Increasing demand for predictive analytics from Asean banks: IBM

By Goh Thean Eu June 18, 2014

Can be a differentiator for banks, in both retail and corporate segments
Most banks in Asia have yet to embrace predictive analytics solutions

Increasing demand for predictive analytics from Asean banks: IBM

MORE banks in South-East Asia, including Malaysia, are looking at how they can use predictive analytics as a differentiator and to better compete against rivals, according to technology giant IBM.

One of the areas banks are looking into is liquidity management, said Ashutosh S. Joshi, payments solutions leader, Payment & Transaction Banking, IBM Banking and Financial Markets, Asean.

“Improving liquidity forecasts is one of the top concerns of corporates, and that is one area where banks can differentiate themselves,” he told Digital News Asia (DNA) in Kuala Lumpur recently.

Ashutosh said that IBM is currently creating advanced algorithms for banks to help their corporate customers better predict liquidity flows.

“For corporates, they need to make payments to their vendors and receive payments from their customers on a very regular basis.

“We have seen that on many occasions, that collection does not necessarily come on the date of the invoice. It may come one day later, or 10 days later. So, because of that, the corporates cannot predict or forecast the exact amount of liquidity they have at any particular point in time,” he said.

Citing an example of a corporate treasurer who is expecting to receive a payment of RM100 million based on the invoice date and is expecting to make a payment of RM90 million to a vendor on the same day, Ashutosh said that the corporate will not have any issues if the payments were received on time.

“However, what happens if the corporate treasurer only receives RM80 million? Then he may be having a challenging time as he will be scrambling and making phone calls to banks for an additional line of credit if he doesn’t have an existing one,” he said.

With a predictive tool, the corporate treasurer can better manage his organisation's payables and receivables.

Ashutosh said that such predictive tools are 'first-of-their-kind' innovations and corporate customers can save interest costs with better liquidity management tools.

Such tools would be helpful and useful for corporate customers, but how do they help the banks themselves?

Ashutosh said that in the competitive banking industry today, such analytics and predictive tools could be the “competitive advantage” that banks are looking for.

“They can get more ‘stickiness’ with their corporate customers. This means potentially more revenue and deals with corporate customers,” he claimed.

Banks still new to predictive analytics

According to Ashutosh (pic), there are three levels of analytics, the first being basic reporting and spreadsheets, basically involving crunching data in a 'silo.' The second level, which is where most banks in Malaysia and South-East Asia are at, involves taking the data and analysing it in a more holistic manner.

“For example: a bank looks at a customer’s credit card spending and finds that the customer has six transactions a year done in a jewellery store. From the data, the banks may conclude that the customer likes to shop in a jewellery store and may offer special discounts related to the store in the future,” he explained.

“The third level of analytics is predictive analytics. This is what most banks are looking into at the moment,” he added.

Besides liquidity predictive tools [that could help them secure more corporate customers], banks are also looking at big data tools to help them capture more retail customers.

The progressive banks are looking at it, putting frameworks in place, and looking at how to use structured and unstructured data,” Ashutosh said.

Nevertheless, he said there are various challenges ahead should banks want to implement advanced analytic solutions, such as predictive cash flow.

"The first is getting data from the various sources as this is subject to what is available and usable. Second, the data needs to be from over a period of at least a year, if not more, as age helps in testing the algorihms to forecast or predict future activity and detect patterns.

"The third challenge, which is also the key value, is the ability to determine appropriate algorithms that are able to predict the cash flow with improved levels of accuracy or probability," he said.

For more technology news and the latest updates, follow us on Twitter, LinkedIn or Like us on Facebook.