No Silver Bullet
Good managers fail when they attempt to use silver-bullet solutions to complex problems. Jerry Weinberg writes in Quality Software Management, Volume 3: Congruent Action that
- There is no silver bullet, but sometimes there is a Lone Ranger.
- —G. Weinberg , p. 1.
Weinberg explains exactly what he means by this, but the reason I like the sentence so much is that it highlights the dichotomy between an emphasis on things (bullets) and an emphasis on people (even a fictional character like the Lone Ranger). Let me take my interpretation one step further and amend Weinberg’s line:
- There is no silver bullet, but there are Lone Rangers who have arsenals of bullets for different situations.
For most managers, the hard part is understanding the different types of bullets and the situations in which each is most likely to succeed.
Ultimately, software projects don’t succeed because of techniques; they succeed because of ideas, individual responsibility, basic principles shared by team members, and collaborative interaction. There is no silver bullet—a truth that applies to mountain climbing as aptly as it applies to software development. A mountain climber knows that there are many different mountains, and that each has more than one route to the top. There are many climbers, each with unique skills and experience. Although better tools have greatly advanced the sport of climbing over the last fifty years, it is still the effort of the individual in combination with the climbing team that makes the difference between their reaching the summit or having to retreat.
Why, then, do so many development organizations say there are no silver bullets for software, and yet at the same time act as if there are? This idea of a silver bullet and the continuing denial that one exists is puzzling. What does “no silver bullet” really mean? The easy answer—at a superficial level—usually is that a particular tool, method, or concept isn’t going to solve all our problems. But this interpretation is too simplistic; the essential next question should be asked: “What about the no-silver-bullet concept has caused it to be so widely discussed in software engineering literature?”
One reason the silver-bullet concept remains a topic in our literature and a tickler in our psyches is that, in an era of constant change, uncertainty, growing complexity, increasing speed, and overwhelming competitive pressure, we crave something simple upon which to build a battlement against the forces of chaos around us. We know a silver bullet doesn’t exist, but we want one anyway. Whether the silver bullet is in the guise of Total Quality Management (TQM) or object technology, its existence injects the illusion of stability into our chaotic, anxiety-ridden world.
Tremendous amounts of time and money have been poured into this quest for stability and the elimination of unpredictability in businesses, in general, and in software development, in particular. Business Process Reengineering was one such silver-bullet effort for which the results never matched the original expectations. Studies reported in the mid-nineties have shown that few BPR projects were successful (for examples, see Tapscott96). Despite everyone’s belief in them, why have so many silver bullets been such a dismal failure?
One answer comes from Ralph Stacey, professor of management and director of the Complexity and Management Center at the University of Hertfordshire in Great Britain, who provides a new perspective on the issue of silver-bullet failure. In his 1996 book, Complexity and Creativity in Organization, Stacey identifies an essential mental shift needed to understand the failure. The paradox is that the silver-bullet solutions to business problems become the source of the very failures that they were designed to prevent.
Stacey contends that current silver bullets—such as TQM, BPR, and the CMM—create a vicious cycle of problem definition and solution. I call this cycle the optimization paradox, and it follows the stages outlined below:
- Some problem (for example, the need to reduce costs, increase quality, or reduce cycle time for some business function) is articulated by business managers.
- The solution (for example, BPR) is based on a deterministic organizational strategy in which the company analyzes the situation and then designs a solution based on an assumption that the business environment is predictable and stable.
- The business environment of course turns out to be ambiguous and unpredictable; therefore, the “reengineered processes” are faced with unanticipated variations that adversely impact the success measures. The dichotomy between the real and ideal worlds increases management and staff anxiety.
- To correct this new problem, management selects models of business strategy that are prediction- and control-oriented, the assumption being that better prediction and control methods are needed; hence the reliance on well-articulated, engineering solutions such as BPR as a vehicle for imposed order.
- The business may respond to the silver bullet favorably for a short period, but because rigid processes exclude data outside “acceptable” limits, critical feedback is denied. Since the process “can’t be faulty,” failure to meet predetermined success measures is blamed on poor execution and therefore “motivation” and “control” efforts are redoubled.
- The failure to achieve desired results is identified as the next “problem” and a new silver bullet is proposed.
The next silver bullet starts the cycle over again. The fundamental reality is that the business environment is ambiguous and unpredictable, but this reality is never seriously acknowledged. The statement, “I don’t know how much it will cost or how long it will take or exactly what we can produce” is not acceptable from someone who wants to be a “real” manager. Managers may pay lip service to constant change, but they are not willing to adjust their fundamental views of management to accommodate the real impact of that change. On one hand, they acknowledge the need for change, but on the other, they still think they can beat the odds and succeed with their “plans” in the face of change—denial in action.
In this problem-solution-problem cycle, millions of dollars are spent for ever-decreasing results. Organizations in this cycle experience a lot of change, but little learning. They are trying to optimize the non-optimizable. Stacey contends that breaking the silver-bullet syndrome depends upon breaking the linear, imposed-order mental model of business and substituting a belief in emergent order.
Most managers use silver bullets in the hope that they will provide prescriptive answers to complex situations. Most soon discover, however, that these prescriptive, optimizing answers get bogged down in metrics, procedures, forms, and slogans—and use of the chosen solution often lasts only until the next fad comes along. Silver bullets can be useful, but they are insufficient. They deal best with the stable parts of the environment, not with the rapidly growing complex portion. In complex environments, use of TQM or BPR or even CMM practices may seem to be the answer, but they are not sufficient for success.