Singapore-headquartered DBS Bank is aiming to revolutionize its data analytics capabilities through a program called Data First. As part of this change program, the bank is looking to tap into a new type of data professional—that of the data translator.

Paul Cobban, chief data and transformation officer at DBS, tells WatersTechnology that under the Data First program, one of the three pillars is culture and capability, where the bank conducts a lot of internal training. Through this program, DBS has identified the data translator as a crucial role for the business.

Cobban explains that a data translator is not a data scientist, but knows enough about data analytics to understand what data scientists talk about. “They’re also anchored in the business so they can translate data science concepts to the business and vice versa. This is a big focus for us, and we want many of our people to become data translators,” he says. Essentially, the role serves as a bridge between the data scientists and business professionals.

[Without help] the data scientist is spending 90% of their time doing things other people can do. It’s the same on the translation side, which is why we think data translators are important because data scientists may not necessarily understand the business.

Paul Cobban, DBS Bank

As a result, they are now both training staff internally to be data translators and are recruiting externally to fill the role.

The role of the data translator is less known compared with the data scientist role. The data scientist role has gained popularity in recent years thanks to an explosion of new datasets available, the ability to relatively cheaply store and analyze this data via the cloud, and the need to seek alpha by finding insights from non-traditional data sources. 

Data scientists have become the answer to turning data into valuable and actionable insights and tend to be quantitatively focused. The problem that has been encountered is that mathematically-driven data scientists can run into problems when dealing with, say, a sales trader or non-quant portfolio manager.

Often, there can be a disconnect between the two in terms of expectations as to what can be delivered and timetables. For example, a data scientist understands the mathematical problems at hand, but may not necessarily understand why or how it’s a problem for a bank, specifically.

In an article published by Harvard Business Review in February 2018, consulting firm McKinsey said translators play a critical role in bridging the technical expertise of data engineers and data scientists with the operational expertise of marketing, supply chain, manufacturing, risk, and other frontline managers. McKinsey predicts demand for translators in the US alone may reach two to four million by 2026. 

Cobban notes that by combining a data translator with data scientists, the latter can be more experimental with their modeling while the translator can make sure that the actual business problem is being specifically addressed.

“The problem may change after going through exploratory data analysis, and we have a formal process to do this,” he says. “To be most effective, it is best to supplement data scientists with other roles, so the data scientists are freed up to do more sophisticated modeling.” 

As the data scientist becomes more valuable, what has also become clear is the need to couple this role with other data specialists. An example would be the data engineer. The data engineer doesn’t have to delve as deep in the mathematics as a data scientist needs to, but will ensure that the dataset is clean enough for the data scientist to use. 

“Otherwise, the data scientist is spending 90% of their time doing things other people can do. It’s the same on the translation side, which is why we think data translators are important because data scientists may not necessarily understand the business,” Cobban adds. 

DBS currently has about 150 data projects running at the moment, focused on improving revenue, reducing costs, or improving customer experience, according to Cobban. Typically, one data translator will be assigned to two data scientists on any given data project.