The role of a data scientist and why we need them

Big data analytics can transform how businesses operate. While marketing and sales businesses cottoned on to this early, more and more previously non-tech-focused companies are realising the benefit of having expertise on board.

“The way you can use data is a way not just to understand customers and products better, but also be organised in a way that makes decision-making faster and allows people to have more autonomy,” says Anthony Fletcher, chief executive of snack company Graze.

Graze has its own dedicated data team, who “democratise data”. Mr Fletcher says: “This involves finding sources of data, cleaning and feeding it into our central cloud-based data warehouse where anyone can access it anywhere, on a variety of devices.

“This tracks literally thousands of different data sources from how fast our factory lines are running, the strategies for posting boxes in the US, 15,000 pieces of product feedback we get an hour, social data, trends in sales, how people are using our mobile site, and so on.”

The data scientists at Graze are also skilled at doing pieces of complex data analysis from attribution, A/B tests, machine-learning or various types of regression. “This often involves not only maths, but being really good at visualising data in ways which really tell the story,” says Mr Fletcher.

Some companies position their data team in different ways. At Graze it reports to the chief executive. “This was highly unusual, but reflected the efforts of building the original infrastructure and creating the right culture around data,” says Mr Fletcher. “The important thing today is they are a team who work across the business. This tends to be a draw for data scientists here as they get to solve so many different types of problems.”

James Parsons, chief executive of digital workforce and consultancy Arrows Group, says: “Data scientists are the rocket scientists of the digital world and the role of the chief data scientist (CDS) is emerging as the influence of data spreads horizontally across business functions.”


Three pillars

A company’s digital capability generally has three pillars, Mr Parsons explains, and the role of the CDS differs from the job of the chief information officer in relation to these three. “The IT infrastructures – platforms and tools – as well as the processes and methodologies – agile development practices – come under the jurisdiction of the chief information officer (CIO). The third pillar of digital capability is a company’s data assets. Exploiting these assets into actionable insights is a function that is starting to outgrow the role of the CIO and this is where the CDS comes in,” he says.

It is the role of the CDS to do the heavy lifting in managing the incoming flood of data. “They ensure that data is assimilated and analysed, then delivered as actionable insights to different stakeholders within the business. Designing and maintaining a standards-based data infrastructure is crucial if insights are to be made intelligible for departments such as sales, marketing and finance,” says Mr Parsons.

So, where should data science be positioned within a company for success? “This all depends on the size and hierarchy of a business,” he says. “For smaller enterprises, the data science function will often fall within the IT team’s remit. But larger enterprises are beginning to create a distinct data science function. One company we work with is creating a horizontal data capability which spans their global operation. This moves away from the siloed ways of processing data – an essential step in achieving data-driven actions across all departments.”

A chief data scientist will be the enabler for next-generation innovations such as machine-learning algorithms, autonomous product development and accurate financial forecasting

Data science is not a new role, but the creation of a chief data scientist represents the sharpening of executive capability for big data solutions. “The speed with which they are able to achieve innovation is of great importance and an increased capability to analyse the data streams flooding into a business is key to this,” says Mr Parsons.

“A CDS will be the enabler for next-generation innovations such as machine-learning algorithms, autonomous product development and accurate financial forecasting as all these are powered by big data insights.”

While companies of different sizes and at different stages of big data adoption may not always be ready for a CDS, the value of data to all businesses is growing exponentially. Bringing on board a new CDS may well be the best way to liberate insights from data for maximum business benefit.

“Having a CDS is of enormous benefit for companies that recognise data as a strategic asset to driving growth and profitability. It also sends a clear message that, as an enterprise, you take gaining value from data seriously,” concludes Sachin Bagla, associate vice president at Infosys. “However, the role requires a board-level mandate and sponsorship to be successful.”