Securing your data science future: QuantumBlack’s Data Science Leadership Breakfast

By Carlo Giovine, Associate Partner and Didier Vila, Global Head of Data Science, at QuantumBlack.

Leveraging the power of data has become increasingly important for modern businesses. However, doing this is often easier said than done. A hyper competitive market for skilled data experts and legacy systems in drastic need of overhaul have left many companies questioning how they can effectively build and harness the potential of data science teams.

QuantumBlack set out to provide answers this summer, bringing together a selection of 18 data science leaders working across a range of businesses, from FTSE 100 companies to startups.

Our aim was to create a forum where leaders felt comfortable to share their experiences and build a framework of best practices that everyone can benefit from.

The first session focused on two key challenges: how to acquire and retain data talent, and how best to utilise it to achieve success. Below are our takeaways from the session — our next event is planned for 27th September, and details on how to apply can be found at the end of the article.


Maintaining your data talent pipeline

Skilled data scientists have become a hot commodity in today’s labour market, with demand far outstripping supply. All attendees at our breakfast agreed that filling vacancies and ensuring existing experts aren’t lured away has become increasingly challenging in this competitive environment.

Attendees discussed common pitfalls when it comes to recruiting data talent, and confusion among data illiterate companies was a recurring bugbear. There was unanimous agreement that many job descriptions often leave candidates with more questions than answers. There were plenty of anecdotes about data scientist job listings which read like they were posted by a leading tech firm, calling for colleagues to build and deploy complex machine learning models — only for the candidate to arrive at the company, find very little data infrastructure and end up focused on creating dashboards.

This confusion also extends to salary structures on offer in the data science jobs market, which our attendees noted were becoming somewhat overheated. Many businesses have responded to the talent shortage by attempting to attract data scientists with increasingly impressive salary packages, which exacerbates the market’s competitiveness.

Solutions were proposed throughout the session. Our attendees agreed that it’s crucial for businesses to understand what they want data scientists to work on before they hire them — not only will this help in sourcing the most appropriate candidates, but it also highlights what foundations need to be put in place before the new hire arrives.

One guest suggested that businesses should be introducing themselves to the candidate pool at an earlier stage by forging better links with academia, either by funding PhD programmes or providing expert lectures to data science and mathematics university courses. Another suggested that retaining skilled staff was the key priority, and that employees should be empowered to pursue their own development.

This chimed with QuantumBlack’s outlook on workplace culture. Our Guild Mentorship programme sees that each of our employees receive close guidance from a more experienced data scientist. Naturally this involves direction on technical development, but we also structure the programme to include regular advisory sessions, where each employee can discuss their vision for their career and the wider business.

By having this deep understanding of each member of staff’s aspirations, we’re able to assign them to projects we know they’re passionate about and would best suit their career vision. Furthermore, this ensures everyone feels invested in the direction of QuantumBlack and nobody is isolated from our business’ senior decision makers.

We take care to recognise that data scientists, like any other employee, will have their own passions that they want to pursue. With this in mind, we invest heavily in research and development and deliberately provide structured time for our data scientists to pursue their own ‘passion’ projects in parallel with client work.


Harnessing the power of your data science team

For many businesses, the confusion doesn’t end once you’ve hired your data scientists. The majority of our attendees agreed that successfully leveraging a data science team can often depend on how an organisation is structured.

Many attendees advised that firms should centralise their data teams in the early stages, particularly during their first foray into data science. With so many aspects to be determined, whether that’s objectives, infrastructure or team capabilities, it’s best to have data scientists working primarily within their own team in these opening phases. As the analytics group reaches maturity by settling on a defined protocol and demonstrating early successes, companies can then accelerate data science throughout their organisation.

Our attendees recognised that achieving this far-reaching, federated model is challenging. However, there was common agreement that it worked best when selected data scientists were designated to interact with wider departments. These individuals are often able to act as ‘data ambassadors’, introducing new concepts to the wider organisation and behaving as a catalyst for change and new ideas from teams throughout the organisation.

Towards the end of our session, the topic of senior stakeholders was debated. There was an overall consensus that boards are positive about the potential of AI, and this general advocacy at the top is matched by a great enthusiasm for innovation from junior employees. However, too many of our attendees have encountered ‘the frozen sandwich’ — an obstinate layer of middle management which is often resistant to change, due to either a lack of skills or a lack of understanding about what data can achieve.

It’s a situation that many of us are all too familiar with, and we debated at length on how it could be addressed. One proposed solution was that data scientists present their projects directly to senior stakeholders for investment or buy in, in a similar fashion to academia funding programmes. This would secure senior champions for data science projects within a business and accelerate progression from the top.


We’d like to thank all attendees for their time and contributions to a tremendously insightful first session. Our next Data Science & Leadership Breakfast is scheduled for Thursday 27th September in London — we’ll be exploring ways to bridge the gap between ‘proof of concept’ phase to production.

It promises to be a lively session — if you would like to attend, please contact Didier Vila, our Global Head of Data Science, or Carlo Giovine, Associate Partner at QuantumBlack, who both oversee the event series.