Productivity, skills and intelligent learning recommendations

Marc Zao-Sanders Jun 15, 2017

The Learning Universe.jpg


The system supporting skills and careers development is inept and that costs the world trillions. But the future for learning is bright: higher quality, justified, data-driven, technology-enabled, culturally embedded, more of it. That future arrives sooner with useful, relevant, high-quality, fresh, intelligent learning recommendations.

1. Today

We suck at developing the skills of our workforce. After decades of debate, the skills gap is going nowhere: by 2020 in advanced economies there will be 95 million workers without the education or skills that employers need. In the UK and US, we struggle with basic skills such as literacy and numeracy. Staff spend less than 5 minutes a day (1% of their time) training.

This is a monstrous problem for global productivity. There are 230 million knowledge workers in the world (17m UK, 63m US). In developed economies, they represent half the workforce. Assume they earn $40k on average (in the UK it’s £35k). That’s a $9.2 trillion wage bill. This massive group spends 28% of their time emailing, 19% gathering information and 14% in meetings – collectively that’s 61% of their working life. 

230m knowledge workers image.png

Take a second to comprehend: this is what half the workforce of the developed world spend most of their time doing. Are we great at it? What if we were 10% better at writing succinctly and persuasively, systematically avoiding interruptions, Boolean search, evaluating information sources, listening attentively, minimising meeting attendee lists. We’re talking about a half trillion dollar global uplift (which would dwarf the planet’s $140bn L&D market).

It’s also a problem we feel on a personal level. Do you know what to learn next? Do you have a good understanding of what it will do for you? How to benefit from it, enjoy it, retain it, apply it? Will it be recognised externally? Will it be recognised, even by you? How often do you come across insights that transform how you work or solve a current problem? How many insights are lying out there, just waiting to be discovered? Do you feel you have enough time to learn? Most people can’t answer many of these questions in the affirmative.

There’s a vast, overwhelming panorama (Elliott Masie’s phrase) of content. There are several million business books (just listed on Amazon – I just checked), 8,000 MOOCs, hundreds of thousands of elearning courses, 140,000 courses on Udemy, 3,000 TED Talks…the whole of YouTube. It’s impossible for us to extract what we need from all that. As Donald Taylor puts it, whatever the requirement 'someone else has probably already created it' (Taylor, Learning Technologies in the Workplace, 2017). Here’s an infographic we created recently to illustrate how much learning content there is: 

Learning Panorama Infographic Final.png

Here’s a different visualisation of this absurdity: there’s a gaping chasm of a skills gap to our left and an enormous mountain of content to our right. We throw a few rocks from the mountain into the chasm, that’s all.

There’s no widely accepted credit system for corporate learning. Badges and CPD / CPE are still not an effective currency. Careers are longer and more varied than ever yet the dominant qualification is still the university degree, taken, for most people, before they’ve even chosen a career.

In corporate training specifically, there’s little innovation. LMSs are mostly built on a technology from well over a decade ago. Most of the learning done there is compliance, and detested. Industry buzzwords – gamification, personalization, social, 702010, Ebbinghaus (19th Century research!), MOOCs, microlearning, learning styles, VR/AR, business alignment, spaced repetition – are adopted unquestioningly by some in the industry and dismissed too quickly by others. Few companies have a deep, genuine learning culture: training is not a boardroom priority; L&D professionals lack the influence they should have; in hard times, L&D budgets are amongst the first to go.

Some of the reasons for this are clear. Responsibility for career development knocks on many doors but is answered properly by none; individuals, employers, professional bodies, governments all have a role to play, but the role’s not as clear as it is at school, college and university. The market is flooded with content because with digital there are virtually no barriers to entry, especially in the consumer market. And the benefits of learning (other than enjoyment and passing tests) are over the long term and therefore difficult to measure and prove. A corollary of this is that the luxury of learning is nearly always trumped by the urgency of work.

On the other hand, everyone wants the right skills (and behaviours, values, attitudes, etc). CEOs constantly complain about shortages in (insert any of: digital, employability, literacy, numeracy, soft, hard, STEM, vocational) skills. UK and US productivity levels are low, and skills are a key driver of productivity (along with enterprise, competition, investment and innovation). The half-life of a skill is the shortest it’s ever been in this digitising gig economy with increasingly meandering career paths; it’s now more important than ever to improve them frequently and efficiently. And certain skills stand out. For example, the rise of technology and automation makes the human aspects of brain capability even more important so complex problem solving will be in even more in demand.

Who wants skills? Everyone. But collectively we suck at cultivating them.


2. Suppose…

But things may be a lot better soon.

Suppose we moved from trying to prove the causal chain of learning-skills-productivity to assuming it. Learning, at the start of this chain, immediately becomes more important. The logic and levels of Kirkpatrick and Phillips are right, but in the real world they are difficult to implement and demonstrate comprehensively. Suppose this assumption catalysed a culture of professional learning beyond the firm. So yes, in firms, from trainee to CEO. But also amongst policy makers, undergraduates, schoolkids and teachers. Not just learning organisations. A learning civilization. What are we human beings, if not that? As John Holt says, 'Living is learning. It is impossible to be alive and conscious (and some would say unconscious) without constantly learning things' (Holt, Learning All The Time, 1989).

Imagine the quality of learning were much higher. That content were very good as a minimum, and often scintillating. Suppose we just used the top 5%, that best content rose to the top, poor content was winnowed out, the bar only went up. That’s still plenty – many, many thousands of assets. And the quality of the top 5% is very, very good (TED, School of Life, FutureLearn, Harvard Business Review, Pluralsight, etc).

