“Done is better than perfect.”
Sheryl Sandberg popularized this Facebook maxim in Lean In as a response to perfectionism that delays contribution. The line values completion because unfinished work cannot serve a customer, invite feedback, or create the evidence needed for improvement. Perfection is often undefined and unreachable; its pursuit can conceal fear of judgment or a demand that a person prove readiness beyond what a role requires. “Done” creates a real object that others can use, test, and refine. In that sense, completion is not the end of quality but the beginning of learning.
Teams can apply the principle by defining what “done” means before work expands. Acceptance criteria should identify the customer outcome, safety threshold, owner, and evidence required for release. Work can be divided into small increments that deliver value without pretending the first version is final. Timeboxes and review dates stop endless polishing, while feedback supplies priorities for the next iteration. Managers should distinguish consequential quality from cosmetic preference so that attention goes to the defects and uncertainties that matter most.
The sentence is not permission to ship unsafe, inaccessible, misleading, or unreliable products. In medicine, finance, security, and other high-stakes settings, a premature release can make later improvement impossible for the people harmed. The appropriate standard depends on reversibility and consequence. Leaders should also resist using the maxim to justify chronic understaffing or to dismiss craft. The choice is not perfection versus carelessness; it is sufficient quality for the next responsible learning step versus delay that adds little value.
Generative AI makes rough drafts abundant, which raises rather than removes the need to define completion. A polished-looking output may contain unsupported claims, privacy risks, or subtle bias. Teams can require verification, human accountability, and disclosure appropriate to the use. They can ship internal pilots before external promises and make rollback possible. Sandberg’s phrase works best as a bias toward action inside clear guardrails: finish something honest and useful, learn from reality, and keep improving instead of waiting for confidence that no process can supply in advance. A release ladder can make “done” proportional to consequence. Internal drafts require clarity and privacy checks; bounded customer pilots add reliability, consent, and support; broad launch adds scale, accessibility, security, and recovery. Each level has a named owner and exit criteria. The ladder preserves momentum without pretending every artifact needs the same polish. It also prevents leaders from moving work into public use simply because a prototype looked convincing during a favorable demonstration.
Sandberg popularized the Facebook maxim in her 2013 book Lean In.
She used it to challenge perfectionism that delays contribution, learning, and advancement.
Clear acceptance criteria and small increments turn completion into a responsible path toward feedback and improvement.
High-stakes work still requires rigorous safeguards; the phrase does not excuse unsafe quality or chronic under-resourcing.
AI can produce polished drafts whose defects are harder to see than their fluency suggests.
Verification, accountability, pilots, and rollback define responsible “done” when generation is easy.
Explore more "Quotes of The Day"
Discover more notable quotes from influential voices across politics, science, business, technology, sports, and culture. Each quote offers insight into how ideas, beliefs, and decisions shape the world around us.