At the heart of performance improvement lies the twin concepts that performance is both measurable and that it can be impacted positively in some fashion, normally by the employer. For companies who employ knowledge workers, measuring performance can pose as much of a challenge as determining how to improve it. When the relative performance of an employee, division or the company as a whole can only be inferred from externalities, performance improvement initiatives often bog down with debate over how to accurately assess the performance and any improvements. Add in bonus or compensation structures that are based in part on performance improvement and the defining of performance can become the focus of the exercise, with improvement regulated to the corner for some day when we have time to think about it.

The challenge for executives in such a situation is to find ways in which to move the discussion away from solely focusing on reliability and include validity. Measures of reliability (or the appearance of it) hold tremendous sway in our society and discussions of performance inherently focus on reliable ways of measuring it. If an employee does x, does y result? If an employee does 2x, does 2y result, or is it only 1.5y? It is in precisely this area that measuring performance and subsequent improvement becomes very difficult for companies that are involved in knowledge work. Unlike for employees that produce x units of a product, the average knowledge worker cannot say with great precision how good his or her performance was on a particular day or week, as the project or task may not lend itself to easy quantification.

When a manager is seeking reliability, variables that are difficult to measure quantitatively such as feelings and relationships are taken out of the picture. By assessing performance improvement solely from a reliability standpoint, companies can end up with strict, reproducible definitions of what an improvement looks like that no one actually feels is an improvement. It may meet all of the criteria of an improvement, but when it encounters the world outside, the result is a resounding flop, with people questioning how on earth that could be considered an improvement.

Contrast this sole focus on reliability with the inclusion of validity. Rather than dismissing variables that are not easily quantifiable, the manager focuses instead on ways to measure the validity of the variables. If as a result of making a change, a division does their job in a fashion that other employees feel is a marked improvement, does the absence of a quantifiable measure of improvement mean no improvement took place? A simple method of combining aspects of reliability with validity is the use of surveys and questionnaires to assess performance improvement. If customers are asked to assess the company’s performance in a particular aspect, changes are introduced and a new survey shows that customers feel the company is doing worse in that regard, clearly there was a performance deterioration. Even in the absence of quantifiable measurements (or occasionally in contradiction to what quantifiably looks to be an improvement) the company can say with strong validity and confidence that the changes did not improve performance.

Another area where companies often encounter problems in performance improvement is confusing an improvement in the activities designed to improve performance with actual performance improvement. This confusion results from the fact that many employees view their job as a matter of completing tasks, rather than outputs. A training manager for a company who is focused on tasks may consider herself to have done her job if the company holds 17 training sessions during the year. The same role looked at from an output perspective views number of training sessions as the means of accomplishing the output of the role, namely unaided recall of key learnings the company wants employees to know.

While an improvement in the tasks done to produce an output may seem worthy of reward, the criteria that matters is whether that process improvement results in output improvement. An enhanced training program that covers off more topics but does not result in higher levels of unaided recall of key learnings by the staff, is in fact not an improvement that matters. In the same manner, the provision of shorter, more focused sessions that results in higher levels of unaided recall is clearly an improvement. Managers need to be careful that the areas of performance improvements being measured and rewarded and in fact that areas that matter to the company.

Surje & Company works with clients to ensure that the performance improvement process remains true to the desired goals of the company and that managers are able to move past the sole focus on reliability to target and achieve valid and meaningful improvements.