[ccl] The origins of 70/20/10

I've blogged before on how the 70/20/10 model became the talk of the town - or at least the talk of the learning and development department. The model (not law or recipe) is intuitive and catchy and states that roughly we learn 70% from experience, 20% through others and 10% via courses.

The model is often traced back to research done at the Center for Creative Leadership (CCL) in 1996, and has since started a life on its own. I'm currently working 4 months at CCL and want to share what I now know on the origins of the model. This will help everyone to better place it. I've seen organisations using 70/20/10 sensibly, and I've seen organisations turning it into a dogma or a justification of budget cuts. I've seen questions on the validity (as a rule of thumb: always be suspicious with models that have rounded figures), I've read people arguing the numbers don't matter, and I've seen it pop up in the discussions on social learning. Whether you are a 70:20:10 advocate or sick of its tyranny, here is what you should know about the 1996 and following studies to make up your mind and have meaningful conversations on this topic.

I'm including a slideshow from my colleague and researcher Cindy McCauley used at a conference board in January. She was so kind to let me upload it:

70 20-10 research (mc cauley)

View more PowerPoint from Bert De Coutere

1- The model is derived from leadership research. 

The initial research "Lessons of Experience" was targeted at managers. It was essentially a survey and the question asked to succesful executives was:

When you think about your career as a manager, certain events or episodes probably stand out in your mind—things that led to a lasting change in you as a manager.  Identify at least three “key events” in your career:  things that made a difference in the way  you manage now.

1. What happened?
2. What did you learn from it (for better or worse)?

In the current discussions on the 70/20/10 model it is used for any type of learning. Maybe that holds as true, but the original research was about leaders and managers. From there it's a small step to other behaviors, or is it? The research didn't go into any other kind of topic or audience. For example, it didn't contrast with unsuccesful managers, it didn't touch upon any other job category, it asked for the key events only, and it didn't ask about an individual skill. Also, it asks about long-term development over a career.

2- The original research had 5 categories.

The outcome of the question above generated 5 categories of top developmental events. Only three of those are now in the 70:20:10 model. Here were the original figures:

Event Category	Frequency (percentage of all events)
Challenging assignments	55.8 %
Hardships	17.4 %
Other people	18.0 %
Coursework	6.2 %
Personal Life Events	2.6 %

If you cut the categories 'personal life events' and 'hardships', the others recalculate pretty close to the rounded numbers of the 70:20:10 model. Challenging assignments than gets 69.8% and becomes 'experience', other people recalculates at 22.5% and coursework at 7.7%.
One reason for cutting the categories of personal life events and adversaries is that organisations would not provide for these, they just 'happen'. When they do happen, they have profound developmental impact on leaders. I think one of the best things the 70:20:10 model does, is it breaks open development from just looking at courses. Just know that the original research went also further than the current 3 categories, so if you want to see the big picture you need to also factor in the developmental events that 'just happen' rather than being provided for.

3- The research was repeated over the years.

The 'lessons of experience' research continued over the years, and leaders kept reporting similar frequencies for the different developmental events in the same categories. Have a look at the slides above, particularly slide 5. It lists the frequencies for 4 kinds of studies. The findings are that for executives the 70/20/10 is a pretty close rounding of the research facts. If you want to look at a broader range of management levels, then the results would suggest more of a “50/45/5” distribution (based on the Lessons of a Diverse Workforce Study). And for succesful woman leaders it was more close to 55-40-5.

So what does that tell us? Is 70/20/10 only for white, Americal, executive males?

Since 2004, the 'lessons of experience' research includes other geographies such as China, Russia and India. You can see the results on slide 6. The latest whitepaper on this research talks about the 'big 5 plus 2': 5 broad global categories, with two specific ones to the local culture/geography. You can download the "Grooming top leaders" paper on the CCL website. In this research, the descriptive data on the developmental activities is much more important than the proportional number. And in CCL's view, it is the whole package that counts.

4- ... and is now applied way beyond the original research findings

Nowadays the rule/model/finding is omnipresent in learning land. It is applied to all learning, for all behaviors and skills, for all audiences novices and experienced people alike, and for long term and short term development. For an excellent article on what the rule became independent from its research grounds please read Charles Jenning's post and see his slideshare presentation.  For me, I think 70:20:10 -if nothing else- reminds us of the holistic way we learn and helps us (re)focus on the different elements of the package, as the sole focus for too long had been on courses. Just that is a tremendeous way forward. Don't make more of 70:20:10 than it is though. (Hint: it is not a law or a general recipe.) Think about balance rather than proportions.


Many, many thanks to my colleague Cindy for the info and for allowing me to publish her slides here.