Financial services analytics: Banks do not combine structured and unstructured data analytics.

April 7, 2014

Squirro was selected to participate in Accenture’s 2014 Fintech Innovation Lab in London.

During this 3 months lab we had a unique opportunity to work on a number of proof of concepts with leading global financial institutions including Barclays, Credit Suisse, Lloyds, RBS…

We decided to use this once in a lifetime opportunity to run a survey of their senior level executives to understand their current use, needs and future plans with regards to unstructured data analytics.

We asked 98 senior executives 19 multiple-choice questions and here are the results and over the next few weeks we will present you the results.

From today until Wednesday we will present the results of this survey at Qonnections 2014 in Orlando.

The first part of the survey consists of establishing the current state of affairs for financial services and unstructured data analytics.

As a starting point we asked both what type of unstructured and what type of structured data financial services used.

Figure 1 shows that customer transactions is the number 1 source of structured data followed by financial data and trading data. Together they represent 62% of the structured data used by financial services.

Additionally in other, different responded included data such as customer experience data or legal data.

Figure 2 shows that Internal reporting, internal communication and external communication with clients are the main sources of unstructured data representing 55% of all unstructured data currently used.

Interestingly of the respondents only 7% replied that financial news is a source of unstructured data in their company.

According to Mark R. Gilbert and Ted Friedman from Gartner: “Under pressure from regulations and potential legal liability, enterprises struggle to achieve a “single source of the truth” — the idea that one set of data can be trusted as being fully accurate and accountable. This has been a challenge because structured data has traditionally been stored in databases, while unstructured data (for example, documents and images) has resided on desktops, in file systems and in content management systems. The drivers for greater unification are clear, and compliance is only one of many forces pushing organisations to achieve a more integrated view of their enterprise data.”*

Their paper dates back from 2006, yet their statements are ever more so relevant in light of the recent libor scandals.

Unfortunately the results of our survey (Figure 3) show that to date little has been done in the financial sector regarding the integration and combination of unstructured data. In fact 82% of respondents declared that they did not combine structured and unstructured data in their company or department.

In conclusion while all companies surveyed used both structured and unstructured data analytics and used different sources of unstructured data, none of them combined those insights.

There is here a clear opportunity for financial services to get better insights from the data they generate.

As we will tomorrow and Wednesday in the presentation of the second and third parts of our research financial services recognise the need to integrate structured and unstructured data analytics but face numerous challenges when finding the right solution.


*Source: The New Data Frontier: Unifying Structured and Unstructured Data, Mark R. Gilbert & Ted Friedman, 31 March 2006