By Bracken, David W.
Training & Development , Vol. 46, No. 6
The total quality movement has pushed benchmarking into the spotlight. Benchmarking is a systematic process for comparing some aspect of an organization against that of a company that is considered to be superior in that area. It requires a company to define its critical operations, to identify companies that perform chosen operations exceptionally well, and then to collect measurements from those companies as well as from its own operations.
For example, when Baldrige-winner Xerox wanted to benchmark its product delivery services, it looked to L.L. Bean as a premier provider in that area.
As benchmarking gradually becomes a common practice in developing total quality cultures, employees' views of their organizational cultures have also become objects for comparison across firms.
Many companies have surveyed employee attitudes for some time through paper-and-pencil surveys. Many regularly compare their survey results to those of other companies. Data bases of employee opinion norms are available through consulting firms and consortia.
Most employee surveys use five-point response scales with "agree"/ "disagree" anchors (strongly agree, agree, neither agree nor disagree, disagree, or strongly disagree) or "satisfied"/dissatisfied" anchors (very satisfied, satisfied, neither satisfied nor dissatisfied, dissatisfied, or very dissatisfied). Other variations of the scales are possible, but companies that want to be able to compare their survey results should use comparable response scales.
Results are typically reported as a single "favorability" index; in other words, as a percentage of respondents who gave the item favorable ratings. To come up with the favorability rating, compilers collapse the "agree" and "strongly agree"--or "satisfied" and "very satisfied"-- responses into a single score.
A single index has the drawback of masking the strength of the opinions. Some survey designers wonder why they bother with five- or seven-point scales, when the resulting favorability score will report only one attitude--almost as if the survey had asked for a yes-or-no choice.
For example, say that three different companies include a particular item on their employee surveys. Respondents are asked whether they strongly agree, agree, are neutral, disagree, or strongly disagree with the statement. The results are as follows:
* At Company A, 40 percent of employees agree and 60 percent are neutral.
* At Company B, 40 percent of employees strongly agree and 60 percent strongly disagree.
* At Company C, each of the five possible answers draws a 20 percent response.
For all three companies, the responses would be reported as "40 percent favorable," though the firms' employees obviously have very different opinions on the topic. It is not unusual to find differences like these lurking inside "equal" favorability scores.
Organizations should be aware of the limitations of norms, the differences between norms and other benchmark measurements, and ways in which the use of norms may be inconsistent with total quality cultures. If management insists on using normative data, the department charged with implementing the request should be prepared to ensure that the norms it looks at are the best available.
Using norms as benchmarks
Certainly the first decision should be whether the use of employee opinion "norms" is consistent with the culture an organization is trying to develop or maintain.
The issue here is one of measurement and content. A norm is an indicator of normal or average performance. In looking at norms, an organization is comparing itself to an "average" company. The implication is that it will be satisfied if it can say that it is better than average. This in itself is inconsistent with benchmarking, which involves a company comparing itself with the best companies in a specific performance area. …