The Heisenberg Uncertainty Principle of Management

In the science of quantum physics, Werner Heisenberg postulated that the more precisely we can measure the location of a particle, the less precisely we can measure its momentum--and vice versa. This principle, first formulated in 1927 is known as "Heisenberg's Uncertainty Principle."

Within management sciences, it seems to have a variant: "The more precisely we try to report on what we've done, the less effective we are at productively satisfying our mandate." Managers love reports. They want to know everything about anything as if reports, pie charts, and trend lines will make them feel more in control of a large department where they can't possibly expect to follow every detail as it transpires--but wish they could.

The Heisenberg Uncertainty Management Principle (HUMP) is self-evident when we thing about it. For one, unless the necessary reports can be automatically generated, collated, and dropped into the manager's inbox without any additional human intervention, the effort to do these things takes someone away from doing real work.

Second, people are going to alter their behaviour to align with reporting expectations. If you're measuring productivity by number of sick days taken, people will start coming to work sick, infecting their co-workers at work, and then leaving early. Measuring average time that a support ticket is open and people will start closing tickets prematurely, or transferring them to another department before they count against the technician's stats.

And lastly, managers with the new power of these reports are likely to ask some underling to explain any exception, variance, or unexpected trend that comes up in a report, thus generating a whole string of analyses and other activity that further undermines the productivity of the department.

Yet measuring and monitoring are important activities in any organization. It is a core management function. You can't change what you can't measure. Measurement, progress, process improvement, efficiency, productivity are all very tightly bound concepts.

How can we measure and monitor performance and results without undercutting our productivity? There are at least two ways:

  1. Only insist on a few reports that make sense. These reports should contain both leading and trailing indicators.
    • A leading indicator is a statistic that predicts a future trend or pattern. For example, an increase in employment is likely a leading indicator that there will be more helpdesk calls in the future (since employees make helpdesk calls and more of them will increase the number of calls), or an increase in budget will likely mean more project based work in the future.
    • A trailing indicator measures something after it happened. How many tickets did we get last month? How many project requests are in the queue?
  2. Generate reports in an automated fashion so that they naturally flow out of the work people execute and the procedures they follow. For example, the service management system should be able to automatically calculate the number of tickets fulfilled on time vs. late. We should also be able to get reports showing the average time and standard deviation between technicians in performing common requests.

The goal of both of these methods is to unobtrusively generate the necessary reports without the technician being consciously aware that their performance is being reported on. We want the technician to perform for the user, not for the reports. The reports should give us useful trending information and highlight for us where we need to modify or correct behaviour. Even then, and unless there is a significant problem in behaviour uncovered, the correction should be at a team or organizational level. Peers can do a better job of adjusting unacceptable performance on the part of one member than an overseer from management. Quality management systems such as TQM, Six Sigma, CMM, ITIL, merit-based pay, and others can help improve productivity and performance with monitoring supporting these systems.

By trying to make the gathering of reportable data part of the "background noise" of the organization rather than a task in itself, we can more accurately report on the actual work being done without having people trying to "play for the manager".