We can all give up – 70:20:10 has arrived | TrainingZone.co.uk

We can all give up – 70:20:10 has arrived

Main points: 
We all know the stats. But what do they actually mean, and how accurate are they? Robin Hoyle turns mythbuster for TrainingZone.
Trainers like a good buzzword. We also like something which appears scientific, concerned as we are that maybe the rest of the world thinks of us as prone to a bit of mumbo jumbo. If we can put numbers in there somewhere – hallelujah! 
Welcome to 70:20:10.
You might have seen the ratios 70:20:10 used elsewhere in relation to organisational training and development. I have come across it spoken about a lot in papers, presentations at conferences and business's learning strategies.
The version I last heard in a board room was "most learning happens on the job – its experience that counts". 
This focus on work-based practice and on-the-job experiences resonates because it makes sense. For many – especially those outside the training department - it represents how they think they learned to do the things they do. It also neatly reinforces the prejudice that 'you can't learn to do your job out of a book or in a classroom'.
Of course, these beliefs are fundamentally true. I don't want to be flown on my next holiday by someone who has only read about how to fly a plane. My children are currently learning to drive – does passing the driving theory test make them safe on the roads? Not if the permanent impression of my finger nails in the palms of my hand mean anything. 
The gap between theory and practice is canyon-esque in proportion.
"As far as I can tell, no one has analysed how people gained their skills and determined that the proportions of experience and input outlined in the 70:20:10 model is actually what happens."
So we can all accept, I hope, that we need a mix of different learning and training opportunities to develop the skills necessary for work. What the 70:20:10 concept says is that we learn 10% of our skills and knowledge through formal training inputs – workshops, elearning, etc – 20% from conversations with more experienced people (coaching, informal learning) and 70% from getting our hands dirty and doing things.
I have no argument with that as a concept – but let's be clear, that's what it is. It isn't – as far as I can see - based on empirical research. (I'd love to be proved wrong – if someone has more information on research which shows this is how things actually happen then please let me know). As far as I can tell, no one has analysed how people gained their skills and determined that the proportions of experience and input outlined in the 70:20:10 model is actually what happens. For one thing, the nice round numbers always seem suspicious to me. In fact a recent study by Harold Jarche suggested that formal training – the stuff that most of you reading this are involved in – is actually less than 5% of the training mix, with more than 95% of the time spent not undertaking formal training and – one would hope – learning from our experiences.
The 70:20:10 learning concept was first developed by Morgan McCall, Robert W. Eichinger and Michael M. Lombardo at the Center for Creative Leadership in North Carolina. The model is specifically mentioned in 'The Career Architect Development Planner', by Lombardo and Eichinger. 
What is interesting when you go to the source of the model is that the focus is not on some ad hoc process by which skills are somehow developed in an organic way through simply doing the job. Instead, 70:20:10 as imagined by its originators requires thorough planning and management of the informal and on-the-job learning if it is to work.
"If the model is misrepresented...then are we inviting our sponsors and senior managers to embrace a training process in which we simply throw our people in the deep end because that's how they'll learn anyway."
Failure to properly understand the concept is the big issue. From 'we have a method which will help people learn very efficiently and develop practical skills using on the job experience' to 'most learning happens on the job, why are we bothering with courses' is a small leap in the minds of the bean counters tasked with finding cost savings.
If the model is misrepresented as '70% of learning happens on the job' – which is not what Eichinger and Lombardo said - then are we inviting our sponsors and senior managers to embrace a training process in which we simply throw our people in the deep end because that's how they'll learn anyway.
The real secret to 70:20:10 is that structure and planning are at the heart of what makes it work. How you do that is the subject of the second part of this article.
Robin is senior partner at Learnworks. He is a writer, trainer and consultant helping global businesses develop people and improve performance

For further reading on 70:20:10 download Cross Knowledge’s free whitepaper: Effective Learning with 70:20:10 by Charles Jennings and Jerome Wargnier


charlesjennings's picture


I've read a couple of posts by you on the 70:20:10 framework over the past months - on the Infinity Learning site here, for example http://www.infinitylearning.co.uk/infinity_images/july/aNewBlend702010.html

It seems to me that you're suggesting a proposition and then knocking it down. That may be good sport, but not terribly helpful.

I've done quite a lot of work with the 70:20:10 framework. Yes, it's a framework or reference model not a recipe. I know of no-one who presents it as being anything more, or of anyone who suggests that the ratios are prescriptive. Of course they're not - it all depends on context.  

However if you read Jay Cross' 'Informal Learning' book, or Marcia Conner's work, or any number of other studies there are plenty of samples that show people learn most of what they need to do their job by doing their job and with the help of others.

