By Ben Richmond
While data may once have conjured images of Excel spreadsheets or filing cabinets, it lives in digital form. It is sophisticated but can be unwieldy. The sources of data are evolving too. Traditional data sources used for onboarding or know-your-customer requirements are slowly falling out of favor to be replaced by new sources, also known as alternative data.
Alternative data is essentially financial and non-financial data that is secured from non-traditional sources. Raw data is scraped from wide and far-reaching sources, including geolocational information, drones and social media information, to help make faster and more accurate decisions. Where once an onboarding process may have looked at bank statements and a person’s work history, banks are now able to make decisions from sources that determine a customer’s sentiment. What do they search for online, for example, and could this impact the products they care about?
A report by Burnmark and CUBE found that over the course of 2021, financial services harnessed alternative data points across the scope of their businesses. The report identified 14 alternative data sources, including drones being used for commercial loans and insurance products, social media data harnessed for onboarding and credit scoring, and behavioral biometrics within consumer risk management.
Is there a customer advantage?
Alternative data goes hand in hand with the digital transformation of financial services. Driven in part by the global pandemic, digital transformation has accelerated. For many customers, digital banking has been an invaluable resource, allowing them to remain financially active without having to leave their homes. There is naturally an argument to say, however, that financial digitization is exclusionary and may isolate those who are less digitally literate, which poses fair lending risks. Further important questions include: Are firms able to say that the data used is unbiased or that technology has not taken on the subliminal biases of the people who have developed it?
The Burnmark Report notes that Philippines-based CIMB Bank is using smartphone data to assess behavioral scores of applicants, drawing on this data source as an alternative to traditional credit review data. This raises an interesting counterpoint about access to banking through digitization. While increased technology within banking may be exclusionary for some, in this instance, the move away from traditional data sources for digital onboarding opens up the financial landscape for many. Those who do not meet traditional, rigid lending or risk criteria may now be able to prove their merit through geolocational data sources or those based on phone usage.
Consent is often raised as a contentious issue, as well as public interest. Where does this data come from, and do banks need it to be effective? Historically, most forms of alternative data have sat well within the public realm or have been consented to through third parties. However, alternative data sources are slowly falling under the remit of emerging data privacy laws from the California Consumer Privacy Act to GDPR. As we saw in enforcement action from the SEC case against App Annie, regulators are closing in on the flagrant use of alternative data without restriction.
Inevitably, banks will need to adopt a model to benefit all. Perhaps offering digitization in the main, with traditional alternatives for those who need it—all while meeting regulatory obligations.
Is alternative data use good for banks?
Aside from the customer advantage, the glut of new and evolving data sources presents limitless bounty for banks—if managed correctly. Automated systems, driven by artificial intelligence and machine learning, can quickly identify patterns within a series of unlinked data sets to make intelligent inferences about customers or market activity.
Dataminr, for example, which runs analytics on Twitter data, uncovered preliminary reports on the Volkswagen emission scandal three days before the market reacted. Other instances include where satellites are used to track the activities of industrial facilities from afar to generate reports on manufacturing activities. Using natural language processing, digital public data can be instantly analyzed to reveal sentiment and context, so banks can quickly identify attitudes toward investment factors, such as ESG.
The devil is in the data, though. If it’s managed well and properly understood, then the rewards are plentiful. However, if managed poorly, banks could very quickly find they have a surfeit of data that offers more challenges than it solves. Unstructured data, in its purest form, must be cleaned, prepared and analyzed before it adds value or provides insights. Internal processes for data management can often be laborious or badly built—while handing it over to third parties can be risky without watertight agreements, processes and policies. Of course, emergent issues are arising around data ethics, especially with regard to personal information. So, as well as meeting customer preferences and regulatory obligations, firms will also need to consider the ethics and biases of the data they harness.
Bad data, bad systems or a combination of the two could mean that banks receive little customer insight, but are on the sharp end of regulatory requirements, from storage to privacy to consent. The question is, are the risks worth the reward?
Will data be a poison or a cure for banks? It all depends on how they use it.
Ben Richmond is CEO of CUBE.
