How to really make an impact in the age of skills

Years ago, I developed a simple matrix where I plotted the complexity of a topic on the x-axis and proficiency levels on the y-axis. I then started to map a number of typical learning interventions like reading a book/audiobook, animations, videos, scenario based learning, Communities of Practice and coaching. Later when I was at Philips, I expanded the matrix to include a 6 level extension of Blooms model and additional type of interventions outside L&D like ‘lead by example”, performance KPI’s, culture & rewards.

The key purpose of this matrix is to help select the best available method to meet your learning objective(s) based on 2 simple elements: (1) the level of complexity of the tasks/processes/activities in scope, and (2) the level of impact you want to achieve.

The key reason I developed the matrix was that I see way too many situations where the learning intervention does not match what either the level of complexity or level of impact requires. And often there is a mismatch in both!

Plotting the type of learning solutions against complexity of the topic and level of impact required can help selecting the a fit-for-purpose solution.

With Cornerstone onDemand acquiring Clustree and some excellent comments being made by Josh Bersin and others it seems the “war of the skills cloud” is full on. This is strengthened by a steady flow of articles stressing the need to upskill people, like from this years World Economic Forum on how to upskill a billion people. However when you look beyond the hype of technology/AI based skill identification, and beyond the immediate implications of that technology (which for me first and foremost benefits recruiters and other HR professionals rather than employees), I am getting really concerned about the direction of the conversations and solutions when it comes to how this will help employees and managers.

First, let me explain that I fully agree with Josh Bersin on many things. Yes, building extensive competency models is a thing of the past. I have seen many companies try; Some are still trying after 10 years, some have given up and some have developed competencies that are so generic and abstract that nobody can actually translate them to something tangible or workable. Secondly, the future of work is all about skills: having the right skills but even more important being able to build new skills quickly. Since many years I am a big fan of the youtube series “Did you know” (I use it a lot in presentations, workshops and training). Every single version of the video states that we simply do not know what skills we will need in the future. So it is only logical that we in HR start to become more skill driven. Not just for identifying the skills required right now, but more importantly, exploring possible key skills of the future and support employees to develop these skills.

That brings me to the third trend that I really like. The concept of skills is very ambiguous. Everybody has a different opinion of what skills means, let alone that we can agree on how we should label skills. Companies like Coursera and LinkedIn record and analyze thousands and even tens of thousands of skills to understand who has and who is developing what skills. So having the means to crowd source these skills is great. No HR department (however big they may be) will be able to manage and maintain these massive lists by hand.

So far so good. Now here are my challenges….

If we really talk about skills (according wikipedia, A skill is the ability to carry out a task with determined results), we miss 2 vital ingredients in the mix: (1) how do we assess skills and (2) how do we provide the right tools to people to enable them to develop these skills? Not just to increase the employees knowledge, but build real skills?

High complexity of key current and future skills

If you look at what skills are in high demand, you will typically encounter things like: Creativity, Persuasion, Collaboration, Adaptability, Time Management, People Management and Analytical Thinking as well as many technology skills like AI/ML, Programming, UX Design, blockchain etc.

Looking at these lists, you can hardly ignore the fact that many, if not all, of these skills can be considered complex and very complex. If you then take my little matrix of complexity vs proficiency level, you will realize that both the assessment of these complex skills as well as building them, requires serious attention and even more serious solutions. And personally I believe we are neither giving this the attention it requires nor are we building the solutions that upskilling needs.

We should not forget that from the perspective of the employee, all of these discussions on the future of work and future of skills can be somewhat overwhelming and challenging. As an employee, most likely you have 3 key questions: (1) what skills do I need to build my career? (2) what skills do I have now and how big is the gap with what I need? (3) what can I do build the skills I need to stay relevant (and continue having a job!).

Challenge #1: How do we assess complex skills?

While many skills models and analyst reports share what skills people require. There are few really useful descriptions on what these skills really mean. What is creativity? And how can I validate if I am creative? There are also few services and solutions available to help employees assess if they have these high demand and complex skills, especially online. Most online assessments are personality tests, not skill assessments. Sure, LinkedIn has launched a skill assessment earlier in 2019. It is a great tool (although non of my skills in my LinkedIn profile apparently can be assessed when I was writing this article) and it will have its use. But interestingly enough, the example they shared in the article is around an excel skill assessment. Now I am the last one to criticize excel. I am using it daily and think I have quite mastered it. I am also somewhat amazed that excel is the Nr 1 in the Coursera Global Skills Index.

