TL/DR
It’s time to end our reliance on unthinking buzzwords and mantras like “design for digital natives” or “gamification” as mental shorthand for innovation. Follow these buzzwords and they lead to myths. 70/20/10 is one such myth and it’s particularly harmful. Not only can we not measure this ratio in the first place, it’s a bad guide to planning L&D activity. We need to drop it - like other buzzwords - and start building our strategy around problems we can solve.
The L&D industry is badly afflicted by buzzwords and their close relations, mantras. Everywhere I see decision-makers looking to implement a buzzword (most recently, the “ecosystem”) and vendors dressing up the same old stuff in new buzzwords every year (ie, pretending your mess of legacy products which do not connect to other services is in fact an “ecosystem”). Like keywords on a CV or a dating app profile, we use buzzwords as a mental shorthand to avoid the need to think too hard. Sometimes this shorthand is benign. Sometimes it’s dangerous.
People like Lori Niles-Hofmann and Peter Riber have started to call the buzzword problem out and challenge it. They had a great game of jeering “buzzword bingo” going on at this year’s Learning Technology exhibition that made it slightly embarrassing for buyers and vendors to use a buzzword without thinking about it a bit first. Recently Sam Moat has written more seriously about the problems with the language we use in L&D and suggested we try to drop all that talk that distances us from real problems and real people.
In this post, I am taking aim at the sacred cow of all buzzwords, the most credible (and for that reason, to me, the most infamous) mantra of them all: 70/20/10.
The reason I’m so passionate about this is that many buzzwords (although not all) are based on total myths. And when something just isn’t true, to use it as a shorthand for innovation becomes very dangerous indeed.
Read Urban Myths About Learning and Education and you’ll learn that if your L&D strategy is mobile-first, aimed at digital natives, minimises formal instruction and is evaluated by Kirkpatrick (see Donald Clark on that here), you have constructed it on the basis of no objective evidence whatsoever.
To plan your L&D strategy based on folklore or your personal religious convictions would be just as effective. Given the greater cultural weight and history behind folklore and religion compared to something like Kirkpatrick, it might even be more effective. (It's worth noting here that personalization is one of the few buzzwords that is actually backed by evidence, such as from Duolingo and Filtered).
Like many (but not all) mantras the idea that we do 70% of our learning on-the-job, 20% through interactions with peers and 10% through formal learning is a myth. Let’s be clear. As Bruyckere, Hulshof and Kirschner demonstrate (see the chapter in the book mentioned above), there is no study that supports these figures. There has never been a situation in which people have self-reported - or it has otherwise been measured - that they learn things in these proportions. 70/20/10 is just as much bunkum as the "theory" of learning styles.
Those who promote 70:20:10 as an L&D strategy acknowledge that the numbers are not important or verifiable. The 702010 Institute puts it like this in a key manifesto document. “70:20:10 isn’t a ‘rule’. The 70:20:10 model simply describes learning as it naturally happens and then offers a means to accelerate and support that learning”.
Indeed, “Although the 70:20:10 model is primarily a change agent, the numbers serve as a useful reminder that most learning occurs in the workplace rather than in formal learning situations. It also stresses that learning is highly context dependent. However, don’t make the mistake of quoting the numbers as a mantra or as fixed percentages.”
This goes on for a whole page. The 702010 Institute repeatedly warns you in its 702010 primer not be mislead by the numbers. So why use the numbers throughout your whole document? Why use the mantra at all? Why name your institute after it? If it’s based on zero evidence - a completely abstract model - how can it possibly be a guide to action? I feel that this is irresponsible.
The underlying reason why there’s no evidence for the ratio is that you can’t measure the big part of it. Whilst you can measure how long people spend in activities we can define as learning (ie the “10”) you could never measure the informal, on-the-job learning in your organisation. It’s wrapped up into how people work. Learning a skill is inseparable from accomplishing things.
Of course, there are many honourable goals signalled by the notion of 70/20/10. Primarily, it’s a reminder to look seriously at whether what you do every day addresses the performance support arena or helps to construct communities of practice where so much vital learning happens. I think it’s a bad way to make the argument, but 70/20/10 is also a push towards resources instead of courses. Towards solving real problems, in other words.
But like with other buzzwords, it’s used and abused as a mental shorthand for innovation. And shorthand, by nature, is ignorant of the complex and nuanced problems a particular L&D professional may need to solve. I frequently encounter people who tell me they are implementing or have implemented 70/20/10. But what exactly have they implemented in the name of this so-called "model" and how could it be proven to succeed if the model itself is a phoney?
Even the emphasis on learning things informally itself reflects an even bigger myth about learning - the "discovery learning" myth. The notion that unguided or minimally guided "learning by doing" or "discovery learning" is the best way to learn how to do something new is a destructive falsehood. The research suggests it is a very inefficient way of learning most things. Learning a completely new skill with minimal guidance is crazy. But we often assume that minimal guidance, as well as being conveniently cheap, is more effective than training.
