The 70-20-10 Model for Learning and Development Explained

By Kevin Kim · Updated Oct 11, 2026

The 70-20-10 model is a rule of thumb for how people develop at work: roughly 70% from challenging experience on the job, 20% from relationships such as managers, mentors and peers, and 10% from formal training. It came out of research on executives at the Center for Creative Leadership. Treat it as a planning heuristic, not a measured law.

That last sentence matters more than the ratio itself. The model is one of the most quoted ideas in learning and development (L&D), and one of the least examined. This guide explains where the numbers came from, what the evidence does and doesn't support, and how to turn the idea into a development plan that works whether or not the exact percentages are right.

What the 70-20-10 model says

The model splits development into three buckets, often called the three E's: Experience, Exposure and Education.

• Experience (70%): Stretch assignments, new responsibilities, problems you haven't solved before, taking over a struggling project, running something from scratch.
• Exposure (20%): Learning from other people through feedback, coaching, mentoring, observing skilled colleagues and working alongside them.
• Education (10%): Courses, workshops, certifications, reading and e-learning.

The Center for Creative Leadership (CCL) describes the rule as emerging from more than 30 years of its Lessons of Experience research, and states that individuals tend to learn 70% of their knowledge from challenging experiences and assignments, 20% from developmental relationships and 10% from coursework and training. CCL also adds a caveat that gets dropped in most summaries: all experiences aren't created equal, so it's worth working out which experiences actually teach which lessons.

One more point from the same source is easy to miss. CCL frames well-designed training as an amplifier that helps clarify, support and boost the other 90% of learning. So the 10% isn't "the part you can skip." It's the part that makes the other 90% stick.

Where the model came from

The research behind the model traces to The Lessons of Experience (1988) by Morgan McCall, Michael Lombardo and Ann Morrison, a study of how successful executives developed. Lombardo and Robert Eichinger later built practical development tools from that line of work.

The ratio itself is a summary of what executives said, not an experiment. Successful executives described the key events in their careers and what they learned from them, and the pattern in their answers became the 70-20-10 shorthand.

How strong is the evidence?

Honestly, thinner than the model's popularity suggests. This is worth knowing before you build a budget around it.

A critique by James Clardy, published in Human Resource Development Review (2018) under the title "70-20-10 and the dominance of informal learning: A fact in search of evidence", argues that the 70% claim should be set aside because of concerns about how the data was collected and sampled. The underlying data came from successful executives in large corporations, relied on self-reporting and may have involved leading survey questions. Clardy still concludes that informal learning matters, but says it can't be left to chance and needs to be examined, structured and managed.

Put plainly, the weaknesses are:

1. Narrow sample. Successful executives at large companies don't represent a first-year analyst, a field technician or a founder.
2. Self-reported data. People recall what felt formative, not what measurably changed their performance.
3. No measurement of learning outcomes. The ratios describe where people say their lessons came from, not how much time or money produces a given result.
4. False precision. Real development varies by role, seniority and company. Nobody's mix is exactly 70-20-10.

None of this makes the model useless. It means the numbers are a prompt to think, and the useful part is the direction they point: most of what people learn at work happens at work, so a development plan made only of courses is missing most of the picture.

70-20-10 compared with other learning frameworks

The model answers a different question than most learning theories. It describes the sources of development; others describe the process of learning or the stages of skill.

The practical combination: use 70-20-10 to decide where to invest, Kolb to make sure experience is reflected on, and the stages of competence to match the support to where the person actually is.

How to apply the 70-20-10 model step by step

You don't need a learning management system or a big budget. You need a clear target and a few honest conversations.

Step 1: Pick one competency to build

Start from a gap, not from a course catalog. Use your core competencies framework or the last performance review to choose one or two competencies for the next quarter. Trying to develop six at once spreads every bucket too thin.

Step 2: Design the experience (the 70%)

Ask: what real work would force this person to practice the competency? Good stretch assignments have three traits: they're slightly beyond current ability, they have a visible outcome, and someone is paying attention to the result. Examples include leading a cross-team project, presenting to the leadership team, owning a customer escalation or onboarding a new hire.