However, assessing excel skills is very different (and much easier) from assessing if a person is skilled in creativity or adaptability. I assume we all can agree that using simple question lists will not do the trick when assessing any of these more complex skills. So why is it still the most prevailing solution out there? We must (and can) do much better than that, so here are some ideas and thoughts.

1. Peer based assessments.

What I like about LinkedIn endorsements is the idea of having peers endorse your skills. Rather than I decide what skills I have, people in my network can endorse my skills, which makes it somewhat of a more objective exercise. Even more, LinkedIn gives additional weight to any endorsements from people who themselves are regarded as experts in that skill area.

I believe the method of peer based skill assessments has merit and can be valuable, as long as we understand the risks: People might endorse others for ulterior motives. They might want to please a (potential) client, they might expect an endorsement in return. A personal example is that I have endorsements on skills from people with whom I have not demonstrated that specific skill while working together.

Despite LinkedIn’s approach to also include the level of expertise in peer endorsements, I still feel this to be an interesting, yet some what risky method if used in isolation.

2. Leader or expert assessment

The next step up from peer assessment (and this could be a big step!) is assessment by a person who is either in a leadership position, has more experience, and is more skilled, or is all of the above: a manager/leader, or a subject matter expert. Leadership development is a huge market and an estimated 95% of companies intent to increase the spend on leadership development programs. A key objective of any leadership development is to provide managers and leaders the skills (!) to assess their direct reports. Unfortunately, it seems that many of the leadership development program fail to make the desired impact, which is concerning if you have created a model that is based on personal assessment by managers.

Personally I believe more in the potential of skill assessments by subject matter experts, who are not only highly qualified in their domain, but also in assessment techniques. Most of the time experts are less biased because they are not involved in the day-to-day work of the assessed employee, and because of their high level of expertise on the skills they assess, sometimes even being formally certified, they are in a good position to perform high quality and reliable assessments.

But experts (especially in future skills) are expensive and rare. Do you prefer your top AI developers spending their valuable time building the latest and greatest cognitive AI algorithms, or performing skill assessments for your workforce?

3. Data based skill assessment

With more and more people working in a digital environment, we capture a lot of data as part of our everyday work. We can harvest that data to establish how skilled people are. Naturally skills that leave a clear digital footprint will be more ‘easy’ to assess. A nice example is how Microsoft is using outlook to assess you time management or networking capabilities. There is research being done even on the application of unobtrusive sensors in the office to detect work stress. These could be used to test your level of resilience for example. In short, more and more data becomes available which provides us more and more insights on the skills our employees have and have not.

For me this is what we should be working towards. It is using real performance data to the max, combining it with technology to analyse and advice the employee. Important to realize here is that in order to make this a success you must ensure (1) compliance with all relevant data privacy legislation, (2) the data is used for which it is intended, and (3) the consent of all people whom’s data you are storing and processing!

More on data and data privacy in a later post, but my personal preference is always that personal data (private and sensitive) should primary be used to support the employee. The aggregated data can then be used for trend analysis by leaders and HR.

ON THE “WAR FOR SKILLS”: It is interesting to see who will win the war for skills as most HR and L&D platforms now have their own skills and certification model. Degreed has skill certification, Cornerstone as said above now has Clustree, Workday has skills cloud, edcast has the skills framework, and not to forget LinkedIn and Coursera (who both publish skills lists for a few years now as they have a lot of data on skills in demand and skills learned and I expect they will start to use that data in all earnest very very soon). So which one should you choose? As with my earlier comments on ‘selecting the best LXP‘ I always advice on thinking strategically on such a selection. What is the strategy, what is the roadmap, how big is their dataset and how good is the AI? Are they willing to truly partner up? Can they help you with not just implementing the technology, but also the processes and in this case building your skill taxonomy? Or do they have (sufficient and high quality) partners to do so? In the context of skills there’s one crucial and decisive requirement additional question you should ask: How easy is it to ‘pick up and move’ skills data and credentials/certifications? From an enterprise perspective, you will want all that valuable skills data to be easily extracted to your analytics solution (or even a data lake if you have one) in order to run valuable analytics or, if you switch technology, to a new platform (which yes indeed could be a competitor!). As an employee, you want to be able to pick up and take all your skills data to a next employer, and coming to you next employer, you will want to simply add it to your personal (skill) portfolio! Whomever will make this work, will stand a good chance of setting the industry standard!

Challenge #2: How do we facilitate skill development?

