No process, no improvement

Business theories come in and out of fashion, and like fashions they repeat themselves. There's nothing wrong with that: each generation of managers generally needs to learn the same lessons.

Process improvement is an idea that's been around at least as long as the Romans ("Adopt, Adapt, Improve"). It's a powerful and generally-applicable idea, and it's been applied in business in many many different guises over the past seventy-odd years. Deming popularised one particular approach based on statistical analysis and called it Total Quality Management. A similar set of ideas and tools has resurfaced as Six Sigma. Key ideas forming part of these appear in the latest version of the international quality assurance standard, ISO9001:2000.

The basic idea is the same: put aside time and effort for thinking about how to improve processes, and implement your ideas.

Things were simpler between the wars, when the basic ideas of statistical process control were being developed in the electronics manufacturing industry. There, the process could (literally) be taken apart and looked at - it was a physical manufacturing process. The term 'Six Sigma' comes from the allowable percentage of failures in silicon chip manufacture (six times the standard deviation), as promoted by the people at Intel who coined the phrase.

But most process improvement these days is aimed not at manufacturing processes, but at business processes. And although many of the same improvement techniques work just as well (and you can find details of many of them here on our web site), there's a fundamental difference between business processes and manufacturing processes: you can't see business processes.

You can see people carrying out various steps of the process, but you normally can't see the whole thing laid out like a production line. People enter information into computers, or onto forms, and they phone each other or have meetings, and business goes on. There's no machinery, no warehouse, and no final inspection area.

Many organisations make the mistake of trying to improve business processes using just manufacturing techniques. It's a little like trying to improve a manufacturing process without turning on the factory lighting: unless you actually make the process visible, it's very difficult to improve. Worse - you have no way of making the change permanent.

Most business processes are undocumented. They are 'understood' by the people who carry them out, and they are sometimes mandated by declarative documents (memos, rules, laws) which do not actually specify a process.

When we come to document processes for our clients, we almost always find that there is more than one process: each person who carries out the process (or part of the process) does it in a different way. Put all of those together, and you've got a process that might be pictured by a fuzzy set of lines on a flowchart.

Improving an undocumented 'fuzzy' process means making assumptions about what the process is. The improvements you make will change the process, for some people, for a while. The improvements are likely to be based on ideas of the process which are to some extent erroneous.

And when the process improvement has been made, there's no way to stop the process from sliding back to what it was before.

The alternative, of course, is to document the process before you try to improve it. Turn on the lights in your factory before you decide how to change the machinery.

This article may be reproduced only with the permission of HCi (email HCi ). Copyright HCi, 2001-2.