Your database holds thousands of declined applicants. Many deserved a “yes.” The problem was never their financial behavior. The problem was your data sources.
Traditional credit scores miss the full picture. A 2024 Nova Credit survey found that 90% of lenders believe access to alternative data would help them approve more worthy borrowers. That same survey revealed 31% of lenders feel traditional credit reports don’t capture a complete view of consumer finances.
This article breaks down how alternative data is transforming lending, what credit data points actually matter, and why the evolution of underwriting processes will determine which lenders thrive in the next decade.
What Is Alternative Data and Why Does It Matter for Lenders?
Alternative data refers to any information used in credit decisions that falls outside traditional credit bureau reports. Think bank transaction history, utility payments, rent records, employment verification, and real-time cash flow patterns.
Traditional credit relies on a narrow slice of financial life. Did you pay your credit cards on time? How much debt do you carry? How long have your accounts been open? These questions work fine for borrowers with well-established credit histories. They fail everyone else.
Roughly 100 million American adults are considered unscorable, credit invisible, or subprime. Many of these consumers pay rent on time, maintain stable employment, and manage their money responsibly. Traditional scoring just can’t see it.
Non-traditional data sources provide the insight lenders need to reach these borrowers without taking on additional risk.
How Are Examples of Alternative Credit Data Used in Underwriting Today?
Fintech firms pioneered the use of non-traditional data in credit decisions. Now, mainstream financial institutions are catching up.
Cash flow analysis tops the list of alternative credit data points gaining traction. By examining bank account transactions, lenders can verify income stability, track spending patterns, and identify how borrowers manage their balance over time. A borrower with consistent deposits and responsible spending habits may qualify even with a thin credit file.
Rental payment history offers another window into financial responsibility. Someone who pays $1,500 monthly rent on time demonstrates the same discipline required for loan repayment. Yet traditional scores ignore this entirely.
Utility and telecom payments work similarly. Regular, on-time payments for electricity, gas, and phone service signal reliability. Credit bureaus like Experian now incorporate some of this data into expanded scoring models.
Employment and income verification through payroll connections gives lenders real-time insight into a borrower’s earning capacity. Rather than relying on static pay stubs, consumer-permissioned data links provide up-to-date information about job tenure and income trajectory.
What Does the Evolution of Underwriting Processes Look Like?
The shift from static credit files to dynamic, data-driven underwriting represents one of the most significant changes in lending since FICO scores emerged in the 1980s.
Traditional underwriting treated credit decisions as snapshots. Pull a bureau report, check the score, apply cutoffs, and make a decision. This process worked when most Americans fit neatly into credit bureau categories.
Modern lending requires a different approach. AI models trained on diverse data sources can identify patterns that human underwriters would miss. Machine learning algorithms process thousands of variables to predict default risk with greater precision than three-digit scores alone.
The regulatory landscape is shifting, too. The CFPB’s Section 1033 rule promises to streamline how consumers share their financial data with lenders. When borrowers can easily grant permission access to their bank accounts, the friction around collecting non-traditional data drops significantly. Three-quarters of lenders surveyed expect this rule to accelerate adoption.
How Does Cash Flow Underwriting Empower Smarter Lending Decisions?
Cash flow underwriting represents the most practical application of these insights for most lenders.
The concept is straightforward. Instead of asking “what does your credit history look like,” you ask “can you actually afford this payment based on what’s happening in your bank account right now?”
Real-time transaction data reveals patterns that credit reports cannot capture. A borrower might show a 650 score due to past issues that no longer reflect their current situation. Their cash flow tells a different story. Steady income, manageable expenses, consistent savings deposits. The score lags reality by months or years.
Empower Finance built its entire model around this insight. The company underwrites individuals using real-time cash flow, nontraditional data, and machine learning to assess credit risk differently than mainstream lenders. Their 2024 acquisition of Petal Card demonstrates how the industry is consolidating around cash flow underwriting as a modern lending core.
Warren Hogarth, CEO of Empower, described the rationale behind the Petal acquisition: “In both companies, we found a shared commitment to harnessing technology and rich alternative data to unlock financial opportunity for more people who merit our consideration.”
What Insights Into the Shifting Focus From Debit to Credit Products Tell Us?
The fintech landscape has seen a notable shift from debit to credit products over the past two years.
