There are not many conversations, presentations or media briefings about analytics that don’t at some point touch on the issue of ‘people and skills’. Convincing business leaders of the merits of analytics given the technology now available is often the easy part. It inevitably leads on to what the barriers are to analytics deployment. And the first of these is normally skills, especially given it’s a global problem.
Ever since the role of ‘data scientist’ was described as “The Sexiest Job of the 21st Century” back in 2012, not much has changed in terms of the demand outstripping supply. The term ‘data scientist’ is often used as a bit of a catch-all for anyone working with data, so it’s worth looking a bit more closely at what some of the roles are.
We spoke to SAS customers in a range of industries to get their thoughts, and this blog sets out some key points arising from those discussions when it comes to tackling the skills issue.
What are the different roles?
Here are a few job descriptions for those working with data and analytics.
- Data Analyst: There are different types of analyst. Might be database focused, and range from SQL programmer through to a much more senior person that understands business problems
- Data Steward: Quite skilled in data engineering and structures and how to handle them; works in IT but embedded often in business, or data science or analytics
- Information Steward: Like Data Steward but more focused on the business side
- Data Scientist: Can be blurred definitions of what this is. After being described as the “sexiest job” people are conscious it’s good for their own marketability to say they are or have been a data scientist. Some may not even be or have been a data scientist in the strictest sense of this role, where technical skills in areas like statistics are required.
- Data Democratiser: Can bridge business, IT and data. There is usually a need for someone embedded in business who can understand data and use it to tell a story to non-technical people.
- Data Strategist: Looking at new data sources and challenging business to make best use of data. Aim to stimulate ideas and use cases.
The key is how to bridge the gap between IT, business and analytics and somehow find people who work across or within these areas. There are very few individuals who can straddle all three effectively, so it normally requires the establishment of data science teams where people have complementary skills.
The key goal for one customer was having ‘a thought partnership’ where it was possible to ask stimulating questions of the business, and then take problem and put it into questions that the analyst can understand and then investigate.
What are the challenges?
An important one is how to attract talent because there’s relatively small pool of individuals and lots of businesses trying to recruit them.
Applicants therefore have lots of choice – they can go and work for companies like Google or Amazon, that sound more attractive than a traditional telco or bank. It means that to attract them to work for your organisation you may need to give them something new or exciting to work on. However, often they will start in roles where expectations get shattered as they’re doing lots of work that seems less interesting than what was promised. But Google or Amazon also need people to do mundane work as well as the more exciting things, so this is a challenge for every organisation.
There was agreement that young data scientists often have inflated expectations. One customer said that there is often a risk of new recruits leaving after just six months if they feel the work they’re doing is not new or ground-breaking. But the reality is businesses need help with day-to-day needs and are not always able to supply lots of interesting new challenges.
Young data scientists often prefer certain codes or languages (and SAS is perceived by some as a less exciting tool to use). Again, the business reality is often not fully appreciated. It was pointed out that when customers have allowed employees to build solutions using new open source technology and then leave, there is sometimes no one left who knows how it works. Established businesses want reliability and some data scientists might be better off working with a start-up if they only want to be pioneering and build bespoke solutions using open source technology.
Attracting people who can work across business and IT often comes down to the structure of the organisation. How do you embed analytics in the business? You need to find people with skills, but then it’s about retention and career progression. You can’t just find people with these capabilities, so an important part is training and developing skills in the areas where they’re needed.
As the market is so competitive, career progression is wanted in a matter of months or just a few years. Sometimes the only progress that can be made is to a manager role, but the employee might have few people to manage or not be suited to people management. One progression that could be made is for an expert to develop into the ‘data strategist’ role described above.
Regarding training and development, you need to work on the hard, soft and business skills. One option is to provide them with rotation across the business, and look at cross-sharing skills between different employees so they learn from each other. That can help create the analyst that is a data democratiser, that can bridge the gap between data analytical and business skills.
What are the incentives to retain employees?
One of the retention tactics was showing the analytics team the value of their output, and giving credit to their role in achieving this.
Sometimes, analytical teams must regularly report what they cost to the business, what was delivered and the value of what was done. If results are positive than can be motivating. Otherwise it’s important to give them a chance to re-align how they work. Keeping people integrated in the business is crucial.
It’s important to know what good looks like. Good is having this ‘thought partnership’ - the business comes to the data science team with a problem and trusts them to solve that problem. Or the analytics team might go to the businesses in a proactive way after identifying a problem in the first place.
Data scientists should see themselves as a part of a continuous business system, and not as a separate and isolated unit.