Ask most professionals how people get good at their jobs and the answer is some version of time: years on the role, cases handled, hours logged. The uncomfortable finding at the center of K. Anders Ericsson's research program is that time is a poor predictor. Experience accumulates automatically; expertise does not. What builds expertise is a specific, effortful, designed kind of activity that Ericsson named deliberate practice, and it is rare almost everywhere, including at work.
This guide covers what Ericsson actually claimed, the popular distortions of the concept, the actual state of the evidence, why everyday work fails to provide deliberate practice, and what it takes to design its conditions into workplace learning.
What did Ericsson actually claim?
The concept comes from Ericsson, Krampe, and Tesch-Römer's 1993 Psychological Review article, a study of violinists at an elite music academy in West Berlin. The researchers were searching for the conditions of optimal learning that a century of laboratory research had identified: full attention and effort directed at improvement, immediate informative feedback on results, and repeated performance of the same or similar tasks. Music training turned out to embody those conditions unusually well. A master teacher assesses each student individually, identifies what that student can improve before the next lesson, and prescribes practice activities to get there. The solitary practice the student then does between lessons, designed and guided by the teacher, is what the authors named deliberate practice. Their central empirical finding came from comparing three groups at the same academy: the best violinists, the good violinists, and violinists training to become music teachers. All three groups had played for roughly the same number of years. What separated them was how much of this one activity they had accumulated: the best violinists had logged thousands more hours of deliberate practice than the good ones, who had in turn logged thousands more than the future teachers. Same instrument, similar years of experience, different amounts of designed practice, different levels of performance.
In a later clarifying paper, Ericsson and Harwell (2019) spelled out the criteria that make an activity deliberate practice. First, the training is individualized and designed by a qualified teacher or coach who can diagnose the learner's current performance and prescribe techniques with established effectiveness. Second, the teacher communicates a goal that the learner can mentally represent, an internal picture of the target performance the learner cannot yet produce. Third, the practice activity provides immediate feedback on each attempt, so the learner can compare the attempt against the goal, reflect, and problem solve. Fourth, the learner makes repeated, revised attempts that gradually approach the goal.
The other half of the claim is what does not build expertise. Reviewing evidence across domains including medicine, Ericsson (2008) reported that length of experience, reputation, and perceived mastery show only a weak relationship with actual, observed performance, and that performance does not necessarily improve with more years in the profession. Expert performance traces instead to training "focused on improving particular tasks" (Ericsson, 2008), with feedback, evaluation, and refinement built in. Experience is what everyone accumulates by default; deliberate practice is the exception that has to be designed.
What is deliberate practice not?
The concept has suffered badly from popularization, and precision matters because the distortions lead to bad design decisions.
It is not mere repetition or time-on-task. Ericsson and Pool (2016) introduced a vocabulary for two of the weaker forms. Naive practice is simply engaging in the domain: playing casual games, or, for professionals, executing the job as demands arise. Purposeful practice is self-directed effort toward specific improvement goals without a teacher's guidance. Ericsson and Harwell (2019) named a third category that sits in between, structured practice, group training designed by a coach but not individualized to the learner. All of these can produce some improvement. None of them is deliberate practice, and the research consistently finds them less effective.
It is not the 10,000 hours rule. Gladwell's Outliers (2008) built a chapter on that number and cited the 1993 paper as its evidence, but Ericsson and Harwell (2019) were blunt about the misreading. There was no evidence for a magical threshold; Ericsson estimated that winning international piano competitions takes closer to 25,000 hours. Gladwell never mentioned deliberate practice at all, and his examples counted public performances and ordinary work, activities that violate the criteria. The popular version says that enough hours make anyone an expert. The research says that hours of the wrong activity make almost no one an expert.
It is not even lifetime accumulation of the right activity. Krampe and Ericsson's study of older pianists, reviewed in Ericsson and Harwell (2019), found that older experts had accumulated roughly 57,000 hours of solitary practice against roughly 18,000 for young experts, yet performed no better. Much of the later practice merely maintained existing skill. Recent practice predicted current performance better than career totals. Expertise does not accumulate like a bank balance; it is a condition sustained by continued, designed effort.
