Fail, Learn, Innovate, Scale, Repeat: Failure as Fuel for Innovation

Nora Hanke-Louw (International Water Management Institute), Susan MacMillan (ILRI Emeritus Fellow and CGIAR Communications Consultant), Agnes Schneidt (CGIAR Portfolio Performance Unit)  

Failure is one of the few certainties in life, yet it remains one of the hardest things to discuss. Whether in science, innovation, or scaling innovations in agrifood systems, failure often carries an unwarranted stigma, seen as a sign of incompetence rather than a stepping stone toward meaningful progress. 

This is particularly true in science, where negative or inconclusive results rarely make it into publications, creating a misleading picture of success and results. Researchers like Daniele Fanelli and Michael Smaldino have noted that many studies rely on questionable research practices to ensure publishable and novel findings. At the same time, work by Bernard, Edwards, and Roy has highlighted how publication biases and perverse incentives distort scientific evidence. The pressure to deliver positive results is high, sometimes leading scientists to alter methodologies or underreport conflicting data. 

Arguably, the pressure to produce results is as high, if not higher, in international development projects, where, on average, $250 billion is spent annually. In a 2018 study, Matt Andrews examined World Bank-funded public policy reform projects and revealed that 50% fail to achieve impact and 25% fail to deliver on time and budget. This is something not often recognized and discussed in the sector.  

CGIAR researchers thus sit at the interface of two high-pressure environments—increased clamoring for funding and decreased trust in the impacts of that funding—where notice of “failure” is not considered an option, even if ill-defined. 

But what if we saw failure differently? What if, instead of treating it as an obstacle, we recognized it as an essential and inevitable part of learning, adapting, and ultimately succeeding? Being pragmatic about a project’s complexity and uncertainty and providing organizational incentives for ensuring success can increase project performance by 17–20% (Ika and Feeny, 2022).  

During CGIAR’s Scaling Week 2024, a rapidly expanding annual community gathering of scientists, practitioners, and funders working to scale agrifood system innovations, the organizers put failure in the spotlight—thanks to CGIAR colleagues Inga Jacobs-Mata and Nora Hanke-Louw of CGIAR’s International Water Management Institute (IWMI). 

IWMI took the initiative to sponsor a “Failure Fest” during this year’s Scaling Week. Inspired by a well-known format called “Fuckup Nights,” the event was brought to life by Marta Cabañas, a seasoned moderator and speaker coach. It was a night of honest storytelling, humility, and much-needed deconstruction of the myths surrounding failure. Embracing and analyzing failures can unlock the potential for breakthrough innovations—not just in individual projects but also in the agrifood systems CGIAR aims to transform.  

The event: Stories of failure, lessons of resilience  

Failure Fest was a breath of fresh air. Three inspiring speakers from CGIAR’s scaling community shared their professional failures, shedding light on the unexpected obstacles that derailed their projects, the moments of uncertainty, and the lessons learned in the process. Their stories were as useful as they were raw. Each story underscored a fundamental truth: Failure is not the opposite of success; it is a prerequisite for it. By creating a safe space for open discussions, Scaling Week provided a rare opportunity for the week’s participants to confront our collective missteps—not with shame but with curiosity and a drive to do better. This atmosphere of vulnerability and reflection encouraged others to share their experiences, building a culture where learning from failure is accepted and celebrated.  

Why it matters: Failure and innovation in agrifood systems  

In research for development, failure is the elephant in the room. Scientists and practitioners often hesitate to discuss what didn’t work, fearing reputational damage. However, not having candid conversations about what went wrong risks repeating mistakes. More importantly, sharing failures fosters diversity of thought, encouraging colleagues and collaborators to open up and share their missteps and learnings. Normalizing discussions around failure creates space for a richer exchange of ideas, where diverse perspectives can cross-pollinate and drive more inclusive, effective solutions. 

Developing innovations—whether in agrifood systems, policies, or technologies—is messy. Many innovations never pass the pilot stage, and scaling innovations at a system level is incredibly complex. This is why embracing failure as an opportunity for learning is critical. It’s about understanding the “why” behind the failure, iterating quickly, and designing user-centric, adaptive solutions that respond to real-world challenges. 

Scaling Week 2024 reminded us that failure is not inevitable but valuable. Children learn by failing constantly, without fear of judgment. Adults unlearn this skill, internalizing unrealistic concerns of risk and imperfection. Research shows that people overestimate risk, making them less likely to take the leaps necessary for breakthrough innovations. 

