Measurement

Most performance improvement programs measure only results. This can tell you whether the program worked relative to measured objectives but rarely why it produced these results. But, to apply this knowledge again or elsewhere, you have to know what factors and performance drivers were most important to achieving the observed results and how to adjust the program for any differences in a new situation. This is the heart of our measurement methodology — measuring and analyzing every important program aspect so that real knowledge can be generated.

We use the term "measurement" to refer to any process that obtains data from a business situation and organizes it for purposes of analysis.

Only very simple programs and initiatives can be characterized adequately by a few variables. Most require details about the business situation (people involved, stakes, history, initial obstacles, resource constraints), the approach itself (how it is supposed to work, steps in implementing it, any preparation needed, resources required, time frame, essential expertise), implementation details (changes made, new information obtained, obstacles encountered), and finally the outcome relative to goals.

Without enough detail, you often cannot determine which factors and performance drivers are most important in producing the results. Without this knowledge, you can't repeat the approach with any confidence of success if anything in the new situation is different. Detail also allows us to determine what went wrong with a program that did not work, or didn't work well enough. This knowledge may suggest changes that should give the approach a better chance of succeeding.

Measuring nearly everything can be a time-consuming, costly process if done manually and without some time-saving techniques. We have a wide range of computer-based measurement tools to make much of this process quick and easy. We have done enough assessments to know what to look for and to be able to fill in blanks from similar situations. From prior assessments, we know what to ignore and where to focus our main effort.

Measurement produces two kinds of data: The first is largely descriptive (text) that is vital for a complete understanding. The second is quantitative, typically subjective, input that makes it possible to do statistical analysis or model situations and programs (see below). We want to know why a program works (or doesn't) in enough detail to be able to describe how to replicate a success in a new situation or to improve the odds of success if the approach is modified and tried again.

Where there is sufficient data, we can employ a variety of statistical techniques to identify the most important factors and drivers in each situation. In other cases, we may be able to adapt and apply one of our behavioral (simulation) models. In new situations, we fall back on our business experience and expertise to figure out what is going on. In all cases, we want to know which aspects of the situation and program were critical to the results achieved.

Not always is it possible to discover from an initial program implementation how it really works or how to replicate its outcome in a new situation. In some cases, even the program's outcome may be hard to assess on the first pass. This is why we prefer to work with collaborative projects where approaches can be tried in various flavors and situations. Each new implementation provides both a better understanding of what is going on and better assessment tools for future implementations.

PLEASE TELL US — Was this material of any value to you?