In 2018 and 2019, two Boeing 737 Max airliners crashed, killing 356 people. The U.S. Federal Aviation Administration grounded all 737 Maxes until the company could fix a design flaw believed to have led to the crashes. Yet the company’s problems continued. Deliveries of its 777X airliner were postponed in 2020; it still hasn’t shipped. And in late 2024, an emergency exit door blew off another Boeing 737 Max during an Alaska Airlines flight.

The company makes the perfect poster boy for this difficult lesson: Failure is bad, but not learning from it is worse. Beyond the unthinkable human toll, this cost Boeing billions in fines and settlements, plus years of intense regulatory and legal scrutiny, in addition to lost revenue and a damaged reputation. Boeing is hardly alone, however. Every corporation and person experiences failure—and a lot of it. As Mark Coopersmith points out in his book, The Other “F” Word: How Smart Leaders, Teams, and Entrepreneurs Put Failure to Work, most new businesses launched in the United States fail, as many as 95% of new products introduced each year fail, and some 68% of IT projects fail to meet their goals.

Yet many companies struggle to learn and apply the useful lessons from their failures. Despite corporate lip service to the idea (Fail fast! Fail early; fail often!; If you’re not failing, you’re not innovating!), businesses tend to tiptoe past the graveyard of unsuccessful projects, rushing on to the next thing.

“Businesses do a really bad job of learning from failure,” says Coopersmith, faculty director and senior fellow, entrepreneurship and innovation, at the University of California, Berkeley and coauthor of The Other “F” Word. On a human level, who wants to admit failure, much less dwell on it? From a business point of view, it costs time, money, and—if it’s public—reputation. It can carry dire consequences: Someone gets fired, the stock price plummets, and employees run for the exits.

But Coopersmith recommends taking a long-term view. “You’ve already paid for this failure, for this information,” he says, “so step back and see what you can take away from that to put to work. What you uncover could be a unique insight that creates the foundation for something different and better.”

Viewed this way, the by-product of failure could be a competitive advantage. And there are hints that applying increasingly powerful AI capabilities could eventually allow companies to learn from it more quickly and easily. Before that can happen, however, organizations need strong cultures that flip the typical script on failure: Rather than paying it lip service and then pointing fingers and quickly burying unsuccessful projects in unmarked graves, this culture establishes a healthy working relationship with it. That means accepting some (but not all) failures, adopting an attitude that they can provide the opportunity to learn and grow, and establishing systems and processes that support such acceptance and learning.

Building the right foundation

Step one is establishing a strong foundation, starting with the right culture. There are several elements to an organization's culture, including its history, leadership, and market. But attitudes about failure also depend on and can vary by industry or even geographic location. In Silicon Valley, failing fast is gospel. In healthcare or nuclear energy, it can be blasphemy. Japanese culture stigmatizes it and traditionally emphasizes avoiding mistakes.

And corporate culture can change over time. Boeing had a culture that encouraged employees to point out problems and discuss their concerns. “It used to be when you raised your hand and said, ‘We have a problem here,’ they’d say, ‘You’re right; we need to fix it,’” says a Boeing quality control manager in the documentary Downfall: The Case Against Boeing. But a merger in the 1990s brought new leadership and a culture that focused on the bottom line. “Boeing quit listening to its employees,” says the manager. “Every time I’d raise my hand, they’d attack the messenger and ignore the message.”

Even if the corporate line is to accept failure, it must be demonstrated for all employees to see. Evan Spiegel, cofounder and CEO of Snap, gives new employees an assignment that they’re sure to fail at. This eases any fear of failure from the very beginning and frees employees to think creatively, he says in a recent Diary of a CEO podcast.

Executives should walk the talk

The entire leadership team should also model the behavior. “That’s the way to build the culture. Leaders shouldn’t be afraid to share information about their failures. They can admit, ‘I don’t know what I was thinking here,’ but they can also point out lessons,” says Coopersmith. For example, maybe they had the right idea but missed an important element in its implementation. Now that the company understands what went wrong, “it could take this insight and launch a new initiative that has a greater chance of success,” he says.

But not all failures are equal—or acceptable. Company leadership sets the tone, defining the level at which they’re tolerated as well as what happens in their wake. “Good failure is done in the name of progress. It’s exploration as experimentation,” says Samuel West, organizational psychology consultant and founder of the Museum of Failure, a pop-up exhibit of unsuccessful products and services. “Bad failures are due to incompetence, indifference, or sloppiness. We should never accept those … in an organization.”

