Don’t Blame the Engineers

In the aftermath of the Columbia space shuttle tragedy in February 2003 and the Great Blackout of August 2003 in the northeastern United States and Canada, concerns have surfaced about whether nontechnical managers are capable of effectively managing the complex engineered systems for which they are increasingly responsible. As one experienced observer recently commented: “The moral I carry from both the Columbia disaster and the blackout is the dismal level of technical expertise of this country’s managers. … The technical side gets derided as detail.” That challenge, of course, is hardly limited to quasipublic undertakings, such as space flight and electric power; it is endemic in the private sector as well, especially in the United States, where training in management has always eclipsed engineering expertise in determining executive progress.

As a management generalist with technical experience in diesel engines, pharmaceuticals, chemical engineering, information technology and electronic payments, I believe generalist managers are more than capable of successfully managing complex technical systems. It is management’s responsibility to know enough and to challenge well enough to understand the true risks that their technical environments embody. Meeting the challenge of effectiveness demands a combination of skills, some of which appear to have been lacking in both the Columbia disaster and the recent blackout.

Preparing To Manage

Anyone who expects to be accountable for the performance of a technology-driven system has to understand the basics of the technology. Anything less is irresponsible. A variety of approaches exists to gaining appropriate competence, from intensive off-site training to part-time exposure and mentoring. Which is most appropriate depends on both the technology and the scope of the job. For example, when I was learning diesel engine skills needed to run an overseas operation, I tore down engines, spent evenings with company experts in heavy equipment and passed hours digesting the details of electrical generators. As a newly appointed chief information officer of a Fortune 300 company, I underwent an intensive one-on-one “boot camp” conducted by a leading IT consulting firm. Whatever the means, the ends are specific and unique to the managerial role: to grasp enough of the technology to understand how things work, what they cost, their inherent risks and the central factors driving value and governing choice. A further objective is to learn the language of the technology so the manager can “talk the talk.” Only when this new language and its nuances have been mastered will the generalist manager earn the respect needed to effectively lead technologists. Defective preparation at the beginning sets the stage for eventually deficient performance when it matters. Either poor decisions will be made, or the technically skilled organization will learn that it must work around an ill-informed manager.

Setting the Right Tone

An acute sense of the organization’s priorities is essential to any manager’s effectiveness. This is especially important to general managers leading technical functions, because of the often subtle pressure to pursue technology for technology’s sake. Effective managers set the right tone by focusing relentlessly on a few key priorities. In most chemical companies, for example, plant safety is a priority, and relentless attention to it is the prime reason for the low rate of lost work, accidents and fatalities in the U.S. chemical industry. Safety is an essential part of the tone set from the very top of the best-managed chemical companies, and it is made a conscious and consistent part of personal accountability, including financial incentives and career consequences.

Nothing undercuts tone setting more dramatically than a manager who does not “walk the walk” or who does not take action when others miss performance against important priorities. Effective managers motivate and reward positive behaviors, and leverage the lessons learned from mistakes and shortfalls. There is an appropriate place for forgiveness — depending on what’s at stake and the specific circumstances — but there is also a time for “public hangings.” Knowing when to use both is essential.

Measuring Performance

The selection of metrics is certainly a key signal to the organization of what’s important, but the generalist manager’s attention to those metrics is critical to generating personal respect, as well as appropriate focus, attention and execution.

But meaningful metrics lose their value if it appears no one’s looking. Subpar performance that is tolerated without consequence sends a message that no one is looking or that management simply doesn’t care. It doesn’t take long for the overriding result to become, “So why bother?” Accordingly, it isn’t enough to measure; managers of technical functions need to demonstrate that they are watching, that action will be taken when performance falls short and that those who deliver against their performance targets will also be rewarded.

Generalist managers have to work harder at uncovering defects in measurement systems. One potential source of problems is smart employees who know how to “game the system.” In one manufacturing measurement system installed for tracking delivery performance, the chosen metric was number of days from order creation to shipment. Performance metrics were indicating outstanding performance at the same time as customer complaints about late deliveries were on the rise. What management didn’t realize was that orders delayed too long were being canceled by plant management, dropped from the system and reentered as new, more recent orders. This kind of gaming happens more often than managers like to admit, because it makes them look foolish. Smart managers — engineers and generalists — will take the time to understand how the metrics really work: where the data come from and the detailed mechanics of delivery. Not only will this level of diligence produce more meaningful numbers, but the hands-on involvement entailed will underscore the seriousness and savvy the generalist manager brings to the job.

