Banks and big data

There are three important reasons for the slow adoption of big data in South African financial services.

• Confusion around defining big data

  South Africa still doesn’t fully understand what big data is. This is mainly due to the – vastly confusing – variety of individual definitions, presentations and use cases presented by virtually every infrastructure vendor, mega-IT-vendor and storage vendor. So, it’s little wonder we are confused.

  Doug Laney coined the original “Three Vs definition” in 2001 which describes big data as a combination of the variety, the velocity (or processing speed), and the volume of data available.

  Applying this definition remains a good technical test of whether or not something can be considered big data. Simply put – if you can use existing SQL type technology, then it cannot be considered big data.

  My preferred definition is related to its primary benefit: the ability to deliver valuable insight in business time. Big data approaches reduce development time and work well for ad hoc analysis – ensuring that answers can be delivered while they are still relevant to the business decision.

• Identifying applications

  Banks are struggling to identify suitable applications, although there are many potential use cases out there. If we look at case studies from international big data leaders, we can see that their customers have moved beyond IT experimentation to identifying business-driven use cases - such as fraud analytics, customer analytics and regulatory reporting.

  One international trend has been the shift away from the centralised data lake to the delivery of departmental solutions.

  For example, a major credit card vendor is combining transactional, spatial, merchant and customer data to identify and manage fraudulent transactions.

  The marketing team, on the other hand is combining customer data with social shares to deliver better targeted marketing – saving money and increasing conversions.

• Lack of skills

  Finally, the lack of big data skills in South Africa remains a real challenge. Systems requires technical know-how that is in short supply our market, and its rapidly evolving ecosystem makes it difficult to keep skills up to date. One only needs to look at the evolution of data analytics on the software – from technical, to self-service, to visual / iterative – to see how quickly it develops.

  This does not mean that every employee needs to become a data scientist. However, by reducing the complexity of data preparation, investigation and even analysis we can get more value from our data scientists. We can also deliver some level of governed self-service to allow the broader business community to take on some of the simpler analytics problems themselves.

There are multiple ways in which big data could both improve and support South African financial business, if they leverage the technologies and resources available.