Employees need to know exactly what constitutes good and bad failures, including the types and sizes of acceptable risk. Amy Edmondson, Novartis professor of leadership and management at the Harvard Business School and author of Right Kind of Wrong: The Science of Failing Well, has identified the criteria for what she calls “intelligent failures.” In general, these have four key attributes: They venture into new territory; have a clear, desired goal; are informed by current available knowledge; and are no larger than necessary, meaning they don’t risk excess resources. “With these criteria in mind, anyone can try something out and feel good about the results even when they fall short of a hoped-for success,” Edmondson writes.

The characteristics of acceptable failures are also determined by the strategy, mission, and risk appetite of the organization. Coopersmith subscribes to the culture of Silicon Valley and venture capital. He says companies should take a portfolio approach, investing in many different ideas and projects, in the hope that some will produce huge payoffs even as many will likely fail. Because technology is developing so fast, companies need lots of eggs in lots of baskets. “You have to go way outside your comfort zone because linear extensions of existing products or experiences aren’t going to cut it, given today’s rate of change,” he says. “There's no one silver bullet anymore. It's silver buckshot.”

However, other businesses may not have the appetite (or deep pockets) for that approach. “The guy who runs the pizza shop on the corner—he doesn't want to fail early and often,” says Jason Freid, cofounder and CEO of 37signals (formerly Basecamp). “Most people who start businesses don't want to fail. They've got to make it work because they may be putting their life savings at risk.” And just because a project, product, or even business doesn’t do as well as expected, it may not be a failure. “It might be a nice 10-million-dollar-a-year business,” says Fried. He points out that labeling something as a failure can negatively affect morale, both organization-wide and for individuals. “People become afraid of doing anything, because who wants to be labeled a failure?”

Such fear corrodes progress and inhibits innovation, says Edmondson, who has spent her career studying the importance of psychological safety within organizations. “Psychological safety describes people’s perceptions of the consequences of taking interpersonal risks in the workplace,” she says. Without it, employees may be afraid to point out potential problems, as happened at Boeing.

Leaders need to prioritize psychological safety and be deliberate in combating fear. “Silence is the default state in organizations,” says Caitlynn Sendra, organizational psychologist and senior research scientist at SAP SuccessFactors.

Creating the right processes and practices

These ideas are little more than wishful thinking unless put into practice. Organizations need to deliberately build processes and communication channels and even create incentives to reinforce and cement what leadership preaches. Consider these best practices as ways of operationalizing a learning-from-failure culture.

Document and report your failures. Few companies (or employees) want to see their failures spelled out in black and white and then broadcast throughout the organization. Yet that’s exactly what they need to do. Only when a company compiles and uses a thorough and detailed record can it learn useful lessons from its failures. At the very least, the record will help the business avoid repeating them. Organizations with high turnover tend to repeat the same mistakes more frequently than those with low turnover, says Coopersmith. “It's not hard to figure out why,” he says. “There may be no one there to say, ‘Hey, we tried that five years ago and it didn't work.’”

Similarly, failures could be repeated if they’re not communicated throughout the organization. West recently came across a company that launched a large AI project. “AI is super exciting, and this department was throwing all kinds of money into it,” he explains. “But the problem was that a different department had already done the same type of project and failed miserably but didn't tell anyone about it. So now the organization has two massive failures, one of which could have been avoided completely.”

Conduct thorough, effective postmortems. A major reason that some organizations don’t learn from their failures is that figuring out why they happened is hard work. “Learning from failure is not straightforward,” says West. “It's complex and takes effort.” Even when organizations regularly hold postmortems, “they usually just go through the motions,” he adds.

Most projects consist of so many different factors that it’s hard to connect cause and effect. It's easy to blame a person or a single factor, such as not enough budget, and move on. “There is a lot of noise in the learning process,” says Kristina Dahlin, professor with special responsibilities at Copenhagen Business School and coauthor of Everybody Fails But Not Everybody Learns: Why Is It So Hard to Learn from Failures? In her book, Dahlin explains the necessity of distinguishing mistakes from outcomes. People and organizations make mistakes all the time, but not all mistakes lead to failure. When it happens, however, the easiest way to explain it is to look back for a mistake to blame.

It's usually not that simple. Dahlin offers an example of an emergency room nurse who gave a patient blood that was the wrong blood type. The mismatch led to the patient's death. Initially, the nurse was blamed, but a closer look revealed a different story. To ensure that patients received the correct blood type, nurses were supposed to scan the barcode on each blood bag before giving it to a patient. But in the urgency of the emergency room, where seconds can be the difference between life and death, they usually didn't have time to do so. The scanning technology was slow, so nurses often skipped the barcode scan. Inadequate technology and a flawed process, not just that single mistake, caused the patient's death. This knowledge not only eased the blame on one nurse but helped the organization correct the problem and avoid similar failures in the future.

