Managing Clevers: How to retain your data science team
18th July 2018
By Steve Kilpatrick
Co-Founder & Director
Artificial Intelligence & Machine Learning
By 2020, IBM predicts that, in the US alone, the number of jobs for data professionals will rise 364,000 openings to 2,720,000. In the first half of 2016, in the UK, the number of data science roles increased by 32%. A quick Google search shows over 10,000 data science roles being advertised on Indeed alone (Jan 2018). Because talented data scientists are in such high demand, they have options. They could leave you for a more nurturing competitor the moment they become dissatisfied with their work environment.
A poll by KDNuggets found that half of all data scientists surveyed stayed in their previous role for just two years. IT Pro Portal identified the scientists’ main complaint being that they are not being tested to their full potential. In short, retaining valuable talent is a skill in itself. The key is to understand the unique ways to engage with the most talented individuals in your organisation, to know what they thrive on and what factors are likely to drive them away.
In their excellent paper, Leading Clever People, Rob Goffee and Gareth Jones identify seven principal characteristics of the ‘Clever’ that make them a different breed of an employee than any other. Let’s take a quick look at those characteristics, and explore how those features play out in the working environment:
1. Their cleverness is central to their identity
This doesn’t necessarily mean that your Clever is an egomaniac (though it certainly doesn’t exclude the possibility). What it does say, however, is that they are acutely aware of how valuable they are, and that sense of their value is what motivates them to do well and to improve continually.
2. They are indispensable to the organisation
Knowing, as we do, how high the demand is for data scientists, and the high levels of attrition in the role, it goes without saying that losing a good data scientist can be devastating for a company. Don’t let yourself be a victim of not knowing what you’ve got until it’s gone. They know what you’ve got in them; undervalue that at your peril.
3. They know their worth
As with the two above points, it’s fairly evident to any data scientist worth their salt that they hold powerful leverage. They command an excellent salary and deliver sufficient value in return. They know you need them more than they need you (there are always more data science roles out there desperate for a Clever like them), so keeping them engaged, challenged, and satisfied through their work and the management culture of your company could not be more critical.
4. They challenge and ask questions
Asking questions and challenging assumptions is what a data scientist does. They will ask questions until they’re blue in the face, even items you may consider very straightforward and easy to answer. This is because they want to be sure that they know everything there is to know before starting a project. They will challenge you to cross-examine assumptions and unearthing the root truth. In essence, they are extracting and analysing data from you! Rather than being annoyed or affronted by the questions and challenges, see them as a symptom of a competent data scientist.
5. They loathe the idea of being managed
Perhaps because they know they are so smart, or maybe because they just don’t want to be interrupted from their (to them) critical work, anything that looks like management will annoy the hell out of them. Because their work is crystal clear to them but makes no sense to you, demands from management will be seen as an inconvenience and a display of your ignorance. If you want respect from a Clever, let them come to you. And come to you they will…
6. When they need your attention, they need it now
Data scientists will spend hours, days, weeks buried in their work, barely coming up for air. But when they hit a roadblock and need something from you, be ready to get it for them. Admittedly, this is not always feasible, but at least receive the message loud and clear, and take the appropriate steps as soon as possible. Don’t forget how valuable your data scientist is.
7. They need to work with other Clevers
Data science has lots of different areas, and the fabled ‘unicorn’ who is master of all is a rare thing indeed. You need, therefore, a team of gifted data scientists, each of whom will bring their own unique specialism to the team. A good data scientist is curious, creative, and a communicator who thrives on learning. Placing them among others of their kind expands the knowledge and skill set of the entire team, whereas a solo Clever in a world of average Joes will wilt and – worse – leave.
So, those are Goffee and Jones’ Clevers characteristics, but there is more you need to know. There’s a set of rules you must follow to lead these valuable employees effectively, and ultimately to keep them with you and satisfied in their role.
1. Smart data scientists are creative
They will not thrive in an environment where bureaucracy and corporate protocols bind them. The Clever needs the freedom to create and to innovate to do their best work. Strict rules amount to the same as management, at which the Clever baulks. Administrative tasks are a waste of your Clevers’ time.
2. Make sure that they have their preferred tools for the job
If they’re forced to work with systems that are not best suited to the way they work, they will not only fail to produce their best work; they will become frustrated and dissatisfied with working for you. There are a range of tools that a data scientist uses, and each of these tends to have its own specific merits – making one tool better for a particular job, and a different tool suitable for another. A good data scientist works with a multitude of different platforms, and the output they produce for you will depend very much on what they are permitted to use. Allowing your data scientist free reign with what tools they use to complete their tasks will yield more than better data analysis, it will make your data scientist a more fulfilled employee.
3. Protect them from having to do mundane tasks
Wherever there is the opportunity to automate something time-consuming and onerous, let them do it. If there is a junior member of staff who could do a simple thing, don’t expect your data scientist to do it. This will undermine their value and steer them away from the valuable, complex tasks for which their unique skills are best suited.
4. Allow failure
All great innovation and groundbreaking work have pitfalls along the way. The failure of one idea is no reflection on the Clever’s ability, and should not be treated as such. Failure is a step in the process to success. A good data scientist knows this well. Data science requires experimentation, and not all experiments yield the results expected. What failure of an experiment teaches, however, can be applied to future experiments, generating positive outcomes. The data scientist is a patient breed and requires your patience in return. Rest assured, that patience will pay off.
5. Allow them the freedom to pursue their own pet projects and private efforts
The former will help them uncover new strategies that can apply to further the expert work they do within the organisation. The latter may be a route to the generation of new business opportunities for your company.
6. Make your expertise known, but don’t use it as a competitive tool
Never tell a Clever about your own skill – they won’t care. Demonstrate it, showing them how your Cleverness complements their own. This is the way to establish credibility with them; by meeting them on their level. As above, hierarchy means nothing – collaboration means more.
Management of any team requires a honed approach with a keen understanding of what makes different individuals ‘tick’. When it comes to the Clever, particularly in the form of an exceptional data scientist, you’re dealing with talent that can either be the making of your organisation, or a disruptive and destructive force within it. In short, to keep a Clever, you must be clever. For a thorough insight into effectively leading and retaining Clevers, I can highly recommend reading Goffee and Jones’ paper in full.
Looking to hire Data Scientists? You’ll want to read these articles: