Scientists and innovation specialists told the Failure in Science Conference at Imperial College London that academia could learn from industry by treating failure as a normal part of research, making decisions as teams and abandoning unpromising projects sooner. They also warned that excessive risk aversion, pressure to present simple success stories and an overreliance on technological fixes could undermine scientific progress and public trust.
Rethinking failure in science
The discussion followed a session on failure chaired by Dr Nejra van Zalk, a design psychologist at Imperial College London. Reporting back to delegates, van Zalk asked why innovation is prioritised in some areas of healthcare but not others, and who makes those decisions.
“Why do we focus on obesity drugs and pump money into that, rather than helping to fix musculoskeletal issues in children who would have much better outcomes over a lifetime?” she asked.
Dr Stephen Webster, director of The Good Science Project at Imperial College London, introduced the discussion as the fifth episode in a series examining “the business of discovery”. He posed a central question: when it comes to failure, could universities learn from industry?
Van Zalk said the issue was how to incorporate failure safely into working practices. Industry, she noted, can cut its losses, abandon projects and quantify risks in ways that differ from academia. “Could we learn lessons from that?” she asked.
Nessa Carey, a PhD scientist who has worked in academia, industry and as a freelancer, said her own career had involved many forms of failure. In academia, these included rejected grants and papers. In industry, programmes can become impossible to continue, and redundancies are a common experience.
“It’s very easy, particularly the first time, to view that as a personal failure,” Carey said. With experience, however, she came to see that such events are often the result of organisational changes and shifting priorities rather than individual shortcomings.
“Your definition of failure changes as your career changes. And also as you get older, I think it changes as well,” she said.
The cost of becoming risk-averse
At the conference, Professor Jem Woods asked whether society had become “hyper risk averse” and what that meant for science.
Carey agreed that excessive risk aversion was damaging academia, but said the problem extended far beyond universities. She argued that younger people were increasingly expected to achieve a sequence of successes: obtain the right grades, enter the right course and follow a career path that progresses consistently.
Carey said she sees this pressure through her outreach work in schools. Students can feel enormous pressure to perform well in their GCSEs, A levels or equivalent qualifications in other countries because they believe each stage determines the next one.
“We’ve built in this terrible sequential success demand on the young,” she said. Social media can intensify the effect by presenting a stream of promotions, certificates and other achievements. LinkedIn, she noted, is filled with people celebrating their latest successes, while the setbacks and uncertainty of ordinary life remain largely invisible.
“That’s not how life generally progresses. You try stuff, you stuff up, you learn from it, and you move on,” Carey said.
When mistakes are treated as failures rather than as part of learning, people may stay within their comfort zones and avoid taking risks. Carey said this was a serious problem for science, which depends on innovation, as well as for creative industries and organisations more broadly.
“People would rather do nothing than take a risk,” she said.
What academia could learn from industry
One lesson academia could take from industry, Carey argued, is to depersonalise failure. She said industry often provides a stronger sense of shared purpose because people work in teams, whereas academic work is generally more individualistic.
In academia, a failed project can be experienced as a personal and binary outcome: an individual investigator has either succeeded or failed. In industry, by contrast, a team may conduct a decisive experiment and discover that a project cannot progress. The team and its leader can then decide to terminate it.
“In industry, there’s much more an emphasis in good companies on ‘did we make the right decision?’” Carey said. If the decision was correct, the project or programme may have failed, but the individual has not necessarily failed.
She said this distinction was increasingly important because modern science is becoming more multidisciplinary and collaborative. Many major research papers are now large team efforts, yet academic rewards remain largely focused on individual achievement.
“It’s a structural issue. It’s not just about people,” Carey said.
She also suggested that academia could adopt some of industry’s management and staff-development practices. Academic environments, in her experience, are often strongly shaped by individual personalities. That can create a close link between professional setbacks and personal identity, making failure emotionally harder to handle.
