Fail Better: The Craft of Iterative Resilience
Created at: September 19, 2025
Try again. Fail again. Fail better. — Samuel Beckett
Beckett’s Stark Invitation
Samuel Beckett’s line from Worstward Ho (1983)—“Ever tried. Ever failed. No matter. Try again. Fail again. Fail better.”—strips encouragement to its barest frame. Rather than promising success, it normalizes failure as the prevailing condition of human effort. In doing so, it lowers the emotional cost of trying and reduces the shame that often blocks the next attempt. Flowing from this austerity is a paradoxical liberation: if failure is the baseline, then action replaces avoidance. The phrase is not cheerful bravado but disciplined acceptance; it recasts progress as a sequence of controlled stumbles. That shift in expectation becomes the foundation for a practical method.
From Philosophy to Method
Carrying Beckett’s ethos into practice, we find a close kinship with scientific fallibilism. Karl Popper’s The Logic of Scientific Discovery (1959) argues that knowledge advances by attempted refutations, not proof. In that light, “try again” becomes hypothesis, test, and revision; “fail again” becomes evidence sharpening the next design. Thus, failure is not a verdict but a measurement instrument. By treating each miss as information, we transform discouragement into data. This procedural humility links existential realism to rigorous iteration.
Prototyping Creativity
Creative fields embody this cadence through prototypes and drafts. James Dyson’s Invention: A Life (2021) recounts building thousands of vacuum prototypes—5,127 by his count—before a marketable design emerged. At Pixar, Ed Catmull describes “plussing,” a practice of iteratively improving flawed ideas without blame (Creativity, Inc., 2014). Anne Lamott’s Bird by Bird (1994) legitimizes the “shitty first draft,” making imperfection a starting line rather than an indictment. In each case, the intention is consistent: degrade the stigma of early failure so feedback can flow. As prototypes evolve, the work “fails better” by revealing exactly which edges need refining.
Psychology of Productive Errors
Moreover, psychology backs Beckett’s cadence. Carol Dweck’s Mindset (2006) shows that a growth mindset reframes errors as pathways to ability, not verdicts on worth. Robert and Elizabeth Bjork’s “desirable difficulties” research (2011) finds that well-calibrated challenges—where mistakes are likely—enhance long-term learning. Likewise, Anders Ericsson’s work on deliberate practice (1993) demonstrates that targeted struggle at the edge of competence drives expertise. Together they suggest that the quality of failure matters. When errors are specific, timely, and informative, they convert frustration into traction.
Designing Safe-to-Fail Systems
Yet not all arenas afford the same latitude. High-stakes fields embed safeguards so learning does not cost lives. Gary Klein’s “premortem” (HBR, 2007) anticipates failure in advance, while Atul Gawande’s The Checklist Manifesto (2009) standardizes critical steps to reduce catastrophic error. In software, feature flags and canary releases let teams test in production while limiting blast radius. Thus, to “fail better” is to architect risk: make experiments small, consequences bounded, and feedback immediate. The method is bold in intent but modest in increments.
Stakes, Equity, and Accountability
At this point, an ethical caveat emerges: the freedom to fail is unevenly distributed. Amy Edmondson’s work on psychological safety (1999) shows that teams learn more when members can speak up without fear, yet accountability must remain. Aviation’s “just culture,” influenced by James Reason’s safety research (1997), balances candor about mistakes with responsibility for reckless acts. Consequently, failing better demands institutional design as much as personal grit. It requires fair cushions, clear norms, and a commitment to learn rather than to punish.
A Practical Loop for Failing Better
Finally, Beckett’s sentence can be operationalized as a loop. John Boyd’s OODA cycle—Observe, Orient, Decide, Act—emphasizes rapid, thoughtful iteration; Eric Ries’s The Lean Startup (2011) translates it into build–measure–learn. Begin by defining the smallest meaningful test, run it quickly, capture concrete signals, and adjust the next attempt accordingly. In this light, “Try again. Fail again. Fail better.” becomes a cadence of shorter bets and sharper learning. The goal is not to avoid failure, but to convert it—reliably—into forward motion.