Setting aside that major issue, here’s how these three fictitious numbers harm L&D strategy:
The “10” denigrates what should be the learning function’s core activity
All learning professionals should play a role in creating quality learning experiences that help people to grow and providing resources that help people. The mantra cautions that this is only “10%” of learning and suggests we need to drastically reduce the amount of time we spend on it. But helping people learn things in a deliberate and contrived way is simply what L&D does. People need us to do it so they can do their jobs.
The current system of training we have may be wasteful, compliance-driven and misdirected at courses instead of resources. But we're not going to fix that problem by assuming formal training itself needs to go away. Besides, if the mantra was right and your people did spend 10% of their "learning time" engaged in some sort of formal learning process that would actually be a huge undertaking to support and optimise, not something to disregard.
The “20” creates more confusion than clarity
The mental shorthand of 70/20/10 also suggests that 20% of our time should be dedicated to running coaching and mentoring programmes because this is where 20% of learning happens. Now, this is indeed so important a measure that implementing a mentoring programme may be the single most effective thing you ever do in L&D. But it doesn't take a day a week of all your entire team's time every single week. We're talking about a big potential return on maybe 2-3% of your time overall. Here 70/20/10 actually makes it seem harder to do than it is.
Or, if we take the “20” to mean learning from others informally more widely, then we’re often describing a problem rather than a solution. A lot of the peer-to-peer learning that happens is not quality mentoring but trivial knowledge sharing which saps the productivity of those delivering the same piece of advice or technical knowledge again and again. So what benefit does allocating "learning from others" 20% in the 70/20/10 model, and prioritising it as such, actually offer? None at all.
The “70” encourages us to implement red-herring “total learning” systems
70/20/10 encourages us to dedicate huge efforts to support “on-the-job” learning. But since we cannot easily define or measure this concept, that focus quickly mutates into a more general idea of supporting informal learning - a term which often overlaps with social learning. For a number of years now, the result has been the deployment of learning solutions which are inspired by consumer social networking sites. The thinking tends to go like this: our learning platform needs to behave like a social network because most learning is informal. We are then surprised when people don’t use this new social learning platform. By now, any large organisation is strewn with the husks of such general-purpose social learning initiatives.
The reason is that your people already use systems to learn informally, collaborate with each other and search for information. Any informal learning is probably going to happen there (where it already happens) not on your new dedicated destination for social learning. Your people will persist in communicating with a varied mix of tools and - since “on-the-job” learning is inseparable from doing your job - that tool will most often be contingent on the task they are accomplishing. The truth is that in our solution-centric efforts to “encourage” on-the-job or informal learning, we actually end up requiring people to come off the job - and the “engagement” we can generate and measure there is often trivial in nature.
Enabling communication and workflow performance support with new technology is a critical conversation in every business and L&D should be a major player. For example, by encouraging the right culture of collaboration and a transparent growth mindset. But 70/20/10 leads people to grasp after that intangible 70% with new "total learning" systems or Orwellian tracking mechanisms which succeed only in killing collaboration off or forcing it elsewhere. Enterprise collaboration is not the exclusive property of L&D. You’re much better off encouraging better uptake of the successful collaboration processes already in place.
To conclude...
My aim here has been to show that the 70/20/10 dogma is so wrong-headed that even if it could objectively be proven to be valid - which it isn't and can't - implementing it would still create problems: you’re attacking the best way to create new skills by prioritising an informal learning process which is extremely resistant to forms of external control and reporting. Like all L&D mantras, no organisation will face exactly the typecast problem the mantra signals. Even if 70/20/10 were true of humans in aggregate, the model would not describe the complexities of your learning culture. We must always look at specific problems.
Here’s a little example: say I'm organising a conference and somehow I know that 70% of the delegate value came from informal, unstructured networking last year. Do I design my entire event around that feedback and devote 70% of my organisational effort and the session time on the day to unstructured networking? I could definitely put my feet up, but I'd be fundamentally abdicating from my responsibility to organise a good learning experience for everyone.
And that's what 70/20/10 leads us to: trying to ignore the problems we have the greatest responsibility for and assuming responsibility for a process of learning which, as Nick Shackleton-Jones puts it, is like breathing.
Communities need tending to, yes. But informal learning does not require a ventilator.
So let’s say we stop using 70/20/10 and all the other L&D mantras that seem to offer perfect statements of our challenges and the solution. How do we replace them? The answer is “I don’t know”. No one knows what specific problems L&D can solve and how until you try and find out. You might start with a project to try to establish what worries and problems most impact your people in their jobs and what kind of learning support is helping to address them. Just don’t expect that learning picture to look like 70/20/10.
Post-script: On 6 August 2018 I published a follow-up to this article addressing some of the themes that came up in the discussion about. You can read that follow-up here.