The figures vary, of course, but it's clear that the Taylorist approach to workforce development that has been in place for the past 60  years at least - with its narrow focus on formal away-from-the-workplace training - is only a part of the story.

Equally, no-one in their right mind would say that formal learning, whatever the percentage, is not important. Of course it is. But all the evidence points to a 'horses for courses' approach. It's just that most learning that transfers into improved performance emerges from self-directed activities in the workplace, not through L&D-led, managed, controlled, designed processes.  

I appreciate that it may be a difficult pill to swallow for many instructional designers that they are not at the centre of the learning universe in their organisation. But that's the case.  We need to learn to learn to live with it, not put up spurious arguments and then knock them down.

Incidentally, the '70' and '20' in the framework are certainly not simply a 'new blend'.  If you really believe that I suggest you speak to some of the organisations have successfully implemented 70:20:10. To do so requires L&D developing a whole new set of skills and capabilities to enrol and engage line managers and line leaders to make sure workplace learning opportunities are maximised.

Another 'incidentally' - although Morgan McCall and his colleagues are credited with the initial work, the 70:20:10 split was originally identified back in the late 1960s by Alan Tough, now retired Emeritus Professor at the University of Toronto, in his work with 'adult learning projects'.  Alan's books are freely available online.

<off high-horse>

We've been using the 70:20:10 idea for the last year or so in our call centre learning & development plan. As you say, it's important not to discount the value that classroom learning brings, but what we do is try to get away from the "magic 7.5 hours" idea where somehow, every possible shortcoming in skills can be addressed in a single training day, and that when they finish that course, the individual is therefore Trained and somehow competent.

So we focus our efforts on coaching and group learning (shadowing, team "buzz" sessions, etc.) and endeavour to make sure that our leadership community (that includes everyone who manages anyone, not just senior leadership) is on board and aware that the learning in the classroom needs to be revisited and reinforced when the learner returns to their desk.

We'd love to get to the point where 70% of the L&D team's time is focused on those things - the reality is that we're not there yet but working on it.

robinhoyle's picture

..as I would expect Charles.  Thanks. 

You assert that no one would say that  'formal learning was unimportant' or indeed that informal learning needed planning, for that matter. Unfortunately I have seen a number of arguments put forward which do just this.  Usually they quote the 70:20:10 model just before they reduce the headcount in the L&D team.

However, in this pair of articles I try not to simply knock down 70:20:10 but outline how L&D teams can use it as part of their practice and can harness the power of well-planned workplace learning. 

But I get ahead of myself and you'll need to read article number 2 for that part.

robinhoyle's picture

I think the planned and appropriately resourced use of 70:20:10 is ideal for call centres and similar workplaces.  I hope you get some more ideas about how to embed the approach in the second part of this article.  Thanks for your comment.

People like something they can hold on to, and training professionals are people. For sure.

As you said, that's why we hang on to 70-20-10. Makes sense. And if it propels us towards more vivid, practical and active training, who is to say nay about it.

Your link to the recent work on informal learning is apt. Many are talking about it. But when Jarche and others urge movement towards  informal learning, and recognition of so much that we learn as informal, what are the implications for professionals who work in learning and performance?

Frank Nguyen and I took that on  in the January issue of T&D, ASTD's monthly magazine. First up, here is an article about formal and informal learning with three extended examples -- http://www.astd.org/TD/Archives/2012/Jan/Free/Jan12_Feature_Yin_and_Yang...

In related work, you will see a tool we developed to help people think about their choices and strategies: http://frankn.net/yinyang/

Hope this is useful for you and your colleagues.


Allison Rossett

robinhoyle's picture



Really useful.  Thanks very much for sharing



Tough's original  ratio, which was used in his dicussion only as an example of a program, is now commonly quoted by many as a 'golden rule' for learning and development. Sorry, no. Not without evidence.

I agree that learning will take place as Tough orginally described, but I have yet to see the definitive evidence, now trotted out at almost every L&D confence I have have attend in the last 5 years, that 70:20:10 is a "rule".  when I have asked the question "where do the numbers come from?" it has drawn heat, but no answer!

At the risk of preaching to many of the converted, research and analysis, to determine where the learning should best take place, is the key to L&D program design. The analsysis and the subsequent design should incorporate strategies that learning opportunities are developed and implemented in ways that ensure that transfer of learning to the workplace is effective - whether it starts in the classroom or on the job; the ratio will always vary program-to-program and individual-to-indvidual (that's what make being and Instructional Designer so much FUN!).

charlesjennings's picture


I don't think anyone here is arguing that data-driven evidence is not the best approach for validating models.

I mentioned in my comments above that anyone who takes the 70:20:10 ratios literally is either not in their right mind or has just emerged from a climate change denial cabal.