Early neobanks focused on debit cards and spending accounts. The regulatory burden was lighter, the unit economics seemed simpler, and venture capital rewarded user growth over profitability.
That changed as interest rates rose and investors demanded sustainable business models. Credit products generate meaningful revenue. Debit products largely do not.
Petal Card and the future of credit for thin-file consumers exemplify this shift. The company built its reputation offering Visa credit cards to consumers without traditional credit scores. Nearly 400,000 consumers were approved using open banking data rather than bureau scores, with the majority having thin or no credit files at the time of approval.
The Empower acquisition of Petal signals where the market is heading. Debit-focused fintechs are adding credit products. Credit-focused fintechs are expanding their data capabilities. The winners will be those who combine both product breadth and underwriting sophistication.
How Do Financial Institutions Use Alternative Data to Expand Access to Credit?
Banks pull roughly 800 million to 900 million alternative credit consumer reports annually in the US for underwriting purposes, according to industry estimates. This excludes uses for marketing and account servicing.
The adoption of non-traditional data varies by institution type. Large banks represent the most significant users. Fintech lenders and buy now, pay later providers also heavily incorporate these signals.
Credit unions have adopted these approaches to a lesser extent, often because they perceive the analytical and technological barriers as too high. But usage is growing as turnkey solutions become more accessible.
A holistic view of borrower finances enables lenders to expand access to credit without proportionally increasing risk. When you can see that a declined applicant actually manages money responsibly, you can structure a loan that works for both parties.
The goal is not to lend recklessly. The goal is to lend accurately. These data sources empower lenders to distinguish between consumers who look risky on paper and consumers who actually are risky.
What Are the Regulatory Considerations for Using Alternative Data?
Lenders must navigate compliance requirements when incorporating new data sources into credit decisions.
Fair lending laws apply regardless of what data you use. The Equal Credit Opportunity Act prohibits discrimination on the basis of a protected class. If your new data sources correlate with race, national origin, or other protected characteristics, your model could expose you to disparate impact liability.
A 2021 Government Accountability Office study found that 13 of 16 mortgage lenders interviewed cited fair lending concerns as a barrier to adopting non-traditional data.
The CFPB has taken an active interest in this space. Their analysis of Upstart’s model found higher acceptance rates and lower costs relative to traditional approaches. But they also flagged potential fair lending concerns, leading to ongoing scrutiny of AI-based underwriting.
Lenders need explainability. When a borrower asks why they were declined, “the algorithm said so” is not an acceptable answer. Financial institutions must demonstrate that their models produce decisions they can explain and defend.
How Can AI and Analytics Reshape Lending Risk Management?
AI models trained on diverse data sources can process signals that traditional scorecards cannot evaluate.
Consider income volatility. A gig worker might earn $5,000 one month and $2,000 the next. Traditional underwriting struggles with this pattern. AI models can assess whether the borrower maintains adequate reserves during low-income periods, whether their spending adjusts appropriately, and whether the volatility trend is improving or worsening.
Behavioral signals matter too. How quickly does a borrower spend incoming deposits? Do they pay bills before or after due dates? Do they maintain consistent savings behavior? These patterns, invisible to credit bureaus, become visible through transaction analytics.
The fintech firms leading in this space have demonstrated that these approaches work. Upstart, Petal, Empower, and others built their businesses on the premise that modern data and machine learning could outperform FICO in predicting default. Their growth supports the hypothesis.
What Does the Future of Lending Look Like for Financial Institutions?
The future of lending belongs to institutions that combine traditional credit data with non-traditional signals in a compliant, explainable framework.
Pure fintech firms have proven that the technology works. Banks and credit unions have distribution, capital, and regulatory relationships that fintechs often lack. The most successful players will bridge both worlds.
Deeper insights into borrower behavior enable better customer experience throughout the loan lifecycle. When you understand a borrower’s full financial picture, you can offer appropriate products, set realistic terms, and provide proactive support when circumstances change.
The reshaping of the lending landscape creates opportunities for institutions willing to invest in data capabilities. Those who continue relying solely on credit bureau scores will find themselves outcompeted by lenders who see more of the picture.
How Can Lenders Get Started With Alternative Data?
Starting small makes sense for most institutions.
Cash flow data requires the least infrastructure change and delivers immediate value. Open banking providers like Plaid offer standardized integrations that package transaction data into scores or attributes lenders can incorporate alongside existing models.
