- Behavioral and device metadata analytics innovator Credolab has unveiled its Income Prediction Model.
- The new offering will enable lenders to estimate applicant income using privacy-consented smartphone metadata. This will help them serve would-be borrowers with limited credit histories and proof-of-income.
- Founded in 2016, Credolab made its Finovate debut at FinovateAsia 2018 in Singapore. Peter Barcak is Co-Founder and CEO.
One of the biggest challenges for lenders seeking to expand into new markets—especially emerging, underbanked, and digital-first markets—is accessing accurate proof-of-income and credit history information. Even in a world in which open banking is embraced—making financial data more accessible overall—customers who have little data to share will remain on the outside, unable to benefit from a growing range of critical banking and financial services.
To meet this challenge, behavioral and device metadata analytics company Credolab has launched its Income Prediction Model. The new offering leverages machine learning to enable lenders to estimate applicant income by using privacy-consented smartphone metadata. The solution analyzes thousands of anonymized behavioral signals that, put together, correlate with income levels. These signals include app ownership patterns, device model and age, and interaction habits. Individual client institutions can train models on their own specific datasets and customize them based on the unique characteristics of their local populations. Importantly, Credolab’s Income Prediction Model never accesses personally identifiable information (PII) or demographic data like age, gender, or education.
Credolab uses proprietary feature engineering to convert raw metadata—collected with explicit user consent via its SDK—into more than 11 million behavioral features. The technology uses selection strategies based on information value, correlation filtering, and gradient boosting to narrow these features into a few dozen highly predictive indicators. The models use elastic-net logistic regression and tree-based ensemble techniques and validate them with out-of-time and out-of-sample testing to ensure both robustness and explainability.
“In many markets, a lack of verified income data is the biggest barrier to financial inclusion,” Credolab Co-founder and CEO Peter Barcak said. “Our new model gives lenders a privacy-safe and statistically sound way to infer income levels using only device behavior. It’s a powerful step toward fairer, faster, and more inclusive credit decisions, especially among populations for whom traditional data simply doesn’t exist.”
Founded in 2016 and headquartered in Singapore, Credolab made its Finovate debut at FinovateAsia 2018. Since then, the company has become the device and behavioral data partner for more than 150 banks, financial services companies, and fintechs around the world. The company’s solutions for risk management, fraud prevention, and insight-driven marketing have delivered decreases of up to 21.9% in the cost of risk and fraud, increases of up to 32% in applicant approval rates, and decreases of up to 28% in the cost of acquisition.
Photo by Christian Dubovan on Unsplash
Views: 99