Beyond the blame game: Building a failure-positive culture  

One of the greatest barriers to learning from failure is the blame culture embedded in many organizations. Too often, failure is met with avoidance, defensiveness, or outright punishment rather than as an opportunity to reflect and improve. Research by Amy C. Edmondson reported that 70–90% of organizational failures are treated as blameworthy, even though only 2–5% result from negligence or misconduct. This fear of failure discourages openness and prevents organizations from truly benefiting from the lessons that failures provide.  

Individuals and organizations must do the following to create an environment where failure can be leveraged for learning.

1. Foster psychological safety: Encourage open discussions where failure is seen as part of progress, not as a career-ending mistake.  
2. Differentiate between good and bad failures: Some failures are preventable and should certainly be avoided, while others—like those resulting from thoughtful experimentation—should be welcomed (see the spectrum of blameworthy and praiseworthy failures in the graphic).  
3. Incorporate failure analysis into decision-making: Instead of swiftly and quietly burying failed projects, analyze them systematically to extract valuable lessons.  
4. Encourage small-scale experiments: Just as scientists iterate through trial and error, organizations should conduct low-risk tests before scaling an innovation. CGIAR’s novel Innovation Packaging and Scaling Readiness approach can help assess an innovation’s current stage, identify gaps in scaling it more widely, and determine which innovations to prioritize for further testing and/or scaling.
   

Graphic by Amy C. Edmondson, Strategies for Learning from Failure, Harvard Business Review Magazine, April 2011. 

A call to action: Share, learn, and grow together  

The biggest takeaway from the Failure Fest at Scaling Week 2024 was the need to speak up about failure, to normalize it, and to use failure as fuel for innovation.

You can help. If you have a failure story that changed your work, please consider sharing it. Post it in the CGIAR Scaling Community group on LinkedIn, or bring it to your teams. Let’s turn failure from something we try to avoid into something we embrace, analyze, and use as a catalyst for real change.  

CGIAR’s new Scaling for Impact Program fosters demand-driven, evidence-based solutions that bridge research and application. Through innovation systems approaches, responsible scaling, and multi-sectoral partnerships, Scaling for Impact will help turn learning—especially from failures—into concrete, scalable solutions and wide, enduring impacts. An apt motto for the new Program might just be: Fail. Learn. Innovate. Scale. And repeat. 

References 

Andrews, M. (2018) Public policy failure: ‘How often’ and ‘what is failure, anyway’? A study of World Bank project performance. Harvard University Center for International Development Paper No. 344, December. 

Bernard, R., Weissgerber, T.L., Bobrov, E., Winham, S.J., Dirnagl, U., Riedel, N. (2020) Fiddle: a tool to combat publication bias by getting research out of the file drawer and into the scientific community. Clinical Science 134, 2729–2739. https://doi.org/10.1042/CS20201125. 

Bik, E.M. (2024) Publishing negative results are good for science. Access Microbiology 6:000792. https://doi.org/10.1099/acmi.0.000792. 

Edmondson, A.C. (2011) Strategies for Learning from Failure, Harvard Business Review Magazine, April 2011.

Edwards, M.A., Roy, S. (2016) Academic research in the 21st Century: maintaining scientific integrity in a climate of perverse incentives and hypercompetition. Environ. Eng. Sci.34, 51–61. doi:10.1089/ees.2016.0223. 

Fanelli, D. (2009) How many scientists fabricate and falsify research? A systematic review and meta-analysis of survey data. PLoS ONE 4, e5738. doi:10.1371/journal.pone.0005738. 

Grimes, D.R., Bauch, C.T., Ioannidis, J.P.A. (2018) Modelling science trustworthiness under publish or perish pressure. R. Soc. open sci. 5: 171511. http://dx.doi.org/10.1098/rsos.171511. 

Ika, L., Feeny, S. (2022) Optimism Boas and World Bank Project Performance. The Journal of Development Studies. Volume 59. Issue 12. https://doi.org/10.1080/00220388.2022.2102901 

 Ranstam, J. et al. (2000) Fraud in medical research: an international survey of biostatisticians. ISCB subcommittee on fraud. Control. Clin. Trials21, 415–427. DOI: 10.1016/s0197-2456(00)00069-6.  

Smaldino, P.E., McElreath, R. (2016) The natural selection of bad science. R. Soc. open. Sci. 3, 160384. DOI:10.1098/rsos.160384.