Getting at the Truth

In complex technical systems, the data themselves are often more difficult to understand and interpret, and the implications further removed from the generalist’s realm of experience. Consequently, getting at the truth can be a continuing chore. Both the Columbia accident and the blackout share elements of this challenge.

Managers freshly encountering complex technical environments need to quickly establish pathways to collecting unfiltered and accurate information. One way this may be accomplished is by holding large and frequent meetings so that a “vertical slice” of the organization has an opportunity to communicate up the chain. The key risk here is that the culture and group dynamics of such a large public forum may inhibit candid communication. The failure of this approach is documented in the Columbia accident report. An alternate approach is to create an official channel through which “bad” or “unpopular” news can be channeled so that the manager has a direct link to an unimpeded communication flow. A key concern here is the volume of messages and the difficulty of distinguishing between what’s really important and the general “noise.”

By far the best approach is to establish a “back channel”: direct personal contact with one or two technical staff who can be trusted to know enough to have real insight and who can provide unvarnished truth, but also have the mature judgment needed to keep things in perspective. As a CIO responsible for the Y2K remediation of thousands of numerical control devices across a large manufacturing enterprise, I instituted a deliberately open tracking system that forced different company units to compete against each other on their progress toward the year-end deadline. While this created the needed impetus for progress, I was aware that it could also motivate competitive managers to overstate their progress. To verify progress at the largest of our plants, I leveraged personal contacts in the engineering department to get a second read on the reported numbers: What was the real rate of progress? How was morale? Were consulting resources adequately deployed, despite the added cost? What other activities were suffering from lack of resources? Would resource constraints lead to a cutback that could put the company behind schedule? This information provided not only verification that added validity to the reported numbers, but also added to my confidence in reporting progress to the CEO.

Challenging Effectively

In the final analysis, the manager of a technical organization — engineer or generalist — must learn to rely on an effective challenge capability to uncover critical faults and concerns. Fortunately, this skill, when effectively practiced, is often more than adequate to the task in both technical and nontechnical environments.

What is required is skilled questioning designed to uncover gaps in logic or data, to ascertain levels of uncertainty, to test critical reasoning and to arrive at as complete a picture of the situation at hand as possible. Three conditions signal the need to instigate such a challenge: when something simply doesn’t make sense or is in conflict with known facts; when there is uncertainty or waffling on the part of the source or presenter; and when it is clear that one is being told what one wants to hear. In any of these situations, the responsible manager must challenge what she’s hearing and pick apart the words and logic until the flaw is uncovered or until things eventually make sense.

Three generic questions are vital to productively challenging any organization to realistically appraise the viability of its decisions, assumptions and facts: How do you know? Why do you believe that? How certain are you?

Once those questions have been fully and satisfactorily answered, the burden of decision shifts squarely to the manager to assess whether the odds are acceptable, given what’s at stake, and how much additional information — in the time available — needs to be gathered to refine the probabilities and reach a higher-quality decision. Note that a “higher-quality” decision is not the same as a “better” decision. A high-quality decision is one resulting from a more informed and more rigorous process, but one that may ultimately result in the same final decision. Conducting an effective challenge process can uncover the soft underbelly in facts, logic and bias, and help steer toward better-quality decisions.

In the final analysis, generalist managers running technical organizations will want to maintain the creative tension between the demand for fact-based, well-reasoned decisions on the one hand and the obligation of “loyal dissent” on the other. While there is certainly room for “cigar and brandy” decisions based on accumulated wisdom, deep reflection and sound judgment, an insistence on facts rigorously applied needs to underlie most of the commitments a technical organization makes. But cultivating the obligation to “push back” provides managers with an effective counter, which encourages the surfacing of concerns and issues even if the facts cannot yet be marshaled in support. Here, managers actively encourage doubt to enter the picture and, with it, the last protection against the hubris that technology can breed, which is so often exposed as unnecessary risk or wholesale folly. Indeed, after thoughtful decisions have been put into practice, the best way to keep things from going awry is to make it acceptable for anyone to raise concerns and challenge even the most fundamental beliefs and commitments. The generalist manager unskilled in all the intricacies of technology can, in this way, recruit the entire organization to check its own calculus. A healthy organization will be able to both give and take critical challenges yet remain intact. And such an organization will handle both routine activities and serious crises all the more effectively when the chips are down, regardless of who’s in charge.