A learning framework that suggests 70% of workplace learning comes from on-the-job experience, 20% from social interactions and mentoring, and 10% from formal training programs.
Key Takeaways
The 70-20-10 model starts with a simple observation: most of what people know about their jobs, they didn't learn in a classroom. They learned by doing. By making mistakes. By watching experienced colleagues. By getting feedback from their manager. The formal training course was useful, but it was the smallest piece of the puzzle. Morgan McCall and his colleagues at the Center for Creative Leadership discovered this pattern while studying how successful executives developed their skills. When they asked these leaders where they learned the most, the answers clustered around three categories: challenging work experiences (70%), relationships and interactions with others (20%), and formal coursework and training (10%). The model has become one of the most widely referenced frameworks in L&D. Not because the specific percentages are scientifically precise, but because the underlying principle resonates with reality. If your L&D strategy is 100% focused on building courses and buying e-learning licenses, you're investing in the 10% while ignoring the 90%.
Experiential learning is the largest component and the hardest to design deliberately. It's what happens when employees face new challenges, solve real problems, and learn from outcomes.
Give employees projects that push beyond their current skill level. Not so far that they fail, but far enough that they grow. A mid-level marketer leading a cross-functional product launch. A developer architecting a system for the first time. A regional manager handling their first international market. Stretch assignments work because they create productive struggle. The brain encodes information more deeply when there's emotional engagement and real stakes.
Moving employees temporarily into different roles, teams, or departments builds breadth. A finance analyst spending three months in operations understands cost structures differently afterward. A software engineer who rotates through customer support writes better user-facing features. Companies like GE, P&G, and Samsung have built leadership pipelines around structured rotation programs for decades.
Real business problems are the best learning material. Cross-functional task forces, process improvement projects, and innovation challenges put employees in situations where they must learn quickly, collaborate across boundaries, and produce results. The key is choosing projects that are genuinely important, not makework disguised as development.
Experience alone doesn't produce learning. Reflection does. After-action reviews (originally a US Army practice) ask four questions: What was supposed to happen? What actually happened? Why was there a difference? What will we do differently next time? Without structured reflection, people can repeat the same year of experience ten times rather than having ten years of progressive growth.
Social learning happens through relationships: mentoring, coaching, peer feedback, communities of practice, and day-to-day collaboration.
Pairing less experienced employees with seasoned professionals transfers tacit knowledge that can't be captured in a course. A mentor shares not just what to do, but how to think about problems, who to talk to, what political dynamics to watch for, and what mistakes to avoid. Formal mentoring programs with matched pairs, structured meeting cadences, and defined goals produce better outcomes than informal arrangements.
The direct manager is the most important learning relationship for most employees. Regular one-on-ones that include development conversations, real-time feedback after key moments, and coaching questions ("What did you learn?" "What would you do differently?") turn daily work into continuous learning. Gallup data shows that employees who receive daily feedback from their manager are 3.6x more likely to be engaged.
People with similar roles sharing knowledge and solving problems together is one of the most cost-effective learning strategies. Engineering guilds, sales roundtables, HR communities of practice, and cross-office knowledge-sharing sessions all fit here. Platforms like Slack, Teams, and dedicated community tools make peer learning possible across geographies.
Formal training is the smallest slice, but it's not unimportant. It provides the foundational knowledge and frameworks that make experiential and social learning more effective.
Formal training is essential when: employees need foundational knowledge before they can learn through experience (new hire onboarding, compliance requirements), skills require structured practice with expert feedback (negotiation, presentation, coding), certifications or credentials are needed for regulatory compliance, or the organization is adopting entirely new processes or technologies. The 10% creates the scaffolding that the 70% and 20% build on.
Since formal training is a small part of the learning mix, it needs to be excellent. Apply spaced repetition instead of single-day workshops. Build in practice and simulation, not just information delivery. Connect training directly to on-the-job assignments so learners apply new skills immediately. Use pre-assessments to skip content learners already know. Follow up with coaching and peer discussion to reinforce key concepts.
The model works best as a design principle, not a budget allocation formula. Here's how to put it into practice.
The 70-20-10 model is widely used, but it's not without valid criticism. Understanding its limitations helps you apply it more effectively.
The original study surveyed approximately 200 successful executives at large corporations. It wasn't a randomized controlled trial, and the sample wasn't representative of all workers. The exact percentages (70, 20, 10) came from retrospective self-reports about how executives believed they learned. Memory-based self-assessment isn't the most reliable research methodology. Subsequent studies have found different ratios depending on the industry, role, and experience level.
For some roles, formal training is far more than 10% of effective learning. Surgeons, pilots, nuclear plant operators, and cybersecurity analysts need extensive formal training and simulation before experiential learning is safe or appropriate. Entry-level employees often need more structured instruction before they can learn effectively from experience. The model works best for professional and managerial roles with moderate existing expertise.
Some organizations use the model to justify slashing training budgets. "If 70% of learning is on-the-job, why are we spending money on courses?" This misses the point. The 10% provides the knowledge foundation that makes the 70% productive. Without it, experiential learning becomes trial-and-error without a framework for understanding what went right or wrong.
The 70-20-10 model is one of several frameworks L&D teams use. Here's how it compares.
| Framework | Focus | Best Used For | Key Difference from 70-20-10 |
|---|---|---|---|
| 70-20-10 Model | How people learn (experience vs. social vs. formal) | Designing blended learning strategies | Foundational framework for learning mix |
| ADDIE Model | How to design training (Analysis, Design, Development, Implementation, Evaluation) | Instructional design projects | Process model for building content, not a learning philosophy |
| Kirkpatrick Model | How to evaluate training effectiveness (4 levels) | Measuring L&D impact | Evaluation framework, doesn't address how learning occurs |
| Bloom's Taxonomy | Levels of cognitive learning (Remember to Create) | Writing learning objectives | Classifies learning depth, not learning sources |
| SAM (Successive Approximation) | Agile instructional design with rapid prototyping | Fast-turnaround content development | Design methodology focused on iteration speed |
| Kolb's Experiential Learning Cycle | Four-stage learning cycle (Experience, Reflect, Conceptualize, Experiment) | Designing experiential learning activities | Deeper model for the 70% component specifically |
Data that shows the impact of balanced learning approaches across organizations.