Why performance appraisal systems are doomed to fail
Many firms have dropped performance appraisal because it’s just not up to the job
If a supervisor gives performance ratings that are fair and accurate, it is almost certain that the employees who receive them will feel that they are being unfairly criticised and undervalued. Photograph: iStock
Performance appraisal is the job managers and employees love to hate. Managers dislike giving performance feedback and employees dislike receiving it. Surveys of satisfaction with performance appraisal and performance management systems routinely report abysmal results. It is common for 80-90 per cent of managers, executives and human resource professionals to report that the performance appraisal and performance management systems in their organisations are failing.
This almost routine failure of performance appraisal systems is not for lack of trying, as I found in reviewing 100 years of research on performance appraisal and performance management in Journal of Applied Psychology. A number of strategies for improving performance appraisal, ranging from rater training and memory aids to the development of better rating scales have been tried, with generally disappointing results.
Organisations such as Deloitte, Adobe and Accenture have either dropped performance appraisal altogether or have radically simplified their appraisal systems.
Other organisations have gone in the opposite direction, adding multiple layers of evaluation by including peer ratings, self-ratings or ratings from customers and clients as part of a multisource feedback system. Unfortunately, there are few reasons to believe that any of these strategies will succeed.
It is time to stop rearranging the deck chairs; the Titanic is clearly sinking. Rather than continuing to tinker with failing systems, we need to examine the structural flaws that cause performance appraisal systems to fail, and use this information to take radical action.
Two problems are present in virtually every appraisal system which cause them to fail: conflicting purposes and conflicting perspectives.
The first structural flaw is that companies and organisations routinely try to use performance appraisal systems for conflicting purposes. I recently analysed data from a survey of several thousand mid-sized and large organisations in 24 countries. Over 80 per cent have some sort of formal performance appraisal system, and the great majority of them use appraisals for two purposes: identifying training and development needs; and identifying candidates for salary increases, promotions, separations, etc.
But these place conflicting demands on performance appraisal systems. Using appraisals for training and development relies on information about individual strengths and weaknesses – ie differences within individuals – while their use for administrative purposes (eg salary administration) relies on information about differences between people.
It is almost impossible to get clear information about differences between people and differences within people from the same set of performance rating. Ratings that show clear peaks and valleys, distinguishing strengths from weaknesses, will end up producing average ratings that are very close to the middle of the scale for just about all employees, making it difficult to distinguish strong performers from average or weak performers.
People almost always rate their own performance higher than do their peers, their supervisors or almost any other set of external raters
Performance ratings that clearly distinguish good performers from average and poor performers tend to produce flat profiles because they focus on overall performance and effectiveness as opposed to individual strengths and weaknesses.
Organisations that attempt to use the same performance appraisal system for training and development and for salary administration often end up with systems that are not particularly effective for either purpose.
An even more important structural flaw involves differences in perspective. One of the most robust findings from research in this area is that people almost always rate their own performance higher than do their peers, their supervisors or almost any other set of external raters. This self-serving bias is a product of a number of psychological processes (such as a desire to maintain a positive self-image and differences in the way we process information about ourselves as against others) that are strongly ingrained and almost universal in scope (in some Asian cultures, self-ratings are closer to ratings provided by others).
The implications for performance evaluation and feedback are devastating. If a supervisor gives performance ratings that are fair and accurate, it is almost certain that the employees who receive them will feel that they are being unfairly criticised and undervalued. One of the reasons for pervasive rating inflation – in most organisations, 80 per cent or more of the employees who are evaluated receive ratings of “above average” to “excellent” – is that supervisors and managers understand that accurate ratings will lead to conflict and to the rejection of feedback.
Using performance appraisals for high-stakes decisions, such as salary administration, promotion and the like is often a fool’s errand
Are there solutions to these structural problems? In our research, my colleagues and I have recommended two substantial changes in how organisations conduct and use performance appraisals. First, decide whether you want to use performance appraisals for training and development or for salary administration, promotion, and the like. Attempting to do both with the same appraisal system will fail.
Our research suggests that using performance appraisals for high-stakes decisions, such as salary administration, promotion and the like is often a fool’s errand. Organisations that follow this path create strong pressure on supervisors and managers to inflate ratings, making it difficult to meaningfully differentiate superior performance from average and even poor performance.
The research suggests that, if you are going to do performance appraisals at all, you should use them solely as a tool for identifying training and development needs. This means appraisal systems should be ruthlessly tailored to that purpose.
Most performance appraisal systems ask supervisors or managers to rate each employee on several performance dimensions. A much better approach would be to rank-order performance areas, regardless of the employee’s overall performance level. For example, suppose you are asked to rate personal care aides on the key dimensions of their job (communication, assisting others, getting information from clients, documenting and recording client conditions and activities). Rather than rating each dimension individually, it would be better to identify the relative strength of the employee on various aspects of performance. A number of ranking strategies might be applied to provide a detailed picture of each individual’s training and development needs.
A strategy like this means organisations need to give up on using performance ratings to determine salary raises, promotions and the like, but it is clear that attempting to use the same system for both purposes almost always ends up making a mess of both decisions.
It is certainly better to decide which task is important and do it well than to continue to muddle the whole process of performance appraisal and performance management.
Prof Kevin Murphy is the Kemmy chair of work and employment studies at University of Limerick