We use cookies on all our websites to gather anonymous data to improve your experience of our websites and serve relevant ads that may be of interest to you. Please refer to the cookies policy to find out more.

By continuing, scrolling the page or clicking a link, you agree to the use of cookies.

How banks use alternative data

24 June, 2020Ouida Taaffe

The digital world is awash with alternative data which the banking and finance industry can use to improve products and services. Ouida Taaffe asks Sankar Krishnan – Executive Vice President and Industry Head, Banking & Capital Markets, at Capgemini – about alternative data, its influence on products, credit and the impact of Covid-19.

What do banks mean by alternative data?

Index card drawersFor over 200 years, banks have been serving customers and gathering data relating to them. Bank application forms were designed to capture elements of data – name, address, salary, amount of mortgage, terms of mortgage etc – for different products.

In the world we live in now, we need alternative data – also called unstructured data.

Examples include social media information, email information, and internet tags. Bank databases do not have this and it is a great way to know your customer better. In particular, it supports product personalisation.

So, if someone is linked to 5,000 influencers, does that help you come to a credit decision?

Yeah, it does help. Because credit scores are about a customers’ ability to pay and willingness to pay.

Credit score models were invented in the 1950s when there was no gig economy, no internet and no social media. Social media can help confirm the borrower’s identity and whether they are worthy of credit and if there are any undisclosed risks.

For example, an insurance company might see from social media that someone races cars, though this was not mentioned on the insurance application.

What about the traditional five Cs of credit – character, capacity, capital, conditions, collateral? Can you go beyond those with alternative data?

The five Cs are still very relevant. But the way you obtain them has changed.

It could be through social media, profiling, postcode analysis, or even sending a drone up and measuring the neighbourhood someone lives in. Knowing a postcode, you can pull data relevant to creditworthiness like never before.

What about being fair to people who feel that banks are not looking at the full picture?

Computer code on a screen

We have to call out what goes into the credit model from a fairness and transparency perspective. 

Digital account opening and transacting is here to stay – especially following Covid-19. Data on digital behaviour can enhance traditional credit models. It is extremely important for some of these data models to be ‘auditable’ for any ‘bias’.



Do you think banks could launch new credit products and expand overall lending by using alternative data?

Most of the large banks are responding to this very well. Recent public examples include:

  • the work of Citi and Google on card and checking accounts
  • Apple’s partnership with Goldman on cards, and
  • the bank inside the Amazon marketplace.

Increasingly, banks and credit card companies are working on strategies to be the ‘intel inside’ of a marketplace. Multi-sided platforms such as Uber, Airbnb, Amazon, Facebook are front and centre in this.

A lot of new credit products are emerging. Examples include the excellent work that Fronted is doing in the UK, or what SoFi is doing in the US with new products.

Thinking about the Covid-19 pandemic, can banks use alternative data to help them forecast the economic future?

Each bank has a credit or asset liability committee and produces daily reports on how factors like GDP, interest rates, FX rates etc affect the balance sheet.

They don’t look at single customers – unless they are super large – but at portfolios and segments of customers. Banks calculate their total value at risk. In other words, a lot of this is still traditional models.

Where alternative data is used is mainly in asset management industry. Crunching this data using AI models is getting cheaper, which means more is used. Examples include drone channel checks, satellite imagery data, spectrum data, telecom data, and smart city data.

How can banks use data to meet regulatory requirements?

Computer data on a screenRegtech is one of the fastest growing areas in financial services. The idea here is to use the latest technology to ensure that the regulatory environment gets better from a reporting and ‘dipcheck’ perspective. 

The challenge for banks is how to keep this data real time. Also, how do you ensure that the regulators are satisfied that the machine learning and AI models and automation models can be audited in real time to ensure accuracy and avoid fines?

Can retail banking services can tap into mobile phone data?

Absolutely. There's a lot of data consumers have control over sharing. But banks can use mobile devices to customise the banking experience.

Over the next three years it should be possible for us to speak into a device, or into a phone, and initiate a financial transaction with a higher amount of security. Already we are able to do that for basic financial services.

Should financial services institutions have access to your personal health data?

Remember that all the data is yours. They can't get anything unless you approve it.

The beauty of the new GDPR regulations is that if they get that data without your consent, boy, are they in hot water like never before!

In the digital world, you're able to control things a lot better, and who gets access to what. There are a variety of new cyber security companies that make it close to impossible to hack data. But it’s an ongoing battle between the ‘hackers’ and the ‘protectors’.

More data enables better personalisation and customisation and more data can also reduce the price we pay for a product or a service. In the end it is a balance.

Related content

See our Centre for Digital Banking and Finance