Designing for failure to drive deeper learning
Helping students struggle with purpose may be one of the most powerful learning intervention we have, it just needs to be designed intentionally
In February 2014, a resurgent José Mourinho—newly returned to Chelsea FC—made headlines by labeling his Arsenal counterpart, Arsène Wenger, a “specialist in failure.” The comment made headlines—not just for its venom, but because it tapped into something deeper: our cultural discomfort with failure.
As a society, we maintain a deeply conflicted relationship with failure. While we often speak of its value for growth and learning, in practice, most of us do our best to avoid it.
Beyond this cultural ambivalence, however, a growing body of scientific research suggests that failure isn’t just an unfortunate byproduct of striving—it may be one of our most powerful tools for learning.
Professor Manu Kapur of ETH Zurich has spent nearly two decades developing the concept of Productive Failure, arguing that failure should not only be tolerated, but designed for.
You could say Professor Kapur is, in fact, a true specialist in failure.
In this interview, I explore his research and the powerful insight it offers: when approached intentionally, failure can unlock deeper understanding and more transferable learning than traditional instruction ever could.
(The interview has been edited for brevity and clarity)
1. What is Productive Failure? How is it different from project-based learning or inquiry-based learning?
Productive Failure (PF) is a learning approach where, before introducing a new concept through direct instruction, learners are first engaged in carefully designed problem-solving tasks that are meant to lead them to failure.
The aim isn't open-ended inquiry but structured exploration—activities that are generative yet constrained—so that the failure itself primes the learner for deeper understanding during subsequent instruction.
Unlike project-based or inquiry-based learning, which are often broader and more exploratory, PF is tightly focused on a specific concept and uses failure as a deliberate preparation for learning.
2. How can AI-driven educational tools integrate the principles of Productive Failure — such as structured struggle and delayed instruction — to prevent over-reliance on automation and foster deeper skill acquisition?
The key lies in developing AI tools that support learners in the generative phase of learning —specifically in helping them engage with problems without immediately revealing the answer.
Instead of providing solutions, the tools we need should evaluate responses and prompt learners to reflect on why their approach might fail. Sometimes, the system might even offer another incorrect answer to deepen that reflection.
The goal is to sustain meaningful exploration that prepares the learner for later insight. Delivering the answer is easy—designing AI that scaffolds the learning process is the real challenge.
3. How do you apply Productive Failure in professional settings, particularly in high-pressure work environments where speed and deliverables are emphasized?
In workplace settings, it’s essential to differentiate between the learning zone and the performance zone. The performance zone is about execution—delivering results quickly and efficiently. But true expertise is cultivated in the learning zone, where there’s room for trial, error, and reflection.
If managers want their teams to develop deep, transferable skills, they must intentionally create opportunities for structured learning—even in fast-paced environments. This means carving out space for experimentation, and accepting that not all efforts will yield immediate results. While performance pressures are real and valid, an overemphasis on output can stifle growth and, over time, erode both talent and morale.
Organizations must consciously balance these two zones to meet their unique demands.
4. Could you walk us through your 4A Framework — Activation, Affect, Awareness, Assembly? How do these stages prime learners both cognitively and emotionally for Productive Failure?
The 4A Framework emerged from our empirical work and represents the core mechanisms behind why Productive Failure works so effectively.
Activation is the first stage—when learners engage with a problem before receiving instruction, they activate prior knowledge, whether accurate or flawed. For novices, this helps surface what they know and don’t know, which is often unclear to them initially.
This leads to Awareness—a recognition of the gap between current understanding and the demands of the task. This awareness is deeply personal and specific, making it a powerful cognitive trigger.
From here, we observe Affect—emotional responses tied to this gap. These range from curiosity and motivation to more difficult emotions like frustration, shame, or guilt. Interestingly, these so-called “negative” emotions can be productive if managed well, priming the learner with a heightened receptiveness to instruction and a stronger mastery orientation.
Finally, comes Assembly—when the learner is cognitively and emotionally prepared, and an expert steps in to structure and consolidate the knowledge. This is the moment where failure becomes productive, as the prior exploration and effort lay the foundation for deeper understanding.
5. Meta-analyses show that Productive Failure nearly doubles learning outcomes compared to traditional instruction. Why does it have such a powerful effect? And if the evidence is so strong, why isn’t it more widely adopted?