How strong is the evidence really?
An accurate treatment has to acknowledge two boundaries. The first is origin: the strongest evidence comes from music, chess, sports, typing, and memory tasks, domains with centuries of accumulated training knowledge, objective performance measures, and full-time learners. Medicine has a substantial simulation-based literature applying the principles. Ordinary professional work is the least studied territory, which is precisely where instructional designers operate.
The second is the effect-size debate. Macnamara, Hambrick, and Oswald's 2014 meta-analysis of 88 studies concluded that practice explained about 26 percent of performance variance in games, 21 percent in music, 18 percent in sports, 4 percent in education, and less than 1 percent in the professions, for roughly 14 percent overall. Ericsson and Harwell (2019) responded that the meta-analysis had substituted a broader definition, counting group team training and even self-reported hours of studying as deliberate practice, and relabeled that category structured practice. Applying the original criteria and excluding studies that did not measure them, their reanalysis estimated 29 percent of variance explained, rising to 61 percent after correcting for the unreliability of retrospective practice estimates and performance measures.
The fair summary is that the variance explained by practice is substantial but contested, and that no serious party claims practice explains everything. What is not contested, on either side of the debate, is the direction of the design implication: activities that meet Ericsson's deliberate practice criteria outperform activities that do not, and unguided experience is the weakest condition of all. Designers do not need the dispute resolved to act on that.
Why does everyday work fail to provide deliberate practice?
Hold the four criteria up against a normal job and the mismatch is systematic.
Work is not designed for improvement. Ericsson et al. (1993) explicitly contrasted deliberate practice with work and play, which are motivated by production and enjoyment rather than by improving a targeted aspect of performance. A manager handling a difficult conversation is trying to get through it, not to refine one component of how they handle it. In Ericsson and Pool's terms, executing the job as demands arise is naive practice.
Feedback is delayed, noisy, or absent. A musician hears every wrong note instantly. A professional who makes a poor judgment call may learn the outcome weeks later, entangled with a dozen other causes, or never learn it at all. Without immediate informative feedback, repetition rehearses errors as readily as it corrects them.
Repetition is unsafe. The situations that most need work, the escalation mishandled, the objection fumbled, the warning sign missed, are exactly the ones organizations cannot let people fail at repeatedly with real customers, patients, or colleagues. There is no retry button on consequential work.
The predictable result is the plateau. Ericsson and Harwell (2019) note that everyday skill acquisition follows the classic pattern described by Fitts and Posner: performance automates within weeks or months as people minimize effort, and improvement stops. This is why experience correlates so weakly with performance (Ericsson, 2008). A plateau says nothing about the person; it is simply what forms when none of the four conditions is present. The same research offers the corrective: plateaus are not fixed and can be overcome when practice is redesigned and guided.
How do you design the conditions into workplace learning?
Honoring deliberate practice in workplace learning does not mean assigning more repetitions or longer hours. It means engineering each criterion into the learning experience.
Well-defined tasks drawn from real situations. The unit of design is a specific, recurring situation where performance actually diverges, framed as a decision the learner must make, not a topic to review. This is the workplace analog of the teacher selecting the passage the student needs.
A goal the learner can represent. Learners need a concrete model of what good performance looks like in that situation, expert reasoning made visible, so their attempts have a target. This is where deliberate practice converges with the modeling and coaching methods described in our cognitive apprenticeship guide.
Immediate, informative feedback at the decision. Feedback must arrive at the moment of choice, while the learner's reasoning is still available for inspection, and must explain what an expert would notice and weigh, not just score the answer. A delayed grade fails the criterion.
Safe retry and refinement. The learner must be able to attempt the situation again, adjusting based on feedback, without real-world consequences. This is the condition workplaces are least able to supply on the job and simulation exists to supply off it.
Progressive variation. As performance improves, situations should become more varied and demanding, so the learner keeps working at the edge of ability rather than automating and plateauing.
