Just as the flapping wings of a butterfly in West Africa can ultimately trigger a hurricane that ravages the U.S., the failure of an inexpensive part in just one machine can batter the financial performance of the company that owns it.

This “butterfly effect” — or a small change in initial conditions that leads to huge changes in results — is much noted in weather forecasting, but often overlooked in manufacturing. On the factory floor, however, seemingly routine or apparently insignificant machine health malfunctions can ricochet throughout an entire company and result in serious bottom line damage.

Let me give you an example based on a real-life case. In anticipation of a major tropical storm, South Floridians stock up on non-perishable food and bottled water. I was in charge of a plant in Florida that purified and bottled water, which was sold to local supermarkets. As we ramped up production to serve our customers in the face of an approaching storm, our labeling machine broke down due to the failure of a $100 bearing.

Diagnosing what, if any, additional damage the failed bearing caused to the machine would take time. Very often, motors and complex gear boxes can be damaged when a bearing fails and rotating parts freeze or break apart. Also, since the manufacturer of the labeler was German, it was going to take at least a few days until a replacement part could be flown in—plus additional time to install it, if the machine was not damaged in other ways.

Meanwhile, since we couldn’t ship label-less bottles, production had to be shut down. In addition to sending home factory workers, which damaged morale, our logistics team had to scramble and arrange for truckers to bring in supply from other bottling plants throughout the country. How do you quickly find hundreds of long-haul truckers, who most likely would exit Florida empty and perhaps get stuck in a storm? The answer involves paying them a lot of money—about three times normal shipping costs. Additionally, since most other plants near the hurricane path were already tapped out the product had to come from much farther away. How about as far away as New York and California? Also consider all of the double handling of the product which may have travelled 1,000 miles only to be offloaded into a warehouse floor to be reloaded to other trucks bound for the stores.

While this was going on, the company’s sales team and executive officers were trying to placate big supermarket customers who had contracted with us for deliveries of water, not excuses. Our communications team also was preparing to counter any negative publicity.

Fortunately, because of our size and resources, we were able to deliver the needed water to our customers pretty much on time. Sure there may have been an out-of-stock here and there but Floridians are used to spot outages in the face of an upcoming tropical storm. But had our company been smaller or dependent on that one line of business, what turned out to be a costly mess could have been an existential crisis. All because we could not predict the failure of that $100 bearing. For owners of any company, therefore, there is a direct link between reliable machine health and desired corporate performance.

Today, advanced predictive maintenance tools can make sure that link is unbroken. In fact companies like Augury, have machine health solutions that go beyond simply detecting malfunctions by telling you exactly how to correct any malfunction that’s been detected. For the many industrial companies who have more than enough big-picture issues to contend with, assurance that little factory-floor issues stay that way sounds like a good deal.

Ed Ballina, formerly Vice President of Manufacturing and Warehousing at PepsiCo, serves as an advisor to Augury, Inc.