Learning creates an advantage in a stable environment. By repeatedly performing tasks, businesses learn to execute them more efficiently—it’s the experience curve effect. But as our world changes more quickly than ever, efficiency increasingly gets trumped by adaptability. And experience, which traps us in a backward-looking perspective, becomes more curse than blessing. In this context, rather than merely adding to and refining existing knowledge, companies also need to become better at detecting and letting go of outdated, irrelevant, or less useful information and practices.

In other words, companies need to become better at forgetting. Often seen as a malfunction of the brain or a sign of decline, forgetting is actually a mechanism that helps us adapt to new environments. In this article, we explore how companies can achieve a balance between learning and forgetting.

The importance of evolvable mental models

Because the world is complex beyond what our brains can encompass, we create an inner world that works for us: a simplified representation of reality, or mental model.^[1]The Imagination Machine: How to Spark New Ideas and Create Your Company’s Future, Martin Reeves and Jack Fueller, Harvard Business Review Press, 2021. As cognitive scientist Anil Seth puts it: “We don’t just passively perceive the world, we actively generate it. The world we experience comes as much, if not more, from the inside-out as from the outside-in.” For example, when we walk into a café, we register that it’s a café—then our brain fills in most of what we expect to be there.

Our mental models guide our decisions. For example, Kodak observed a rising demand for digital photography, but because it evaluated this information through the mental model of an analog photography business, it responded by doubling down on its core business rather than embracing the opportunity to capture a new market.

Businesses are directed and sustained by the mental models of their people. Over time, mental models become manifest in processes and structures—further entrenching themselves. Stories of corporate downfalls are often stories of entrenched mental models and the real-life complexity they engender: leaders at Kodak (as well as Blockbuster, Nokia, and others) relied on memories of what made them successful in the past and were unable to recognize or lacked the organizational agility to adapt to shifting trends.

To stay adaptive in a turbulent world, companies must make their mental models evolvable. A salient example of this is Amazon’s “Day 1” philosophy: By treating every day at the company like the first day of a startup, the firm fosters organizational forgetting. Employees are encouraged to liberate themselves from legacy practices and to embrace an open mindset, which has helped Amazon pivot to new sources of growth and advantage again and again (from online bookstore, to everything-store, to marketplace, to cloud computing provider, and beyond).

The balance of learning and forgetting

In theory, because of natural employee turnover, businesses have an innate forgetting mechanism. However, firms have set up deliberate practices to retain information, through their processes and structures, knowledge management systems, and (written and unwritten) cultural norms. This was key to success in an era characterized by a more stable competitive environment, in which accumulating information and using it to refine mental models and processes enabled outperformance. In a digital world, wherein the costs of information storage are negligible and the ease of retrieval is greater than ever, it is tempting for companies to overlook the downsides of information retention: an entrenchment of mental models and practices, which becomes dangerous as their relevance and usefulness declines. In today’s turbulent times, deliberate forgetting mechanisms—which help firms get rid of outdated information and practices—need to be established as a counterbalance.

For inspiration on what a system of learning and forgetting may look like, we can turn to the foraging processes of animals: When obtaining food, animals use learned information about food locations, predator behavior, and seasonal variations. But because resources eventually deplete and environments change, relying on this information alone will decrease their odds of long-term survival and procreation. As a result, animals have evolved a desire to explore beyond the boundaries of their known environment to identify new resources they can harness in the future. Exploration provides signals regarding which pieces of existing knowledge are outdated or irrelevant and should be forgotten—thus enabling updates to mental models and actual behaviors.

How companies can implement forgetting

To replicate this balance of learning and forgetting, businesses must establish deliberate forgetting mechanisms that counteract the existing systems geared for information retention.

Structures and processes: Companies should put in place procedures to regularly review and prune structures and processes. This does not only serve to counteract the gradual buildup of complexity, but will also ensure that “ways of doing things” do not get so entrenched they can no longer be easily changed. For example, at Netflix, managers are incentivized to regularly eliminate distracting or complex rules. As former Netflix Chief Talent Officer Patty McCord says, “the longer the legacy, the longer the history, and the deeper the habits, the harder it is to change. […] Large corporations realize that they can’t innovate […] and they’re not nimble.” As a result, she has learned to “question everything and scrub the stuff that didn’t matter..