Avoid assignments that are just more of the same work. Repetition builds speed, not capability.

Step 3: Line up the people (the 20%)

Name specific people for specific purposes. A manager gives feedback on the stretch assignment. A mentor outside the reporting line offers perspective. A peer who already does this well becomes someone to shadow. If you're building the mentor side, see mentoring skills, and if the learner tends to deflect input, check their coachability first.

Step 4: Choose the formal piece (the 10%)

Pick the course, book or certification that supports the assignment, ideally just before or during it. A negotiation workshop the week before a major contract renewal gets used. The same workshop taken eight months earlier gets forgotten.

Step 5: Build in reflection

This is where most plans fail. Schedule a short review at the midpoint and end of each assignment: what happened, what did you learn, what would you do differently? That's essentially the reflection stage of the Kolb cycle, and it's what turns experience into learning. Without it, the 70% is just work.

Step 6: Check results against the competency

After the quarter, ask whether the behavior actually changed. Use observable evidence: work output, feedback from colleagues, the next review. If it didn't change, adjust the mix. The model is a starting point, and your own results are better evidence than the ratio.

Template: 70-20-10 development plan

Copy this table into a doc and fill in one row per competency.

Keep the plan to one page. If it needs more, the goal is probably too big.

Examples by business function

These are illustrative plans, not case studies, showing how the same three buckets look in different roles.

Notice that in every row the experience column comes first and is the most concrete. The course is the smallest and most replaceable part.

Do's and don'ts

Do

• Treat the ratio as a guide. If your best development comes from a 50-30-20 mix, that's fine.
• Make the assignment specific. "Gain leadership experience" isn't an assignment. "Run the Q1 planning offsite" is.
• Pair experience with feedback. Experience with no feedback is slower and sometimes wrong. Learners can easily draw the wrong lesson from a result.
• Time formal training close to use. A lesson you can use this week is more likely to stick than one you might need next year.
• Support the manager. The 20% depends on managers who coach. Build that expectation into continuous learning habits across the team.

Don't

• Don't cite 70-20-10 as science. It's a heuristic built on executive recall, and a sharp reader will know that.
• Don't cut training to 10% of budget just to match the ratio. The ratio describes sources of learning, not how to split a budget. CCL itself presents training as an amplifier of the rest.
• Don't throw people in the deep end unsupported. A stretch assignment with no support is a sink-or-swim test, and many people sink.
• Don't apply it identically to everyone. A new hire learning the basics needs more structure and instruction than a senior leader learning a new domain. The four stages of competence are a useful guide here.
• Don't skip reflection. Work alone doesn't guarantee learning.

When the model works best, and when it doesn't

The model fits well when you're developing experienced professionals who already have the fundamentals and need to grow into bigger scope: first-time managers, newly promoted leads, people moving into new domains. Real work gives them a context that no classroom can reproduce.

It fits less well for beginners who need to build foundational knowledge first, for regulated or safety-critical skills where trial and error is unacceptable, and for small teams where there simply aren't enough stretch assignments or mentors to go around. In those cases, structured instruction and supervised practice should take a bigger share than the model suggests.

A related caution: a growth-minded learner will get more from the same assignment than a defensive one. If your team treats mistakes as failure, the 70% won't produce much learning. Building a growth mindset is often the precondition.

Key Facts

• What it is: A rule of thumb that most workplace development comes from experience (70%), relationships (20%) and formal training (10%).
• Origin: Research at the Center for Creative Leadership, which describes the rule as emerging from its Lessons of Experience research; Lombardo and Eichinger built development tools from the same line of work.
• Evidence base: Interviews and surveys of successful executives, using self-reported data. A 2018 critique by James Clardy argues the 70% claim should be set aside because of data collection and sampling concerns.
• What the numbers describe: Where people say their learning came from, not how to divide time or budget.
• CCL's own caveat: Not all experiences teach equally, and good training amplifies the other 90%.
• Best use: A planning heuristic for building stretch assignments, feedback relationships and targeted training around one competency at a time.