Moving the Goal Posts



Watts S. Humphrey

This library item is related to the following area(s) of work:

Process Improvement

This article was originally published in News at SEI on: June 1, 2000

In this column, I talk about the nature of process improvement and why it is such a dynamic and challenging field. The future will be much like the past in many respects, but it will also be very different. However, as we look ahead, there are some reasonably reliable guides that can help us to address the problems we will face.

Brownian Motion

Just about every time I visit an engineering organization, the people tell me, "We’re different." Of course they are. We are all different, but what is surprising is how often truly different organizations behave in the same way. However, it is also surprising how often seemingly similar organizations, when faced with nearly identical conditions, behave quite differently. People are both predictable and unpredictable. Much like Brownian motion in physics, there is no way to precisely predict individual behavior. However, on average, overall behavior is highly predictable. So, what does this mean for process improvement? Essentially, the following:

  • First, there is no one best way.
  • Second, every situation is different. Each solution must consider the people and their backgrounds, beliefs, and circumstances.
  • Third, the lessons of the past are the only practical guides we have for the future. While we cannot predict precisely what will work in any specific case, we can establish highly reliable general guidelines.
  • Fourth, the principles behind the quality movement are just as sound today as they were in the past. Those who argue that the new Internet age changes all the old truths will continue reliving the same history that many of us have painfully survived.

What these lessons tell us is that a single-minded approach to solving any human problem will almost certainly be wrong, if not for everybody, at least in many cases. There is no single best answer. People are extraordinarily creative, both in the ways that they solve problems and in how they create problems. Therefore, we must recognize that problems will change, and we must continually seek newer and better ways to address the problems that we face at each point in time.

When the Problems Change, the Solutions Must Also Change

The other day, I read the following newspaper headline: "The quality of U.S. automobiles lags behind Japan and Europe." As Yogi Berra once said, this is "déjà vu all over again." After 20-plus years, can quality still be a problem for Detroit? It almost certainly is, and the best way to tell is that the General Motors board of directors cut executive bonuses. That is a guaranteed way to get management’s attention.

GM, Ford, and Chrysler have been working on quality improvement for more than 20 years, but they still have about 150 defects per 100 new cars. However, unlike 20 years ago, these are not primarily manufacturing defects. Most are design problems. Detroit solved the quality problems of 20 years ago, and if the Japanese had not kept moving the goal posts, Detroit would be in fat city. But the world did change, and Detroit is still dead last in the quality sweepstakes.

The world changes, and it does not change all by itself. Everything we do changes it. In another lesson from physics, Heisenberg showed that you can know a particle’s location or its velocity but not both. When you measure one, you change the other. People are just like that. As soon as you fix the process, the problem changes. Does that mean that we should give up? Not at all; it just means that we cannot relax. We must keep thinking, and resist the temptation to blindly rely on the solutions and formulas of the past. Continue to follow the same principles, certainly, but don’t blindly follow the same path. Sooner or later it will lead to a dead end.

Finding the Goal Posts

While process problems are often unique, they all stem from human failings, and these are common to all of us. Because the same human failings have persisted through the ages, we cannot expect to eliminate them. The process improvement challenge is to devise ways to live with and compensate for normal human behavior. We must recognize, however, that soon after we compensate for a given set of failings, human nature will find creative countermeasures. So, in spite of all our efforts, the battle for improvement will continue indefinitely. Hopefully, however, technology will keep improving and each step will move the goal posts a little further down the field.

Human Failings

While software professionals are marvelously creative and highly energetic, we sometimes feel lazy or want to take a break. We are also a race of procrastinators, and when we can’t avoid or put off some difficult or unpleasant task, we try to replace it with an easier task or get someone else to do it. If we find that we still must do the job, we tend to do it as quickly and superficially as we can get away with. This means that for every complex and difficult task, the process improvement challenge is to devise ways to get people to consistently do their work in a highly professional way.

What makes this so challenging is that once we figure out some way to do this, it is only a matter of time before people devise a clever way around our fancy new process. Take estimating, for example. The Capability Maturity Model® (CMM) calls for engineers to be involved in and agree with the project estimate. However, soon after an organization puts a new planning procedure in place, some group will almost certainly find an estimating method that uses expert estimators or complex and arcane tools. Experts will then make the estimates and the engineers won’t be involved. Even though this destroys the intent of the planning process, unless processes are defined very carefully, people will adopt new practices that conform to the letter of the defined process but not to its intent.