First, it's worth noting that this is still relatively recent science. I began working on Productive Failure in 2006, and it was only after the initial publication in 2008 that replication studies and broader experimentation took off, eventually culminating in meta-analyses. So, while the evidence is robust, it's still emerging compared to centuries of entrenched pedagogical norms.
As for its effectiveness, the power of PF lies in its alignment with fundamental learning mechanisms—the 4A Framework: Activation, Awareness, Affect, and Assembly.
These aren't exclusive to PF. They're general mechanisms of how people learn. But PF is uniquely designed to engage these mechanisms deliberately and effectively.
For instance, PF fosters more meaningful activation of prior knowledge than direct instruction, reducing cognitive load. It also carefully designs for affect and awareness, preparing learners emotionally and cognitively for deeper understanding. That’s why it works—but changing teaching culture takes time.
6. Are there circumstances where Productive Failure is not the best approach?
When might direct instruction or other methods be preferable?
Yes, particularly when learners already have a high level of expertise. In such cases, direct instruction can be more efficient.
Imagine a football coach giving a lecture to Premier League managers—their background knowledge is so extensive that structured exploration isn’t necessary.
Productive Failure is most effective for novices, where the cognitive and emotional preparation it fosters can meaningfully support new learning. But as expertise grows, the instructional approach must adapt accordingly.
7. Many forms of classroom instruction fail to produce transfer of learning.
Why do you think Productive Failure leads to better transfer across contexts?
Productive Failure leads to better transfer because it fosters deeper learning and more flexible encoding of knowledge.
Think of it like a LEGO set: if you follow a step-by-step manual to build a spaceship, you may struggle to repurpose those same pieces to create an airplane or a car, because you've only seen them organized in one fixed way. The learning is rigid, tied to a single configuration.
But if you're given the same components and an open-ended prompt, you're forced to explore multiple arrangements. That process allows you to see the versatility of each part.
The same happens to “mental LEGO sets”. PF exposes learners to diverse problem-solving paths, helping them encode knowledge in ways that are more adaptable and transferable to new contexts.
8. This approach seems to require a cultural shift around failure, effort, and growth. How do you "re-norm" failure in learning environments, whether in classrooms or organizations? Is this ultimately a cultural or mindset challenge?
Re-norming failure starts with clear communication of expectations.
Learners often bring preconceived ideas about success and failure into educational or professional settings, so it’s essential to reset those assumptions upfront. You must explicitly define what a psychologically safe space for failure looks like—make it clear that the aim isn’t to get the right answer, but to explore, experiment, and generate even flawed solutions.
For example, you might tell learners that the goal is to come up with as many plausible but incorrect approaches as possible. What matters is not correctness, but the process of thinking deeply. When people understand what’s being measured—and why—they begin to view failure not as a setback, but as an integral part of learning.
Ultimately, this is both a cultural and a mindset shift.
9. Looking ahead, what’s next for Productive Failure research?
What are the most exciting frontiers or unanswered questions you’re exploring?
The first exciting frontier is expanding our understanding of the kinds of resources learners bring to the table in the learning process. Productive Failure is already effective at activating cognitive resources, but we're now exploring how to engage other dimensions—like movement and physical interaction—through the lens of embodied cognition.
Another major development is our work on a new Productive Failure learning app for mathematics. It’s designed to guide students through generative exploration before they encounter formal instruction, essentially priming them for deeper learning. We're still in the early stages with a research prototype, but the potential to scale this kind of experience is something we’re deeply excited about.
I hope you enjoyed the last edition of Nafez’s Notes.
I’m constantly refining my personal thesis on innovation in learning and education. Please do reach out if you have any thoughts on learning - especially as it relates to my favorite problems.
If you are building a startup in the learning space and taking a pedagogy-first approach - I’d love to hear from you. I’m especially keen to talk to people building in the assessment space.
Finally, if you are new here you might also enjoy some of my most popular pieces:
The Gameboy instead of the Metaverse of Education - An attempt to emphasize the importance of modifying the learning process itself as opposed to the technology we are using.
Using First Principles to Push Past the Hype in Edtech - A call to ground all attempts at innovating in edtech in first principles and move beyond the hype
We knew it was broken. Now we might just have to fix it - An optimistic view on how generative AI will transform education by creating “lower floors and higher ceilings”.