It's plainly absurd to expect that any research would find these exact ratios - we're dealing with human behaviour here.  Many decisions are taken on the basis of the Pareto principal - but only a fool would stand up and argue that the ratio must exactly fit 80:20 in order for Pareto to be accepted. Only someone who doesn't know their scientific deduction principles from their statistical analysis would adopt that stance.

Sometimes I feel that 70:20:10 must be seen as 'threatening' by some L&D professionals - out of our comfort zone, not sure how to apply the framework, take away our jobs, cut our budgets etc. etc.

Here's some data to chew on.  It's not a peer-reviewed paper, but it is a report of a survey of more than 200 managers that was carried out by ComRes and analysed by Peter Casebow and Owen Ferguson of GoodPractice in Edinburgh in 2010. This study found that by far the most frequent and effective learning activity carried out by managers is 'an informal chat with colleagues'. 82% of managers will consult a colleague at least once a month, and 83% say that it is very or fairly effective as a means of helping them perform their role when faced with an unfamiliar challenge.  The raw data is there and the other findings, too.


If you analyse this data will you uncover the magic 70:20:10 ratio?  Of course not, but you will see that the vast majority of learning in this sample is gained through experience, practice, and conversations -  non-formal learning - fitting into the '70' and '20' categories.

There are quite a few other studies that reveal the same patterns.

So, where does that leave us?  "researching and analysing where the learning should best take place being the key to L&D program design" in your words?  

That's part of the story, although I would argue that what you are describing is a Training Needs Analysis, one of the tools that has got L&D into a situation where the latest Corporate Leadership Council 'L&D Team Capability' survey (51 organisations, 350 line managers across the world) reports that only 23% of line managers/leaders agree that their L&D department provides an effective service, and only 14% would recommend to their colleagues that they engage with L&D to help them solve their problems (with 52% being 'active detractors' - i.e. 52% would actively discourage colleagues working with L&D). 

Analysing this data one would have to conclude that carrying out TNAs followed by 'program design' is clearly insufficient and maybe we should look at other models.

But the problem isn't due to taking a research-based analytical approach, it's due to the in-built mind-set of many L&D practitioners who think that  'program' or a course' or some form of 'blended intervention' holds the key to solving all performance problems. It's what I call the 'done to' mindset.  The idea that learning is something that gets 'done to' people by learning professionals.  Even your terminology reinforces this - 'program-by-program'.

A culture of real, continuous learning will only emerge in organisations that can get rid of the blinkered idea that learning happens best (and only) when managed by L&D professionals. 

Behind a lot of the squealing I hear from L&D people, and calls for data behind 70:20:10, I see practices going on that are so inefficient and ineffective that were they in any other part of the organisation the practitioners' time would have been called a long time ago.

Personally, I have found the 70:20:10 model useful.  Many others do, too. Until someone provides a better one (and TNAs are certainly not in theball park here)  am happy to use it as a reference model.  Key elements of Einstein's 'law' of relativity weren't proven until 2011 - some 95 years after he postulated them. In the intervening years a lot of science was successfully caried out using Einstein's theory without the definitive proof.  That seems a good enough precident for me.

<off high-horse again>

70-20-10-- a rather good direction to go, depending on the work, worker and workplace. Always.

The answer isn't any one thing, not even a wonderful thing like immersive simulations, such as the ones that airlines use. They are great. But in and of themselves, they are just part of the mix. When an emergency came up, it was team work that made the difference. A few well trained pilots worked together. They even solicited views from a "trainer" pilot who happened to be on board.

I agree with Charles Jennings above. Just look at what he said: ...survey of more than 200 managers that was carried out by ComRes and analysed by Peter Casebow and Owen Ferguson of GoodPractice in Edinburgh in 2010. This study found that by far the most frequent and effective learning activity carried out by managers is 'an informal chat with colleagues'. 82% of managers will consult a colleague at least once a month, and 83% say that it is very or fairly effective as a means of helping them perform their role when faced with an unfamiliar challenge.  The raw data is there and the other findings, too.


What's the advice? Certainly, it's not that the answer is always informal chats, although they sound like good practice to me. Add them to the mix. Tailor that mix based on analysis. Please check out some tools associated with my analysis book.

Congratulations on sparking the debate. Alfie Kohn (in his book Punished by Rewards) summed this whole thing up perfectly:

"There is a time to admire the grace and persuasive power of an influential idea, and a time to fear its hold over us.
The time to worry is when the idea is so widely shared that we no longer even notice it; when it is so deeply rooted that it feels to us like plain common sense.
At the point when objections are not answered any more because they are no longer even raised, we are not in control.

We do not have the idea. It has us."

-- Mark Wayland The Last 3 Feet

Back to top Back to top