A few weeks ago I had the pleasure of sharing my thoughts and experiences on data driven L&D at a Coursera Breakfast event in Amsterdam at the Amsterdam Business School. There was a nice introduction by Coursera on their latest global skills index. Professor Peter van Baalen, the Dean of Digital Learning and Professor in Information Systems of the Amsterdam Business School reminded us again of Blooms 2 sigma problem and that the promise of digital learning is to achieve the impact close to that of 1-1 tutoring without the huge costs associated with doing this on a large (corporate) scale in the traditional face-to-face format.
During the final presentation I shared my vision on how the digital transformation of companies and learning opens up huge opportunities to harvest the wealth of data generated through digital upskilling, what I consider to be a data driven L&D organization and how to build one. I also showed some of the examples of work done at Shell and Philips as well as what we are doing at Novartis with their CLO around data and analytics to support their ‘go big on learning‘ initiative.
Whenever I share my thoughts, I always try to emphasize the enormous opportunity I see to systematically and strategically use data to improve L&D and demonstrate business value. When the most successful companies in the world make data and analytics the heart of their business strategy, I always wonder why we do not do the same in our L&D strategy. After the sessions, we had some good conversations on digital and data, and as all participants had an interest in the topic, the conversations were not much about the why, rather the when.
What took me a little bit by surprise, was that most people felt their organization was not ready to start the data and analytics journey.
I personally do not think you should wait working on a data and analytics strategy for L&D until you are ready. Things are moving so fast that you might never ever be “ready”. But every day you postpone, is a day where you miss harvesting the value of data. On top of that, strange as it might sound, data and analytics are actually very nice topics to ‘just do it’ and adopt a very agile approach where you quickly launch a Minimal Viable Product and build a process that allows continuous improvement over time as people start working with the data and think about what stories they can tell with the data and what additional data they would like to see. Both at Shell and Philips, we did not go through an extensive requirement analysis as we knew that people would struggle finding the right requirements at the start of the work and we would most likely end up with too many requirements to handle. Even if we would build them all, a large portion of them would not be used. By bringing together the knowledge of the business, the learning processes, the LMS and how it handles data, as well as some simple data modelling techniques, we were able to design and develop powerful dashboards that generated a lot of insights nobody had access to before.
Examples of early data insights
– Course completion trends; where in the organization (by location and organization hierarchy) are most and least courses completed?
– Course utilization: what courses do people start but not finish?
– Training hours: what is the average number of hours spend on training per employee?
– Popular topics: What is the top 10 most popular courses (excluding mandatory training)
We took the same approach with data (quality) management; starting with some obvious checks we wanted to do on a regular basis and develop data quality reports on those items.
At first it can feel quite harsh to just start creating reports and insights on both training consumption and data quality. It creates a lot of transparency on not just what goes well, but also what goes wrong and where: Underutilized classes, expensive e-learning modules that are only completed by a handful of people, missing information like training hours that is crucial to measure the time people spend on formal training. Thankfully it also shows what goes really well! But when you only talk about data management and reporting, it’s very hard for people to image what it really means, or what it will look like. When you start developing some initial simple standards and reports and share these on a regular basis, people will become interested and take notice. It will make the ‘vague’ world of data and analytics much more real and tangible.
One tip: Time and time again it shows that people are very competitive, so the one advise I can give on getting people’s attention is to create internal benchmark data. Having simple comparisons of data points between the teams will do wonders….nobody wants to be the team with the lowest scores!