CredoLab looks to alternative data to assess credit risk
Traditional credit risk scoring methods already had their warts before COVID-19, but the pandemic has exposed a key flaw, which has lenders looking for better ways of assessing risk.
Enter CredoLab, which evaluates smartphone and web metadata to produce digital scorecards allowing lenders to make better decisions, especially for thin-file and underbanked people.
More than one in four Americans do not have enough credit history to be properly scored by traditional bureaus. Many people with low incomes find it difficult to open up the bank accounts which will help them start to establish that history.
There’s a better way, CredoLab CPO Michele Tucci said. CredoLab uses POS decisioning among other methods to help companies find reliable customers from those whose credit worthiness cannot be accurately assessed by traditional PII methods. With the pandemic driving more and more commerce to digital channels, such solutions are needed.
Over the past two decades Mr. Tucci, CEO and founder Peter Barcak and the growing CredoLab team tested different forms of alternative data and learned over time that if you had the right tools you could extract valuable insights from users’ smartphone data. The clients agree as CredoLab has assembled a roster of 70 clients in 21 countries over the last five years.
That success was undoubtedly a factor in CredoLab’s recent successful $7 million Series A round, which was led by GBG, a provider of identity location, location management and fraud and compliance management solutions, Walden International and Access VC.
Key to CredoLab’s identity is responsible data use, Mr. Tucci explained. They don’t scrape from unknowing customers and only access data after receiving operating system permission. The data they assess doesn’t include name, gender, profession, age and other personal identifiers.
The types of available data depend on where you are located, Mr. Tucci explained. In North America, alternative data is limited to transactional data like utility bills and bank transactions. That can create friction for the customer because some systems will need to ask more questions to make a decision. In parts of Southeast Asia online banking penetration is scarce. You still end up with plenty of data but it’s not as diverse as you want.
Another option is to buy data from companies who generate it for a desired market, but that tactic creates its own problems, Mr. Tucci explained. Most companies are hesitant to give it up so they charge a bundle. And because they only have a sliver of a particular market you have to purchase sets from three or more providers to develop a representative sample.
If only there was a reliable and economical source of data that is available in all areas of the planet. Wait, there is, and that’s the volumes of data that can be gleaned from smartphones.
CredoLab has an ambitious expansion planned in the coming months and a look at their target markets shows how far mobile penetration has occurred in society. They began with an assessment of 100 countries in both the developed and developing worlds and considered factors including GDP growth, GDP per capita, smartphone penetration and the number of unbanked.
The five chosen markets are the United States, Mexico, Nigeria, India and Indonesia. While there are characteristics unique to each of those regions, CredoLab’s solution is scalable and applicable to all because it accesses the same types of data in the same way, relying on the consistent patterns users display whether they are in Bangor or Bangalore.
Mr. Tucci believes CredoLab delivers the right value proposition for lenders impacted by the COVID-19 pandemic. In addition to the 35 million credit invisible Americans (sub-500 FICO scores), there are the millions more granted payment holidays, which means lenders cannot report customers to credit bureaus.
“This means credit bureau data in the United States is not reliable for the next six to nine months,” Mr. Tucci said.
Lenders therefore have a choice – lend only to the top 10 per cent and provide a horrible customer experience to the rest while seeing their customer acquisition costs skyrocket or consider other solutions, or develop better scoring methods.
CredoLab’s offering fits the bill as it has been tested in the parts of the world beginning to emerge from the pandemic and shown not to lose its predictive power. It was also tested before and during the pandemic and the predictive power of the digital scorecard did not change, Mr. Tucci reported.
In a rapidly changing environment solutions need to be implemented quickly, Mr. Tucci said. CredoLab can have a digital onboarding application with a cloud-based decision engine go live for a client in less than 60 days.
“This is the new normal,” Mr. Tucci said. “With more digital originations there will be more openness to testing new solutions.”