on performance rating distribution for large companies. Even more, if I did, I think I would advise you against relying too heavily on information like this - for all the reasons you mention above and a number of others. As you wisely point out, an ideal performance rating distribution (in a world of accuracy around performance measurement) should uniquely reflect an organization's population and particular performance challenges. As one consultant to another, we know that our clients often seek the easy answers they hope to find in "benchmark" information like this, and avoid the difficult work (and challenging discussion) necessary to figure out the path that best reflects their strategy, needs and reality. Our job, as their external advisors, is to not allow them off the hook too easily, but rather to help guide them in finding the answers that are right for them.

Good luck!

[Bernie]

Ann
Many thanks for the quick response. In general I agree and this client is particularly wedded to benchmarking. The dilemma is that in order to get them to seriously think it through I need something more than my logic and good looks (LOL)! At the moment they are focusing on GE as a potential model, even though I have pointed out that GE hardly consistently practices what Jack Welch preaches at levels much below senior manager -- they simply couldn't replace those engineers and specialists in today;s labor market.

So I am still looking for a repository or research paper that has examined this question.

[Ann Bares]

Bernie:

Let us know if you find that information somewhere!

[Bernie]

Ann
Dick Grote gives a five rating scale in Figure 6.1 on page 148 (of his book on Forced Ranking) together with his recommended starting point for a distribution. But unfortunately he provides no other references or examples. Interestingly I would have imagined that the shape of this curve should reflect the actual distribution of performance in well measured fields like sales, service calls, underwriting, etc., and that therefore compensation folks may have a lot to say about it. The notion that it is skewed distribution, i.e., fewer people in the bottom two categories (~10 to 20%) than the top two categories (~20 to 40%)seems problematic and without logical or empirical foundations since it assumes an asymmetical difference in mobility rates - which may or may not be the case.

[Rick]

Bernie: According to research by Corporate Leadership Council (Oct 2004, Trends in Performance Management Rating Scales): Sun re-allocated it's ratings to mirror GE's 3-point rating scale (Top-20, middle-70, bottom-10)