Insights allow lenders to enhance decision-making without replacing proven approaches. You don’t need to abandon FICO. You need to supplement it with signals that capture what FICO misses.
Service providers in this space range from broad data aggregators to specialized analytics firms. Some, like TrackStar, focus specifically on identifying credit errors and hidden opportunities within existing applicant databases. Their Revelar solution uses machine learning trained on tens of millions of dispute outcomes to find the 720s hiding within the 650s, turning declined applicants into qualified borrowers.
Lenders must evaluate vendors carefully. Ask about regulatory compliance, model explainability, and integration requirements. The right partner makes adoption straightforward. The wrong one creates risk.
Frequently Asked Questions
What counts as alternative data in lending?
Alternative data includes any credit-relevant information outside traditional bureau reports. Common sources include bank transaction history, utility and rent payments, employment verification through payroll connections, and real-time cash flow patterns. The key distinction is that these data points reveal financial behavior that credit bureaus don’t capture.
How does cash flow underwriting differ from traditional credit scoring?
Traditional scores look backward at payment history on existing credit accounts. Cash flow underwriting examines current bank transactions to assess whether a borrower can actually afford loan payments based on income, expenses, and liquidity patterns. This real-time view often tells a different story than static bureau scores.
Is alternative data legally permissible for credit decisions?
Yes, but with important constraints. The Equal Credit Opportunity Act and Fair Credit Reporting Act apply regardless of what data you use. Lenders must ensure their models don’t produce discriminatory outcomes, even unintentionally. The CFPB has scrutinized several AI-based underwriting models for fair lending compliance.
What regulatory changes are driving the adoption of alternative data?
The CFPB’s Section 1033 rule promises to streamline consumer-permissioned data sharing. When borrowers can easily authorize access to their bank accounts, the friction around collecting cash flow data drops significantly. Three-quarters of lenders surveyed expect this rule to accelerate adoption across the industry.
How do lenders get started with alternative data?
Most institutions start with cash flow data because it requires the least infrastructure change and delivers immediate value. Open banking providers offer standardized integrations that package transaction data into scores or attributes lenders can incorporate alongside existing FICO-based models without full system replacement.
What are the main barriers to adoption?
Fair lending compliance concerns top the list. A GAO study found 13 of 16 mortgage lenders cited fair lending as a barrier. Technical integration complexity and model explainability requirements also slow adoption, particularly for smaller institutions without dedicated data science teams.
Can alternative data help lenders find opportunities in existing databases?
Yes. Predictive technology can identify which previously declined applicants likely have correctable credit issues or whose current financial behavior doesn’t match their bureau scores. This approach converts leads you’ve already paid for rather than requiring new customer acquisition.
How does Revelar fit into the alternative data landscape?
Revelar focuses on a specific application: identifying credit errors and hidden opportunities within existing applicant databases. Using machine learning trained on tens of millions of dispute outcomes, the solution finds borrowers whose true creditworthiness is masked by correctable errors. It’s not replacing your underwriting model. It’s surfacing qualified applicants that your current model missed.
Key Takeaways
- 90% of lenders believe non-traditional data sources would help them approve more worthy borrowers who currently get declined
- Cash flow underwriting examines real-time bank transactions to assess actual repayment capacity.
- Roughly 100 million American adults are unscorable or credit invisible under traditional models.
- AI and machine learning enable lenders to process non-traditional signals that bureau scores cannot capture.
- Regulatory compliance requires explainability and fair lending testing when using new data sources.
- The Empower acquisition of Petal Card demonstrates industry consolidation around cash flow underwriting.
- Financial institutions pull 800-900 million alternative credit reports annually for underwriting.
- Successful adoption starts with supplementing existing models rather than replacing them.
- Predictive technology can identify which declined applicants likely have correctable issues affecting their scores.
Ready to Unlock Hidden Borrowers in Your Existing Database?
Alternative data isn’t just about future applicants. It’s about the qualified borrowers already sitting in your decline file.
TrackStar’s Revelar solution uses 30 million data points from 20 years of dispute outcomes to identify which applicants in your database likely have correctable credit errors. Stop paying lead aggregators for new prospects when qualified borrowers already exist in your CRM.
Book a discovery call today to see how Revelar can turn your existing data into new lending opportunities.