Just how much and what type of information should be included in a postmortem? It’s tricky, at least at the beginning, to identify all the possible factors. Among the factors Sendra suggests examining: What were you trying to do? How did you approach it? Were enough support mechanisms in place? Was the project staffed adequately and supplied with the right talent? What tools and technology were used? Was there interpersonal conflict on the team? Did it have the support of leadership?

“The more you document, the greater your ability to go back and say, ‘If we had supplied that project with people with these skills and given them these tools and these resources, it might not have failed,’” says Sendra. “Over time, a lot of things can change: business context, partnerships, technology. That’s why the documentation is so valuable. Maybe a project failed in the past because the technology was ahead of its time. It might actually be worth revisiting, given technological advancements and new capabilities or offerings at the company. Maybe the right time has arrived.”

Indeed, timing can be everything, and plain old luck is never predictable. David Heinemeier Hansson, CTO of 37signals, points out that when he and Fried published their 2013 book, Remote: Office Not Required, “We thought the market was right. But no, it wasn’t,” he says. “If we had published the book [in 2020], we would have hit the wave.”

Such detailed documentation might also highlight what contributes to success, offering useful lessons on what to do in addition to what not to do. “You can look for the common factors that make for a successful hire, for example, or what factors contributed to the success of a product launch,” says Hansson.

It can also help organizations scrutinize the projects they have planned. If employees feel psychologically safe, they might even use premortems. As espoused by cognitive psychologist Gary Klein, this happens at the start of a project rather than the end. At the meeting to kick off a project, the team is asked to imagine that they’re at the end of it and it’s been a huge failure. They then consider and write down every reason they can think of for why it failed.

Recognize the value of lessons from failure. Google, often recognized for conducting excellent postmortems, doesn’t necessarily consider employees who tried but failed at a project as failures themselves. In fact, Google considers their experience as valuable, says Coopersmith. After an unsuccessful project, “Google will say, ‘Now we have these extra people. Who wants them?’” he says. Sometimes another department faces a similar problem in a different market, where the experience is useful. “They don't let that knowledge die; they don't fire everybody. They actually put it to work elsewhere within the organization.”

Some companies reward productive failure. French energy company Engie not only recognizes that learning from failure is important but has established a “best failed ideas” award for employees who attempt an innovative project but don’t succeed. Nominated by the company’s director of innovation, contenders give three-minute speeches to the judges on what they learned.

Technology may help to learn fast from failure

Once a business has implemented the right processes and is sustaining the right culture, using technology could help it learn faster. Powerful AI processing could quickly analyze project data and deliver insights on what went wrong and why and how it could be avoided the next time. It also could quickly surface what all successful projects have in common.

Companies already have internal knowledge management systems with some of this information, says Coopersmith. “They can look back and see what they've tried along similar lines in the past,” he explains. In addition, with increasingly powerful AI search, they might be able to research what other companies in their industries have tried.

They might also be willing to conduct more experiments, as scientists do, to test different hypotheses. Biotech and pharmaceutical companies already use AI to run through many combinations of different types of chemicals and molecules to dramatically accelerate the development of new drugs. And sophisticated simulations speed the design of physical components and products. “The use of AI in all these areas enables us to avoid failure or to identify it early,” says Coopersmith.

Technology is revolutionizing the collection and analysis of data in other areas. In human resources, for example, companies tried for years to build databases of employee skills. But often, by the time the information on people’s skills was collected and entered into the database, it was already outdated due to employee turnover. With AI technology, companies can collect and validate information, infer capabilities, and understand in real time the skills their people have.

The HR example also shows how and why collecting details on failures can pay off. Some companies gave up on the idea of a skills database because it was difficult to keep up to date. “Think about someone suggesting a skills database today, and they’re shot down because the company tried it a few years ago and failed,” says Sendra. If the previous failure had been fully documented, “you could look at what went wrong last time and what has changed. Well, the technology has developed enough so that now this could be a fruitful and worthwhile endeavor.”

In a few years, business leaders may shake their heads about how companies used to make the same mistakes, failing repeatedly, because they didn’t have the technology to dig through all that data.

Maybe. But technology won’t help unless organizations truly embrace the right culture, which allows them to do the hard work of picking apart, openly discussing, and documenting failure, all within a psychologically safe environment. “This is a critical business imperative,” says Sendra. “If you need a reminder about how important it is for employees to feel safe, all you have to think about is Boeing’s experience.”