Companies can also experiment with setting expiry dates as they create new structures and processes—a practice already applied in the context of projects or committees.

Knowledge management systems: Companies need to continuously clean, prioritize, and update information in their databases. For example, in an interview with a Meta engineer, we learned that the company’s internal knowledge database, which stores documentation from HR processes to coding rules, assigns each document a score from 0 to 100 that reflects user feedback and freshness. This incentivizes frequent updates and revisions, indicates to the reader how useful a document is, and makes outdated or obsolescent documentation naturally harder to access (as it is sorted downwards based on the score).

Culture: Forgetting entails not only letting go of the theory (for example, a process or document prescribing a practice) but also of the habit—as humans tend to stick to past behaviors. Implementing forgetting therefore requires fostering an understanding that mental models influence decisions, and that making them malleable has advantages for adaptivity and, thus, for long-term competitive advantage.

To make mental models malleable, make them explicit, for instance, by spelling out the implicit assumptions and heuristics guiding decision-making. This will help people understand that they are just models that can be changed. AI can help firms make their mental models explicit: By being fed information from meetings, documents, or internal communications, language models can identify and formulating an organization’s unspoken beliefs and assumptions.

Awareness of your mental models can also provide inspiration. For example, Starbucks grew up when there were two dominant mental models in the U.S. coffee business: You either sell instant coffee, fast and conveniently, or you sell coffee beans to be ground at home. Starbucks CEO Howard Schultz overturned these assumptions and imagined a new type of coffee company that offered a space for socializing, but at which coffee—and the smells and noises of the coffee grinder—remained the draw.

Making mental models malleable also means fostering an openness to change, which can be accomplished by deliberately considering information that will challenge existing mental models: anomalies, the unfamiliar, or issues that you do not understand. For example, in 2000, Google—whose search results were then limited to simple pages of text with links to websites—observed an anomaly: A spike of searches for the green dress worn by Jennifer Lopez at the Grammy Awards. They realized that image search would be required to address this type of query and set out to develop this feature, which has since become a staple.

Microsoft’s success story of the past decade exemplifies this shift toward a mindset that embraces change. The company transformed its culture from one in which knowledge of the past was a valuable asset for gaining power, to one in which being curious and generating new ideas is encouraged and rewarded. Says CEO Satya Nadella: “If you take two people, and one of them is a learn-it-all, the other one a know-it-all, the learn-it-all will always trump the know-it-all in the long run, even if they start with less innate capability.” To foster its learn-it-all culture, Microsoft created dedicated time for growth (a quarterly learning day) and initiated “hackathons,” which led to the development of products like Microsoft Teams.

Calibrating forgetting

How can companies find the right balance of learning and forgetting? Turning to biology, we identify two crucial factors.

Turbulence of the environment: A more dynamic context means that staying up to date is more beneficial, which will require more frequently initiating new learning cycles. Research has shown that a higher rate of environmental change means animals track their environments more closely—through exploration—and give more weight to recent experiences at the cost of older, outdated information. For companies, the same logic applies, which is why successful players in quickly evolving digital markets often look back on a history of significant business model pivots. For example, both Flickr and Slack started as video game developers before becoming pioneers in photo sharing and office communication, respectively.

Emphasis on the short-term: In foraging, the payoffs of exploration lie further in the future than those of harvesting from known food sources. As such, in times when there is greater importance on the short term, behavior is adapted to de-emphasize exploration and forgetting. For example, birds are more likely to stick to food sources they know to be reliable when feeding their young, minimizing the time and energy required to find food, the uncertainty of food quality or availability, and the risk of exposure to predators. Accordingly, companies looking to foster long-term advantage should calibrate more toward forgetting, while those looking to increase immediate payoffs should focus on using and refining their current knowledge.

By creating a cycle of learning and forgetting, companies will be better able to adapt to changing environments, evading the mental entrapments of the past. But the real winners of tomorrow will go one step further, complementing the new-found power of forgetting with an improved learning capability. For example, firms can harness each execution of routine processes to learn more about what else their customers might want or what they might want next. As an example, consider Mars, which, after observing that customers who bought their chocolate bars also bought spicy snacks, created a “Spicy Snickers Bar,” which became an instant hit in China. This powerfully illustrates that firms can treat execution as more than merely an opportunity to identify marginal efficiency gains.