Whenever I join an L&D team and look at the data in their learning systems, it always amazes me that we actually not seem to be very good at capturing what skill is covered by the training at what proficiency level. Even worse, we state rather frequently that a training promises to teach us a skill and then turns out to be a page turner with voice over and some video. For me this is a key reason why every L&D department should have a data strategy: we need to define exactly what data we want to capture to make sure we accurately define training in terms of content, skill(s) addressed and proficiency levels achieved. I actually think that you need to take data into the heart of your design to ensure that you develop programs and activities that are modular enough in terms of skills and proficiency levels to allow them to be curated into personalized learning journeys. This to prevent an already experienced project manager to be forced to first go through basic project management topics and activities (that he/she already knows and is familiar with) whenever taking an expert level training on project management. I’ve started to refer to this as data driven addie, or DADDIE, and will come back on this soon.

In the context of skill development we must be really think through how we can best facilitate: What methods are both efficient and effective? The good news is that we have many examples and solutions available already. My diagram from the top of this article can help to create clarity and structure.

1. Gamified learning

All successful games in history have something in common: they found the right balance between challenges and rewards that makes you want to learn and improve. Whether it is the ability to throw and excellent curve ball in Pokemon Go (my kids’ favorite!) or build megacities in SimCity (my old favorite), in order to progress in the game, overcome the challenges and reap the amazing rewards….you will need to develop and improve certain skills. Now the nice thing about games is that you have infinite attempts (if time permits) to practice over and over again, and make many mistakes.

Game design principles (the art of game design – my favorite book on the topic – it’s a classic and a must read!) like this can be applied to learn new professional skills as well. A pity that most commercial games learn us skills that are not necessarily useful in the workplace, but there are many really excellent examples of tolls and apps that seem to have mastered this: Duolingo for when you want to learn a new language skill, the Khanacademy that started with Algebra and now has many more subjects (although not all of these allow you to learn a new skill), and codeacademy for coding and data analysis skills are my favorite examples.

Not all gamified learning programs need to be in the form of a full blown serious game. Developing a really good serious game requires a lot of time and investment and is something you would only do to teach large groups very critical skills. A less complex way of bringing in some of the game elements is for example with scenario based learning, or more simple puzzle-like games.

2. Augmented and virtual reality

I once missed out on an opportunity to work as a training manager for a high tech company that makes extremely complex machines and faced the challenge on how to train their customers on operating these machines without any loss of time. It would have been a great opportunity to go full on with virtual reality learning! However there is not much research available on the impact of virtual reality training. While research does suggest that VR training is more impactful than learning from a computer screen, the real question is if VR can be good enough to replace skills development with real people, machines and/or situations. Perhaps the lessons we have learned in using flight simulators to train pilots are the most significant: is shows (already in 2010!) that while not being able to fully replace the actual thing, flight simulators do dramatically reduce training costs. With big data and AI, the potential of AR/VR to develop complex skills is even bigger looking at what Samsung is doing with their NEON product.

3. Non typical learning activities

I believe there are a lot of possibilities for employees to develop their skills that are outside the traditional scope of L&D. Stretching assignments, temporary assignments, public speaking and writing and having a manager that is truly supportive in skill development and takes a personal interest in developing the skills of their subordinates. No doubt that much of this is already happening. Just not in a way that allows us to capture the sometimes amazing results that these methods can achieve in a way that allows easy integration with our training based data and analytics.

4. Analytics and AI

At the top right of my matrix you will find analytics and AI. It was not there when I first created the model 6 years ago. But now that we live in such a complex world, and we work in even more complex environments, it will be a matter of time before we will need the help of an AI based coach. In a strategy paper on learning analytics I’m preparing for a client, I’ve introduced a value chain that starts with making decisions based on intuition (basically our personal capacity to process data and derive conclusions from it) and ends with cognitive AI. Cognitive AI is the type of artificial intelligence that aims to develop AI that has the human ability for cognition and has the ability to “perceive, understand, correlate, learn, teach, reason, and solve problems faster than existing AI solutions”. The example I use is a fully personalized assistance that helps you develop skills: “Todays presentation was much better as I noticed the attention span of the audience was up 75%. Your reduced elevated heart rate and stable eye movements showed that you were much more relaxed and confident. However, you did struggle with explaining the concept of cognitive AI so let’s practice a bit more over the coming weeks”


The attention to skill based HR is great and much needed. However if we want to take building skills in our organization serious we need much more than technology. Doing this right is not just a matter of buying new technology and 3rd part training catalogs. It is much more complicated than many technology providers wants us to believe. (1) we need to have a clear strategy and approach on assessing current skills that is efficient, effective and easy to ‘pick up and move’, (2) we need to use the right instructional design methodologies for supporting skill development, not just rely on video and simple online training, (3) we need to look beyond the traditional scope of L&D and (4) we need to prepare for AI.

You will never learn a new skill by just watching videos

Leave a Reply

Your email address will not be published. Required fields are marked *