The True Cost of Guessing in Problem Solving

By Scott Whitbread

Some years ago I worked on a big problem in a food manufacturing plant. Fifty heat exchanger units, used to heat and sterilize a liquid food product, were prone to leaking through a mechanical seal that allowed them to rotate. These leaks had been appearing for as long as anyone could remember, costing the company millions and posing a food safety risk. When a leak got bad enough, the unit was taken out of service, and its seal rebuilt, temporarily correcting the issue but causing substantial cost and disruption in the process.

As tends to be the case with such long standing, unsolved problems, there were a number of competing theories about how it was occurring and therefore how to solve it permanently. The most complex solution was proposed by the director of engineering. He had concluded that the units, as designed, could simply never perform without leaking. He suggested replacing the entire seal mechanism on all fifty units with a new technology. His guess looked good because the units had never performed without leaking (at least within the organization’s living memory). The downside was the multi-million dollar price tag of the new technology. Plus, it would take over a year to implement and may cause other new problems since the technology was, as yet, untested in this type of application.

The simplest solution was proposed by a maintenance technician. He guessed that the pressurized food product was steadily overcoming the force that pressed the two seal surfaces together: a coil spring. He suggested using a spring with additional coils -- a stronger spring -- to increase the force being applied and counteract the pressure of the food product trying to escape. His guess was attractive because springs are cheap; just a couple of dollars each. However, few people, including the technician himself, believed it could make a significant difference. For starters, there wasn’t room inside the spring assembly cavity for any extra coils. And how could such a trivial change make any difference on such a difficult problem; surely it was immune to simple solutions like that.

There were other theories too, related to the equipment and parts suppliers, and the operations and maintenance practices. The various champions of these different guesses were locked in an epic battle for the direction the organization would take in addressing the issue. Other people took comfort in the fact that no one expected a rapid improvement (it was a very hard problem that had defied resolution for years) and that there seemed to be some promising ideas to pursue.

This is where many businesses get stuck: there’s plenty of ideas, but insufficient confidence that any of them are correct. How do you decide which of them to pursue first (or at all)? Each wrong guess you invest in wastes time and money, diminishes the political standing of the idea’s champion, and further solidifies the problem’s reputation of being “unsolvable”. While guessing at solutions feels irresistibly productive, it usually does more to stall progress than to promote it.

How can you escape this dilemma? You need to approach problems with a new objective: investigate the problem to understand it root cause, and, in the meantime, forget about the solution. Once the root cause is well understood, the solution -- be it a new technology or a new spring -- will become evident; no guessing required.

For this team, such a process began by closely investigating the seal mechanism and scientifically determining what variables affect it. They found that much of the time when the seals leaked, the seal components were still in good order, and were able to rule these out as potential causes. Next, they studied the pattern of failure and found that leaks occurred more often at a higher operating pressure. This suggested the problem may lie in the competing pressures between the liquid product trying to escape and the seal trying to contain it.

They framed the problem in terms of these competing forces: the liquid’s pressure sought to part the seal interface, while the spring sought to keep the interface closed. Three variables control the spring’s force: the number of coils, the width of the metal strip used to make the coils, and the stiffness of the metal. While there was no room to add extra coils to the spring, and a stiffer material would be more expensive, it was possible to make the metal strip wider in order to apply greater force.

New springs were fabricated from a wider metal strip and within a few months all the old springs had been replaced. The leaks were virtually eliminated and the problem never returned. By seeking to understand the problem’s root cause -- an insufficient force against the seal interface -- the team found a simple solution: increase the width of the spring’s material.

So it seemed the maintenance technician was half right. Though, in the end, who was “right” is a moot point because simply guessing at solutions had produced a stalemate. Though the technician’s guess had practical merit in this particular case, without a deeper understanding of the problem and its root cause, the team couldn’t pick it out from other, poorer guesses. As such it was left behind, including by the technician himself. The business sat on a correct guess for years but testing everyone’s guesses -- including replacing the system entirely -- would have been far too expensive. Such is the true cost of guessing in problem solving.

What simple solutions is your organization endlessly contemplating for lack of confidence that they will work? What understanding of the underlying problem’s root cause is required to unleash (or abandon) them?


Like this article? You might like:

Ramping up a New Facility in Half the Time

Why Asking Stupid Questions is the Smartest Thing to do

Where Simple, Elegant Solutions Go to Die