SINGAPORE (July 29): When Peter Barcak, co-founder and CEO of CredoLab, thought about his next move after selling off his stake in Platinum Bank in Ukraine, his attention turned to an issue he had faced as a banker. How does one address the problem of lack of data when assessing new customers for credit facilities?
“How [do you] improve the way banks do [the] underwriting, their ability to judge who is going to be a good customer or bad? That was an inspiration to look at alternative data,” Barcak tells The Edge Singapore.
He wanted to improve the situation in emerging economies, in particular where traditional sources of credit data, such as credit bureaus, were not available. He started CredoLab in January 2016 in Singapore, with the aim of using alternative data, or non-traditional sources of information, such as from mobile phones, to score the creditworthiness of potential banking customers.
Using Singapore as a base, CredoLab plans to tap the unbanked masses of Southeast Asia as demand for credit grows. Middle-class households in Southeast Asia are expected to double to 163 million by 2030, according to a McKinsey Global Institute report, “Future of Asia”. Of the 1.7 billion adults in the world who are still unbanked, 95 million are in Indonesia alone, according to the Global Findex database.
Access to financial services is one way out of extreme poverty. For instance, the unbanked would struggle to start a business or grow one without financing.
CredoLab is but one of many start-ups using smartphone data to provide credit scoring, and the industry looks set to expand as the population of the unbanked is massive.
“The persistent problem is that lenders have no proper tools to make a good [assessment]. Consequently, what they do is reject an applicant because they don’t know [anything] about the person. The person doesn’t have a credit card, doesn’t have a record in a credit bureau,” says Barcak.
Such rejections not only cost the applicants time, but are costs to lenders as well. “Also, it damages the relationship because the customer is leaving without [getting credit],” he adds.
What CredoLab does is provide lenders with an alternative credit scoring model that utilises data from an applicant’s smartphone. Applicants have to install a white-labelled app, which pulls metadata and data from their smartphone to determine if the applicant is expected to be able to repay his or her credit on time.
“Because not everyone has a credit card or a credit bureau, but almost everyone carries a smartphone. A smartphone is a great source of data, because it contains your daily activity,” says Barcak.
“It is very important to mention here that we access only metadata: only information about information. So, we don’t have access to any private or sensitive information,” he adds, highlighting that the potential customer would still have to approve the app’s access to the data.
“Based on metadata, we are still able to develop a very sophisticated solution that helps banks and lenders in this region improve their model’s predictiveness.”
CredoLab typically runs a data collection phase over one to two months to amass sufficient data to develop a “scorecard” for loan applicants. It will use more than 50,000 data points collected from metadata across SMS, calendars, applications and storage, to determine an applicant’s behavioural characteristics. Each lender would get a tailor-made solution from CredoLab, utilising data mined from their customers.
“In order to develop a scorecard, we need to correlate data we have access to [through the applicant] and performance data [from the lender], which is a simple indication from a lender that [the customers are good or bad]. Then, we look for similar behavioural patterns among bad and good customers to come up with a bad and good scorecard,” says Barcak.
“Our decision model system provides a score within a couple of seconds, so the lender can make a decision instantly.”
According to Barcak, both creditors and borrowers have benefited from CredoLab. Lenders have reported a 20% higher approval rate for loans/credit card applications leading to more borrowers benefiting, while seeing a 15% dip in non-performing loans and a 22% decline in fraud cases.
“Every month, we help to process credit products in the volume of over US$500 million ($682 million),” he says.
CredoLab has 50 clients currently, spread across Southeast Asia in markets such as Vietnam, the Philippines, Thailand, Myanmar and Indonesia. These range from large national banks to regional banks and consumer finance firms, though the start-up has declined to name its clients.
CredoLab is also embarking on an expansion into India, Africa and South America. “These [areas] have similar issues as consumers in Southeast Asia, such as [a lack of] access to regular financial services at a reasonable cost. Our initiative is part of financial inclusion, helping customers to have access to standard banking services, so we take this very seriously,” says Barcak.
Besides geographical expansion, CredoLab is looking into expanding its product offerings as well. “We want to expand and make our products more visible, immediate and more convenient. Just last month, we launched a new product, Credoapply, an application-form platform that any bank or lender can use for the digital onboarding of new customers. We will continue to deliver new products and innovations,” says Barcak.
CredoLab has raised US$3.1 million in funding so far from venture capitalists such as Walden International, FORUM and Fintonia Group. Investors also include financial service lead generator GoBear, which the start-up sees as a strategic partnership: CredoLab makes their leads better while it provides them with a new data source.
“We are looking for investors who can bring in value, besides financial value, and help us increase our client base or innovate our products by adding new data sources,” says Barcak.
