CHRO Day 2023: HR leaders learn how to recruit – and retain – the right data team


Attendees hear why CHROs need a multidisciplinary approach in the age of AI.

Organisations need to leverage data-driven insights to optimise HR operations, improve employee engagement, and drive performance. However, it’s not an easy feat to find data teams that add real value. Kosta Kontos, a lecturer on the Data Science Leadership for Executives course at the Graduate School of Business (GSB) at the University of Cape Town, shared with CHRO Day attendees how to recruit, structure, and retain a high-performance data team.

According to Kosta, technical courses on data science are ubiquitous, but courses aimed at executives who want to lead a data-driven organisation are hard to find. There is a clear gap in the market, and the two-day GSB course teaches executives like CHROs how to lead the implementation of data science solutions so they're aligned with their organisation’s strategy.

Start with ‘low-hanging fruit’

He emphasised that this process is not easy, so a multidisciplinary approach is needed to find and retain the right data team. Through a 10-step framework and four pillars (data, people, tech, and maturity), Kosta advised attendees they start with “low-hanging fruit” – to pick something they know will succeed – because exco buy-in is critical from the start. It’s also important to design a strategy that creates value and that everyone can access.

For his presentation, Kosta focused on the ‘people’ pillar. He said that recruiting the right team often starts with the data scientist as the first hire, as these people can combine technical disciplines like machine learning, model training (feeding the data) and automation with reporting and analytics, and they understand the broader picture. Other considerations include what technology and infrastructure will be used (tech), how results will be measured, and how scaling will take place (maturity).

Hire the right data team, from job-spec to wrap-up

For hiring a data scientist, you need a clear job spec. Kosta shared his advice on how best to do this.
First, describe the data science role and identify what opportunities you want to start with. For this field, you need people who have a bit of experience. List the tools required on the job spec – but keep in mind that the person you hire doesn’t have to do it all.
A rule of thumb is to prioritise someone who is proficient with one tool, but can learn to use more on the job. Kostas believes the salary should be stated upfront because data science is competitive (and data scientists are usually paid well).

Kosta said a first interview should be quick, so you can “spot the nonsense” and make sure there’s a data science expert in the room. In the follow-up interview, you (and a data expert) can go into detail by asking the candidate to show an example of previous work with a walk-through. This interview can be up to two hours; also use it as a culture-fit barometer, so don’t shy away from tough questions. The final interview should be a wrap-up, where HR makes a job offer ideally on the day, or in the next few days – remember, data science is competitive, and a candidate likely has irons in more than one fire.

Kosta noted you can build data science teams internally: this offers more control over structure and is cheaper in the short-term, but often you must upskill to add value. He added that it’s hard to attract and retain talent from the outside, as banks offer high salaries, and the rest simply can’t compete.

Another option is to leverage consultants, as they offer a fast time-to-value, but cost more. If you opt for this route, ensure that whoever you use has a track record: ask for references and experience.

Kosta noted that consultants are great for testing a data science idea, if you have one thing to implement, or if you’re just starting. Many consultants can help your organisation identify opportunities for free so you can build a relationship with them. Just pay attention to contract clauses that cover IP ownership, how the handover would work, and exit assistance if a consultant shuts down.

Centres of excellence and ‘brilliant jerks’

Kosta added that a useful recruitment tactic is for an organisation to build a centre of excellence, as it’ll allow the business to compete on culture instead of salary. He explained that a centre of excellence is a standalone space where data scientists work together, celebrate successes, and where teams can give talks so everyone in the organisation can learn about what’s happening. Prioritise quiet study time, which is rare in large, open-plan corporates, and one or two recognised data science experts can work in the centre to serve as role models.
Kosta closed off by urging CHROs to “hire the best people, but avoid brilliant jerks”. In other words, some candidates are brilliant on paper, but they end up costing you more. How to spot these people is tricky, he says, as many data scientists are introverted and tend to shy away from confrontation, but they usually have an answer for everything, they don’t think things through, and they’re hard to fire.
However, Kosta concluded, with a well-structured and talented team, organisations can unlock the full potential of data to drive success.


Related articles

CEO Avian Bell's top tips for succession planning

Incoming CEO of Quantumed Avian Bell shares how a positive organisational culture, strong skills transfer and a solid succession plan enabled him to rise to the top rung of the corporate ladder.