Industry, she said, is not free of strong personalities, but its organisational structure can make it easier to separate professional failure from personal worth.
Fail early, fail cheaply
Carey said industry had become increasingly effective at identifying “killer experiments”: tests that can quickly show whether a project is viable. This supports the principle of failing fast, early and cheaply rather than continuing to invest in a project that cannot reach its intended endpoint.
She contrasted this with what she described as a common problem in academia, where researchers may continue working on a favourite target, process or enzyme because they have built a reputation around it or because it can still produce a few papers.
Researchers may also continue a line of work because a PhD student needs data for a thesis. Sometimes this persistence is justified, particularly when curiosity-led research produces unexpected discoveries. But Carey said there was also a risk of continuing simply because abandoning the work would feel like failure.
She did not argue that academia should copy industry in every respect. Instead, she said academic researchers could benefit from a more detached question: “Would we carry on if we were starting now?”
Innovation, trust and communication
Woods also discussed the role of science in addressing major societal challenges. He referred to the emergence of “hyper disruptive technologies”, particularly those affecting how information is managed and how people access and use it.
He linked this to what he called the BBC’s “post-truth society” and argued that science still had important questions to answer about how information is generated, perceived and used by communities and individuals.
Carey said trust had declined not only in science but also in politics, writing and the media. Scientists, she argued, must continue to conduct research with the highest standards of probity. They should also challenge one another appropriately and professionally, and communicate more clearly with the public about both the value and the limitations of science.
She warned that public engagement and science communication had increasingly become additional responsibilities for academics. Because academic workloads rarely shrink when new duties are added, communication can become another box to tick.
This creates a risk that scientists will feel pressure not only to conduct good research but also to produce research that can be presented in a simple and appealing way. Institutions are also under constant pressure to explain the direct, tangible benefit of their work to humanity.
That pressure can encourage researchers to overstate their findings or present complicated evidence in overly black-and-white terms. Such stories are easier to communicate, Carey said, but they can damage public confidence when they are later challenged.
“We shouldn’t be scared of saying it’s complicated,” she said. Overly simplistic communication can undermine trust not just in one study or researcher, but in science as a whole.
The danger of trying to “tech” our way out of problems
Carey said she had worked in innovation for about 25 years and still found it as difficult as when she began. Some challenges have become more urgent, including artificial intelligence, which she described as a technology that may outrun its creators and whose consequences remain unclear.
Another difficulty is identifying which of the vast amount of research being published today has the genuine potential to make a meaningful difference to a particular problem.
Carey cautioned against defining innovation too narrowly as the search for the next major device or drug. Technological innovation may not be the solution to every social or environmental problem.
She pointed to the increasing emphasis and funding directed towards carbon capture as a response to climate change. “Carbon capture at scale is an entirely unproven technology,” she said, adding that it was not clear whether it could solve the climate crisis.
By contrast, she said, there is evidence that switching to electric cars can make a substantial difference, yet the UK is not legislating to support that transition appropriately.
Carey also criticised the billions of dollars that pharmaceutical companies spend developing treatments for chronic health conditions that may be closely linked to social inequality and poverty. More equitable healthcare and food systems, she argued, could deliver a much greater return for society, although the benefits would take longer to appear and would represent a government return on investment rather than a conventional capitalist one.
Her concern was that major failures could result from repeatedly trying to apply technology to problems that do not require technological solutions. Some issues may instead demand smaller, less visible and longer-term interventions that generate little profit for companies.
“We will ‘tech’ when we shouldn’t be ‘tech-ing’,” Carey said. “We will try and ‘tech’ very expensively when actually we could have made smaller investments with much more impact.”
She also criticised the tendency among scientists to promise a technological solution rather than acknowledge when social or political action is needed. Social scientists may already understand what interventions are required, she said, but those solutions are often not implemented.
“That, I think, is a failure on the part of science to pretend that we are the ones who will solve everything,” Carey said.
Source: www.nature.com


