New research finds doing both well is particularly difficult when firms operate in cycles, an environment that is more common than we realize.
Article continues below
"Learning is usually about figuring out the new and forgetting the old," said Scott Rockart, an associate professor of the practice at Duke University's Fuqua School of Business.
"That's efficient in settings like technology. We know we're not going back to cathode ray televisions, for example. But a lot of what we do instead goes through predictable cycles, during which things become more or less important over time. In those cases, to move on and forget is a mistake."
Cycles in business come in a variety of forms and vary widely in length and depth. Some cycles, like those in fashion, are generational. Other cycles are much faster, driven by natural seasons, the weekly calendar, and the rise and fall of economic activity.
When cycles are brief, it's easy to remember how to thrive in different conditions.
"What people learn, and what people need to do well, differs at different points of a cycle," Rockart said. "Imagine you're a waiter, and during the week you are working with families, taking time over a meal, whereas on the weekends your customers are college kids looking to party.
"In order to be a good waiter during the week, you introduce yourself, you take time to talk, but on the weekends you are moving fast and not spending as much time with each customer. The waiter knows they need to act differently in those different settings. They're constantly being reminded because the cycle is so fast."
But at firms, Rockart said, these cycles can last years.
"So it's very easy for individuals to forget what they used to know, and even more importantly, to lose confidence in the importance of what they used to know," he said.
"If you're a manufacturer, when things are booming, you need to make sure you can get supplies, hire workers, and expand a factory.
"During a downturn you may have to do very different things: make sure your customers won't fail on credit; that your inventories are lean and your costs are controlled; and reduce your workforce."
Rockart's new research, "Learning in Cycles," is forthcoming in the journal Organization Science. He worked with Kristin Wilson of the University of North Carolina.
They studied community banks across the U.S. to see how well institutions learned over the course of real-estate cycles. The researchers looked at 40 years of government data from more than 3,000 banks across all 50 states.
While experience is usually associated with learning and improvement, they found that banks with more experience during an upturn in the real-estate market generally underperformed relative to their peers during subsequent downturns, and vice versa.
"When banks came back from being in another phase of the cycle," Rockart said, "they weren't as good as they were before. Most importantly for managers to consider, we found the culprit was not simply forgetting."
Loss of knowledge was not well predicted by how long a firm had been in another phase, but by how intensely firms had worked during that other phase.
Rather than simply forgetting, firms appeared to be actively overwriting prior knowledge, even when it would certainly become important again later.
"This might be efficient in the short term, but over the long term it means firms are far less effective than they should be," Rockart said. "The implication for managers is that they need to ensure the practices necessary to succeed in one phase of the cycle are protected during other phases."
Because cycles are inevitable, even if the timing is unpredictable, firms need to acknowledge when they operate in a cyclical environment and adapt accordingly, Rockart said.
"The thing about cycles is that they repeat," he said. "When we're in a downturn, we know that at some point an upturn will follow. Firms need to recognize that they learn certain things for a reason and will have to do them again in the future.
"You can be smarter about not doing certain things during an upturn, but you can't just go blindly through and assume that if something's not needed right now, that you can wipe it from your memory and your protocols." ■