Suppose we made content relevant. Meaningfully and usefully personalized. Learners get the learning they want and need, when they want and need it. Content, technology and data combine to enable us to make better choices in many walks of life (Trip Advisor, Twitter, Spotify, Netflix, GoodReads - see graphic below) – why not in learning? More tangibly, this means getting the learning closer to work: infused in workflows, linked to appraisals, spaced appropriately into inboxes, relating to current user circumstance.

Imagine that content also felt fresh. Drawn from a diverse, creatively curated pool of high-calibre content, it could surprise and delight. Maybe you don’t yet know that a basic awareness of confirmation bias would help you make significantly better decisions and listen to colleagues more empathetically for the rest of your career. Maybe you don’t yet know that getting to grips with $ signs (absolute references) in Excel takes seconds and that that investment will repay you thousands of times over. Maybe you don’t yet know that you’re a 3-minute video away from adopting your own hand-drawn visuals into your business thinking and communications and making a much bigger, human impact. In the near future, you will.

Suppose we really understood the relationship between roles, skills and learning. Roles reveal skills needs. Appraisals reveal skills gaps. Learning helps to fill them. Imagine this is measured and calibrated. Appraisals could also reveal hitherto unknown areas of skills abundance. Workers would find it easier to acquire the right new skills as well as to apply existing skills. Suppose that as technology, society and economies change, available skills can be repurposed and augmented to meet new challenges effectively - to the benefit of employee, employer and economy. Imagine we all had a clear, career-oriented, development path of sumptuous content. Assess skills needs. Assess skills levels. Fill skills gaps with the right learning. Tie back to tangible benefits (remuneration, promotion, qualifications).

Suppose we learned more. For much more than the current 1% of our working life. Suppose we understood and harnessed curiosity better. That learning became an enjoyable, positive habit. Suppose we understood the mix of learning that works best for each of us: in-person, synchronous, online, video, written, interactive, reflective, on-the-job. Suppose that mix were accessible for everyone. Suppose it were delivered wherever and whenever you need to see it.

Suppose we better understood our minds and brains. As we come to control more and more of the physical and digital world through human intellect (increased automation), we’ll need to understand that intellect better. If we all had a better handle on neuroscience, heuristics, mindfulness, mood, learning preferences, fulfillment, happiness, habit, addiction, distraction, bias, curiosity and how to apply learning, we would make better decisions for ourselves and for others.

Imagine that we anticipated new technologies free from the hysteria and hype. That we each got to a considered, sensible position. For example, there is a lot of hype about AI. Clearly, the onset of artificial intelligence will affect job roles. Job roles vary considerably. Some are better suited to automation and AI than others. So aspects of those roles will start to be carried out by AI over the coming years. The human’s role will then evolve into something more…human. Very few jobs will be consumed in their entirety by AI, still fewer (maybe none) by actual physical robots.

There’s plenty of evidence that much of the above is starting to happen. Some of what I’ve said here has been argued for by the likes of Bersin and Masie. Technological advances will drive a lot of the change. Our learning industry has been slower than most to adopt consumer-grade design, machine learning, immersive reality, etc, but it’s happening now. The LMS – at least in their current form – will continue to fade.

The consumer-employee now expects consumer-grade experiences at work and will get them soon.


3. Our role: intelligent learning recommendations

Filtered makes intelligent learning recommendations. We do this to enhance the skills and productivity of the world’s knowledge workers. We do this because in a world stuffed with people, things and information, pointing has become more useful than producing. So we are to learning what Netflix, Amazon, Airbnb, Spotify and Google are to films, things, accommodation, songs and webpages respectively.

 Recommendations graphic.png

Recommendations can’t solve all problems but they are special.

Recommending is an inherently human behaviour. We make countless recommendations every day: have you thought about asking James what he thinks about this strategy – he wrote a whitepaper about it last year; try right-clicking; it’s worth checking those numbers against this source; have you seen how this website solved that problem?; how about this space for that event you’re running? Or, closer to learning: have you read this article on digital transformation in the public sector?; since you’re so keen on getting that promotion, how about working towards this qualification?; watch this short video on compassionate leadership please, boss. These range from in the moment, off-the-cuff through to more formal interactions such as appraisals.

Recommendations usher in a content meritocracy. Poor content doesn’t make it. Average content gets limited exposure. But outstanding assets systematically find the individuals that need them. Intelligent learning recommendations galvanise the existing ecosystem rather than try to rebuild it.

We think we’re in a good position to help galvanise. We’ve been unraveling the link between roles, skills and learning for almost a decade. We’ve created thousands of learning modules and served them intelligently to almost a million users. We recently won a UK government grant to build the IP behind our new recommendation engine, globalfilter. Our arsenal includes machine learning, collaborative filtering, natural language processing, learning lexicons, semantic nets, consumer-grade and high-discoverability design, a deepening understanding of learning and a genuine passion for it, our own first-hand experiences of learning. We are a team of 35 based in Shoreditch, London, focused on this problem.

Lifting the skills and productivity of the world’s 230m knowledge workers is a lot of work. But the prize is huge. The companies that intelligently develop their workforces with the right skills in the today’s digital, algorithmic, world will thrive. We aim to be one of those and to catalyse that transition. If some of the above is of interest to you and you’d like to know more about what we’re doing and thinking, please be in touch ( Customers, clients, investors, partners, competitors, government, experts, future colleagues, future friends…all welcome.


José, Don, Team, thanks for the nudges.

About Marc Zao-Sanders
Marc started his career in strategy. He then applied the skills learnt there to a number of small businesses including Pure Potential and Over the course of this period he began to realise the shortage of basic business skills in the work place and wanted to do something about it. And so the idea of Filtered was born. Marc is now Filtered's managing director.
Read more by Marc Zao-Sanders