What this Means for Process Improvement

What this means for you and me is that process improvement must not be directed at only the process. The principal objective must be to change human behavior. However, to change human behavior, we must consider and compensate for normal human failings.

For example, we now find that even in CMM Level 5 organizations, people have learned to compensate for their new processes. In some of the Level 5 organizations I have visited, the measurement and process analysis work is handled by the process and quality groups, and the engineers continue to work essentially as they did at Level 1. This totally misses the point of CMM Levels 4 and 5, which is to have engineers use data, not just gather and report it. This implies that even the goal posts defined by the CMM levels must be moved to keep pace with our rapidly changing technology.

In the last analysis, to improve engineering performance, organizations must change the behavior of the engineers and their managers. If you find that some change has stopped producing the desired results, find out why and then devise another improvement to solve the new problems.

The Implications for the Future

We have made great strides in the last 10 or more years, and we must continue to build on our successes. However, the goal posts are moving, and the problems we will face in the future will almost certainly be different from those of the past. Think of it this way: You could build a 10-foot boat in your garage, but a 1,000-foot ship would require entirely different tools, technologies, and processes. Similarly, in transportation, going from 3 to 300 miles per hour requires several changes in technology. In the software business, we think nothing of factors of 100. We use the same tools, methods, and processes for a program with 10,000 lines of code (LOC) as we do for a 1,000,000-LOC programming system.

Our ability to master the software-intensive technologies of the future will be largely guided by the ability of engineering teams to match their behavior to the more demanding tasks they will face. We cannot expect that our current tools, technologies, and processes will be adequate in the future. The challenges will keep increasing, and we must continually evolve our methods to keep pace. We must think of process improvement in multi-dimensional terms and include the educational system, as well as industry. An informed customer community will also be important, and we must consider all levels of the engineering organization: executives, managers, teams, and engineers. Much as in the automobile industry, we must retain the solutions of the past, but we must broaden our perspective to consider all the relevant aspects of the problem.

While it is always risky to predict the future, some trends are now pretty obvious:

  • Systems will get larger, more complex, and more integrated.
  • Engineering teams must also become more highly integrated.
  • Compatibility, reliability, usability, privacy, and security will be increasingly important.
  • While schedules must be as short as possible, they must be absolutely reliable.
  • The quality of every engineer’s personal work will be even more important than it has been in the past.


In writing papers and columns, I make a practice of asking associates to review early drafts. For this column, I particularly appreciate the helpful comments and suggestions of Noopur Davis, Jim McHale, Don Mcandrews, Julia Mullaney, and Marsha Pomeroy-Huff.

In closing, an invitation to readers

In these columns, I discuss software issues and the impact of quality and process on engineers and their organizations. However, I am most interested in addressing issues that you feel are important. So, please drop me a note with your comments, questions, or suggestions. I will read your notes and consider them in planning future columns.

Thanks for your attention and please stay tuned in.

Watts S. Humphrey

About the Author

Watts S. Humphrey founded the Software Process Program at the SEI. He is a fellow of the institute and is a research scientist on its staff. From 1959 to 1986, he was associated with IBM Corporation, where he was director of programming quality and process. His publications include many technical papers and six books. His most recent books are Managing the Software Process (1989), A Discipline for Software Engineering (1995), Managing Technical People (1996), and Introduction to the Personal Software Process(1997). He holds five U.S. patents. He is a member of the Association for Computing Machinery, a fellow of the Institute for Electrical and Electronics Engineers, and a past member of the Malcolm Baldrige National Quality Award Board of Examiners. He holds a BS in physics from the University of Chicago, an MS in physics from the Illinois Institute of Technology, and an MBA from the University of Chicago.

The views expressed in this article are the author's only and do not represent directly or imply any official position or view of the Software Engineering Institute or Carnegie Mellon University. This article is intended to stimulate further discussion about this topic.

Find Us Here

Find us on Youtube  Find us on LinkedIn  Find us on twitter  Find us on Facebook

Share This Page

Share on Facebook  Send to your Twitter page  Save to  Save to LinkedIn  Digg this  Stumble this page.  Add to Technorati favorites  Save this page on your Google Home Page 

For more information

Contact Us


Help us improve

Visitor feedback helps us continually improve our site.

Please tell us what you
think with this short
(< 5 minute) survey.