Blog article
See all stories »

Banking industry must think 'out of the box' to solve the data science skills gap

Arthur C. Clarke, writer of the famous science fiction epic 2001: Space Odyssey, famously remarked that any sufficiently advanced technology is indistinguishable from magic. The problem with magical things, however, is that we often ascribe them miraculous powers that are beyond their actual capabilities while ignoring the fact that far greater powers reside in a much humbler vessel – the human cranium.

Data analytics provides a great example of this disconnect. Every organisation knows that it needs to unleash the power of the information that it holds, while simultaneously being painfully aware that there is currently a worldwide shortage of data scientists. (A LinkedIn Workforce Report for the US earlier this year identified 151,000 data scientist jobs going unfilled in that region alone.) Instead of incurring the expense of hiring and retaining these experts, it may seem sensible to ask if there is some magical technological substitute – a ‘data scientist in a box’.

But can large, complex organisations like banks really swerve the dearth of data science skills and unlock the insight they need, all from a magical technological box somewhere in the cloud?

 Asking the right questions

In truth, there’s no such thing as a data scientist in a box; nor do should there ever be. Machines can replace us in many spheres of activity but, as anyone who remembers HAL from 2001: A Space Odyssey knows all too well, we should never expect them to take total control of our most important missions without human supervision.

Marketing provides a great example of machines’ limitations. Communicating with consumers and, indeed, designing new products and services for them, requires a comprehensive understanding of customers themselves. A bank can’t load all its data into a box and hope that it spits out the ‘right’ answer to such complex questions such as how to attract new customers or improve the banking experience. Some common sense is required, and no matter the speed at which technology can take over the heavy lifting, it is still the case that rubbish in means rubbish out when it comes to data-based insight.

Instead, marketers in the banking industry must understand their audience and their needs intimately; they must be able to segment them into groups or personas while, of course, mapping these to the bank’s strategic aims – all before they even begin to get to grips with the data. A technological “no code” solution can bring enormous benefits to this process of design, testing and iteration, enabling non-specialists to turn masses of data into useable insight. But it can never replace data scientists, who bring their own invaluable, indispensable magic to the process, ensuring that data is applied to problems correctly.

 What do data scientists actually do?

There’s still a degree of misunderstanding of data scientists’ role in an organisation. They are not data-evangelists-in-residence, nor are they there to add their own unique insight to existing projects. Data scientists will, in all likelihood, know little or nothing about the actual business in which they’re engaged. They wouldn’t be able to tell you about merits or demerits of a new mortgage or credit card; nor will they understand the banks’ marketing objectives.

Their job is to make sure the data is delivered in a way that it can be applied (and trusted) by non-specialists in marketing or other areas of the business engaged in the process of evolving new initiatives.

Marketers, for example, use intelligent engagement platforms to understand their audiences and to test new campaigns among carefully defined groups. This is a great example of the democratisation of data that’s been enabled by a new generation of powerful tools that enable anyone to become a “citizen developer” without a degree in computer science. Yet if these tools are being fed with erroneous metrics, then the results will obviously be inaccurate – perhaps catastrophically so.

Perhaps a better way of thinking of data scientists is as “data stewards” who are responsible for collating data from numerous silos, checking and deduplicating, defining models and metrics, and ensuring they iron out the tiny errors that can add up to an unbalanced sample. Data scientists also make a direct impact on the big picture, improving business outcomes through implementing data-driven techniques, such as predictive and prescriptive analytics, throughout the organisation. Being able to link the IT and human worlds, data scientists are vital to driving forward the digitally enabled business.

For example, a bank might need to identify a group of people who do not have an active financial product such as car insurance. Without a data steward, the data from which the non-specialist is working might include someone whose insurance was cancelled after one too many crashes. Technically, this person doesn’t hold a financial product – but that doesn’t mean that they’re the right person to target with marketing about renewing their insurance.

Driving relevant interactions

It’s only once this important work has been completed that non-specialists can get on with the task in hand. And for this, they will need a ‘box of magic’, in the form of an intelligent engagement platform that can stitch together accurate data from multiple sources to gain a true picture of customers and drive relevant interactions with them.

The dashboards, analytics and self-service capabilities that make it easy for a non-specialist to design and test campaigns from pre-built use cases are all built on a solid foundation of accurate data. In this sense, the data scientist operates at the very beginning of the chain that, ultimately, leads to improved offerings, more personalised communications and an enhanced customer experience.

It may seem like the modern world is run on magic, but anyone who has worked with a modern intelligent engagement platform will realise that the combination of smart people and easy to use technology is the most powerful one. Both the data scientist who sets up the system and defines predictive models and the marketer who orchestrates the delivery of relevant customer interactions in real time are both essential to the process of driving business value.


Comments: (0)

Doug Gross

Doug Gross



Member since

10 Sep 2019


New York

Blog posts


This post is from a series of posts in the group:

Analytics in Banking

This is for discussion and sharing of views on trends, practices and views in analytics in banking and financial industry

See all

Now hiring