Contact Us
  • Blog

The talent challenge in data analytics

The right people are in great demand, but the supply of them falls short

With the advent of AI and machine learning, organizations are maturing their data-analytics strategies. They are unlocking dark data, creating knowledge graphs specific to their needs, and generating real business insights, such as next best action recommendations. As is often said, data is the foundation for AI.

In our work with leading enterprises around the world, we've learned that harvesting that data foundation requires a few key competencies – and talent is a critical one to get right. Enterprises need to have the right people – ones who can innovate business models on the back of data insights and set up a data strategy and infrastructure. Moreover, the right people can discover, wrangle, model, and contextualize data and insights.

That was the consensus of a panel of experts that I recently hosted on Genpact's Journeys in Transformation webinar, “Getting ready for AI, data lakes or a big swamp." During the discussion, our talk inevitably turned to the challenge of finding talent in a market where good people are in high demand and equally short supply.

Amaresh Tripathy, an analytics services leader at Genpact, noted that at the fundamental level, data scientists are critical to success. But even more importantly, he added, are those who understand the two sides of this new world – the business side and the data-science side – and can converse fluently with people from both. ­

Tripathy sees a deep bench of data scientists who can spin extensive webs of data into theoretical models. At the same time, though, there are far fewer people who can take a model and translate it into a working reality in business on an industrial-size scale.

These “bilinguals," as Tripathy dubbed them, are people who are comfortable with both analytics and domain – for instance, people who understand machine-learning modeling or a forecasting model, as well as finance and accounting processes and operations. When helping enterprises set up a foundation for analytics and AI, Tripathy first assesses the number of bilinguals they have.

At the same time, data-savvy talent is not easy to find, of course. The demand for data science means much of the talent in data engineering has moved into higher-profile roles, as Bhaskar Ghosh observed. As a result, said Ghosh, partner and CTO at 8VC, a data-analytics-focused venture-capital firm, the so-called boring work of data plumbing, integration and cleansing – data engineering – is where organizations most need talent. But it is also where talent is least available.

So how do we address the problem? Rather than try to solve the issue solely with the existing labor pool of talent, he thinks it's “a generational problem." Ghosh works with computer science departments at leading universities to encourage data analytics curricula and experience.

Similarly, Tushar Shah, vice president of enterprise service platforms at PayPal, stressed it's not just who you're trying to hunt down, but where you look for them. Shah said PayPal started to consider different geographies where local talent might best address what its needs. Beyond China and India, Shah scavenges for talent in countries, like Singapore, that are starting to play a big role in blockchain. The lesson: firms must also tap into this raw-talent market whenever possible.

Finally, Barbara Stortz, Microsoft's data science leader, focused on properly defining the kind of talent that you are seeking. “Data science is too broad a grouping," she cautioned.

Like the word “doctor," the term "data scientist" can mean a lot of different things, she explained. Heart surgeons and dermatologists are specialists. In the same way, you should think about your talent needs in terms of the specialty you want.

And as data science evolves and organizations become more bilingual, the designations we assign to tomorrow's jobs – and the expectations accompanying them – will become ever more specialized and demanding.

To learn about other important ways to keep your data lake from turning into a swamp, check out the full Journeys in Transformation roundtable discussion here.