Guided scenarios are one direct implementation of this specification. In AliveSim's Guided Scenarios, learners who have already acquired knowledge through traditional methods encounter well-defined decision tasks drawn from real situations, receive expert feedback at the moment of choice, retry safely without real-world consequences, and move through progressively varied situations, the same four conditions just described, built into the AliveSim experience rather than left to chance.
Human coaching still matters, and the theory says exactly where. Designed scenarios supply the four conditions inside the experience, but the first criterion also describes a longitudinal role: someone who watches an individual's performance over time, diagnoses idiosyncratic weaknesses, decides what to work on next, and adjusts the program as the person develops. That is a manager's or mentor's job, and no scenario library replaces it. The realistic division of labor is that designed experiences deliver the feedback-rich, safely repeatable attempts that calendars of human experts never could, while human coaches use the freed capacity, and the performance data those experiences generate, to guide each person's longer arc of development.
References
- Ericsson, K. A. (2008). Deliberate practice and acquisition of expert performance: A general overview. Academic Emergency Medicine, 15(11), 988–994.
- Ericsson, K. A., & Harwell, K. W. (2019). Deliberate practice and proposed limits on the effects of practice on the acquisition of expert performance: Why the original definition matters and recommendations for future research. Frontiers in Psychology, 10, 2396.
- Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363–406.
- Ericsson, K. A., & Pool, R. (2016). Peak: Secrets from the new science of expertise. Houghton Mifflin Harcourt.
- Gladwell, M. (2008). Outliers: The story of success. Little, Brown and Company.
- Macnamara, B. N., Hambrick, D. Z., & Oswald, F. L. (2014). Deliberate practice and performance in music, games, sports, education, and professions: A meta-analysis. Psychological Science, 25(8), 1608–1618.
Related questions
Is the 10,000 hour rule real?
Not as popularly stated. Malcolm Gladwell's 2008 book Outliers presented 10,000 hours as the threshold for expertise, citing Ericsson's research. Ericsson himself rejected the rule: his data showed no magic number, and he estimated that winning international piano competitions takes closer to 25,000 hours. More importantly, Gladwell never used the term deliberate practice, and his examples counted activities such as public performances and ordinary work that violate its criteria. Ericsson's actual claim was about the type of activity, not a quantity of hours. Accumulating time in a domain guarantees little; accumulating designed, feedback-rich, effortful training is what the evidence connects to expert performance.
What is the difference between practice and deliberate practice?
Most of what people call practice is repetition of things they can already do. Ericsson and Pool distinguished naive practice, which is simply engaging in the activity, such as doing your job as demands arise, from purposeful practice, which is self-directed effort toward specific improvement goals, and from deliberate practice, which adds a qualified teacher or coach who diagnoses the individual's performance, sets goals the learner can mentally represent, prescribes activities with immediate feedback, and supports repeated, refined attempts. The distinctions matter because the evidence for strong performance gains attaches to the designed, individualized forms, not to time spent in the domain.
Does deliberate practice work for soft skills?
The strongest evidence comes from music, chess, sports, and medicine, so honesty requires saying that interpersonal skills are less studied under the strict definition. But nothing in the conditions restricts them to technical domains. If an organization can define what good performance looks like in a difficult conversation, present realistic situations that call for it, give expert feedback at the moment a learner chooses a response, and allow safe retry, the conditions are met. The practical obstacle is not the theory; it is that workplaces rarely make expert judgment about conversations explicit enough to feed back immediately, which is exactly what designed scenario experiences exist to fix.
Did the Macnamara meta-analysis debunk deliberate practice?
It complicated the picture rather than debunking it. Macnamara, Hambrick, and Oswald's 2014 meta-analysis found practice explained about 14 percent of performance variance overall, and far less in education and the professions. Ericsson and Harwell replied in 2019 that the meta-analysis had used a broader definition, counting group training and even hours of studying as deliberate practice. Reanalyzing only studies that met the original criteria, they estimated 29 percent of variance explained, or 61 percent after correcting for measurement unreliability. Both sides agree extended practice is necessary for expertise and that its quality matters; the unresolved dispute is how much variance it explains and what role other factors play.
Published July 8, 2025 · Updated July 15, 2026 · 10 min read