Unstructured data is currently attracting a lot of interest among banking innovators. What exactly is it, and why is it important?

Fintech2030 Q&A – A series of Q&As with leading experts on the future of fintech in payments, banking and capital markets

Bob Legters is chief data officer at FIS® and a regular commentator for Forbes® on data strategy and its application across all industries to inform and improve customer experiences.

Whereas structured data is highly organized and formatted in a way that’s easily searchable in databases – such as social security numbers or bank account numbers – unstructured data has no predefined format or organization, is often text-heavy and can be located in an endless number of systems and technologies. For example, this would include information stored in consumer documents, video files, voice recordings or social media posts.

As well as being useful to correlate and corroborate data from structured sources, unstructured data is extremely valuable because it provides additional information, context and insights into people’s behaviors and preferences. The problem, however, is that it is often difficult to locate, process and analyze effectively.

Therefore, unlocking the value in unstructured data is going to be a gold rush of sorts for the next decade. I fully believe that determining the sources and value of unstructured data efficiently will become a key tool for businesses in understanding changing customer needs and staying ahead of future trends.

What are the main applications of unstructured data for banks?

According to recent research by Fintech Futures®, approximately 80% of banking data is unstructured. This presents a huge opportunity for banks of all sizes, with two principal applications: fighting fraud and improving consumer experience.

In fighting fraud, structured data in the form of passwords and login IDs has been in use for over 30 years. However, because the likelihood of fraud is increasing with the use of mobile devices, banks are beginning to combine unstructured data, such as device numbers and geolocation, to prevent fraud by helping identify when someone attempts to log in from an unfamiliar device.

When it comes to improving customer experience, a key trend we are seeing among financial institutions is the blending of traditional data with new sources. Combining these new insights with information that already exists, such as transaction data, profile data and scoring data will allow banks to create a more accurate and comprehensive picture of their customers, allowing them in turn to develop new and tailored product offerings.

Experian Boost® is a good example of this. By adding typically unstructured data to the structured credit score process, more informed pictures of the consumer can be created, often times providing a better score for consumers. In the case of Experian Boost, consumers are able to add additional information, such as how much they pay in rent or the total cost of their utility bills. This information is then used to boost a customer’s credit score, mitigating risk for the credit bureau and creating a financial benefit for the customer through improved scoring.

What are the main applications of unstructured data for banks?

Most banks, regardless of size, are currently overwhelmed by the amount of structured data they already have. Combining, cleaning and simplifying the storage of this existing data will be the first task.

In recent examples, I have worked with banks attempting to collect the structured data that exists across all the various software used in the bank. Much of the information in each system is the same, such as names, account numbers, etc., and there is some unique information each system creates and maintains, as well. Getting all of that information together and normalized is a foundational step for structured data. Using that combined data to help with consistency, automation through AI and machine learning and to build risk or marketing plans is a key goal for most banks. To do this for the average bank, we will be supporting over 35 data sets across multiple platforms.

The other big challenge for the use of unstructured data, generally, will be compliance with regulations such as the General Data Protection Regulation® (GDPR) or California Consumer Privacy Act® (CCPA), which limit the collection, use and storage of certain data types.

This means balancing convenience with privacy and security will be an issue that financial institution leaders will need to consider in detail. This can include using unstructured data with limited permissions and in anonymized forms until the regulations evolve to reflect the larger industry trend of unstructured data being combined with structured financial data.

However, banks can start preparing for the future now. In addition to organizing their existing data, financial institutions of all sizes should start thinking about what types of unstructured data will be useful for them and where to get it from.

Of the financial institutions globally that are currently trying to do something with unstructured data, most are using third-party providers and partners to source the additional data sets, rather than trying to create their own versions of unstructured data. This is a trend we’re likely to see increase over the next 18 to 24 months.

How can we expect to see banks using US data in the next two years?

In the next two years, the biggest change we will see is the much-increased use of unstructured data to reduce risk and fraud, using data from a third party, such as a geolocation data.

After that, we will then start to see more complex interactions, such as the blending of credit file data or even social media data, to create a fuller understanding of customers. I expect this to lead the AI and ML innovations in the near future, but getting this type of capability into mainstream use will take some time. One thing is for sure: Using this new data will be a highly effective way of developing new services and connecting with more customers in this increasingly digital age.

For FIS, this means there’s a clear market need for assisting banks as they combine this new data with their existing data sets to help drive growth and engagement, one that has also developed into a service offering itself. For example, our Ethos offering is already working with banks of all sizes to combine data sets, enabling them to develop a 360-degree view of their customers’ banking habits to enhance their online customer experience and develop new product offerings.

How can we expect to see banks using US data in five years and beyond?

As banks become more sophisticated in combining data with third-party sources, they’ll also become more sophisticated in collecting unstructured data from their own customer interactions, such as call center recordings information from smart speakers.

Over the next five years, we will see banks of all sizes using unstructured data, as well as parallel technologies such as natural language processing and AI, in a hyperpersonalized way. This will not only improve the customer experience, but allow connectivity with customers across all channels – including mobile, apps and voice – simultaneously.

These new data-driven tools will help focus the spending of banks, and more personalized offers for financial products will emerge. The combination of financial data and unstructured data can help predict the changes in life cycles that occur as consumers exit college and move into revenue earning and debt creation and then help navigate the development into investments, family planning and retirement.

And within the next decade, we will see these services combine. In the future, my smart speaker won’t just find me the best rate on a car loan. Instead, it will transform into a banking virtual assistant, using multiple data sets to not only apply for the best loan for me, but recommend if using a credit card would be savvier and even tell me how to boost my credit score and earn loyalty points at the same time.

This is why the future of unstructured data is incredibly exciting.

All third-party trademarks are the property of their respective owners and not affiliated with FIS.