Harvard Business Review may have dubbed being a data scientist the “sexiest job of the 21st century”, but that hasn’t necessarily made it easier to find analytics experts. There is still a huge skills gap in data science.
Recently I met with Andreas Vermeulen who is Director of Technology and Head of their Analytics and Automation and Digital Consultancy Services at consultancy company Sopra Steria. He is using a variety of approaches to filling the skills gap, including improving the skills of existing team members, rather than just recruitment.
“I am responsible within my team for our official education program. I'm sending people back to university to improve their skills, and we have noticed that this has knock-on effects. Individuals will go and learn something, and then share it with other people. This meant we naturally started developing strong mentoring relationships.”
Andreas Vermeulen
Working with universities
Other possible approaches to the skills gap include working with universities to provide opportunities for students to gain experience of working in industry. This is an approach that we have also taken at SAS with the SAS Academic Programs, which help students to gain skills and qualifications. Andreas agrees that this can work really well.
“I’m doing some research as part of a university team, but I’m also supervising students. I’m actively involved with at least three universities right now, and supervising students at both MSc and PhD level. Most of my students so far have been in data engineering or data science, but this year, I also had one doing operational research. It’s become a thing—but there are also benefits to the company. Sometimes the students can try stuff that’s not working for me, and we also provide ideas for projects, effectively sponsoring their research. We started doing it at Sopra really because we wanted to wipe out the whole digital skills shortage, but once they’ve got their degrees, it seems a natural step to think that they might consider working here too. It works both ways. They know who we are, and what it’s like to work here, and we know what we’re getting.”
Andreas comments that the work has so many benefits for the business that he has started to organise further collaborative work with universities on a more formal basis.
“We are getting a lot more collaborative. Over the last 18 months, people have started to look at solving common problems together. We go on quite a few hackathons and events like that. I am also trying to connect to more universities, so that we can help them to make good ideas into commercial solutions. We find we are having conversations with PhD students about an idea, and then eight weeks later, we might be deploying their idea as a new solution inside a bank. We also do collaborative work where we would bring university folk in to help with a problem.”
Filling the work experience gap
One of the biggest issues in recruiting is the gap between qualifications and experience. A university qualification is essential for a job in analytics—but it still doesn’t mean that you can do the job. Andreas agrees that qualifications are no match for experience.
“If you were to ask me where my team sits on a scale of 1 to 10 for digital readiness, I’d probably put it at about 6 or 6.5. I may be being a bit harsh, but I really feel like they need more experience. I’ve got lots of PhDs and MBAs in the team, so they have the skills, but they don’t have a feeling yet for when something’s not right. You can only get that through practice. I keep stressing to them that they need to talk to each other when things aren’t going right. Don’t keep it to yourself. Share the experience, so we can all learn from other people’s mistakes. The other aspect is that they can make things too complicated. Sometimes you just need a simple decision tree. They also want to make their models perfect, and there often isn’t time in business.”
Addressing the data literacy gap
Despite work by companies like SAS and Sopra Steria, there are still large skills shortages. Andreas notes, however, that these are not simply shortages of data scientists.
“We need more people who understand data. I've got a lot of clients who just can’t manage data. You talk to them about correlations, and they glaze over. We need to solve this problem as we start to be more data-driven, or we are going to leave a lot of people behind. As data scientists, we need to get better at explaining models.”
We’re making progress on working with academia at the high end of the skills bracket, were contributing to widening those data literacy skills with government agencies through the totally free STEP data literacy Programme, but still feel as an industry the practical application is where a gap exists.
