Buy Here Pay Later Charge Offs: Backend Readiness & AI Solutions

Buy Here Pay Later Charge Offs: Is Your Backend Ready for the BNPL Boom's Aftershock?

BNPL is no longer merely a pandemic-era story. The CFPB reports that U.S. BNPL—generally zero-interest loans repaid in four or fewer installments—has grown rapidly since 2019, while noting material limits in outside measurement of use and loan stacking. CFPB, “Consumer Use of Buy Now, Pay Later and Other Unsecured Debt” (Jan. 13, 2025; reviewed Oct. 8, 2026)

These issuer-reported Q3 figures have different scopes. Klarna reported Q3 2025 GMV of $32.7 billion, up 23% like-for-like. Affirm reported approximately 6.1 transactions per active consumer as of Sept. 30, 2025; it defines the measure as the average number of platform transactions by an active consumer during the preceding 12 months. Sezzle reported Q3 2025 GMV of $1.0 billion, up 58.7% year over year, and labeled its purchase-frequency measure 6.5x versus 5.4x in the prior comparable period. These are company metrics, not a common market-wide measure. Klarna Q3 2025 release (Nov. 18, 2025; reviewed Oct. 8, 2026); Affirm Form 10-Q, quarter ended Sept. 30, 2025 (filed Nov. 6, 2025; reviewed Oct. 8, 2026); Sezzle Q3 2025 results (Nov. 5, 2025; reviewed Oct. 8, 2026).

The front-end of the business is firing on all cylinders. But as I analyze these reports, I see a hidden challenge building on the backend.

The Hidden Challenge: Delinquency Lag

I see a backend visibility issue rather than a claim about what every scoring model can or cannot do. The CFPB found that BNPL lenders typically had not reported loans to nationwide consumer reporting companies in volumes large enough for external observers to measure persistence in use, cross-firm loan stacking, or total unsecured debt. Its study used a matched sample of applications and originations for U.S. pay-in-four loans from six large BNPL firms and corresponding de-identified credit records; that evidence does not establish a result for every BNPL product, lender, or scoring model. CFPB, “Consumer Use of Buy Now, Pay Later and Other Unsecured Debt” (Jan. 13, 2025; reviewed Oct. 8, 2026)

Here is my breakdown of the Q3 trends and why we built Debt Catalyst to solve the specific risks they expose.

1. The Shift to Long-Term Liability (Klarna)

Klarna reported that U.S. Fair Financing GMV rose 244% year over year in Q3 2025 and described the offering as transparent, fixed-term installments. The release does not supply the article’s “typically 6–12 month” range, so I do not treat that range as verified. Klarna Q3 2025 release (Nov. 18, 2025; reviewed Oct. 8, 2026)

The Risk: Moving from short-term installments to longer repayment schedules can increase exposure to changes in a borrower’s financial circumstances. Whether that risk materializes varies by the actual product terms, underwriting, and borrower facts.

The Debt Catalyst approach: Our team describes the Macro-Context Widget as a risk-review workflow that can incorporate macroeconomic inputs. I am not presenting independent validation that it forecasts defaults or reserve needs; lenders should assess model inputs, governance, and performance on their own portfolios before use.

2. The "Super User" & Debt Stacking (Affirm & Sezzle)

The issuer-reported engagement metrics above are not interchangeable: Affirm’s 6.1 figure is a 12-month average per active consumer, while Sezzle’s release labels its 6.5x figure “overall consumer purchase frequency.” The cited Sezzle release does not supply the article’s claimed top-10%/90-day cohort, so I do not treat that cohort claim as independently verified. Affirm Form 10-Q, quarter ended Sept. 30, 2025 (filed Nov. 6, 2025; reviewed Oct. 8, 2026); Sezzle Q3 2025 results (Nov. 5, 2025; reviewed Oct. 8, 2026).

The Risk: Multiple outstanding loans may be relevant to affordability, but the question requires product- and borrower-level analysis. The CFPB’s matched U.S. pay-in-four study was designed to measure cross-firm loan stacking because it was not observable from credit records at sufficient scale; I do not infer debt distress from transaction frequency alone. CFPB, “Consumer Use of Buy Now, Pay Later and Other Unsecured Debt” (Jan. 13, 2025; reviewed Oct. 8, 2026)

The Debt Catalyst approach: Our team describes Financial Durability and the Spending Power Index (SPI) as pre-charge-off segmentation concepts. I am not making an independently verified claim about their inputs, accuracy, ability to calculate discretionary income, or ability to predict delinquency; lenders should validate any model on their own population and under their own governance standards.

3. The Regulatory Pivot (Sezzle)

Sezzle’s Q3 2025 release says its corporate strategic project costs included professional services tied to “the exploration of pursuing a bank charter.” The release does not identify an Industrial Loan Company, say that Sezzle “hired consultants,” or set a 2026 application date. I therefore treat a charter as exploratory rather than settled. Sezzle Q3 2025 results (Nov. 5, 2025; reviewed Oct. 8, 2026)

The Risk: Moving from a fintech partner model to a chartered bank invites a level of scrutiny that most BNPL data infrastructures are not ready for. Regulators will demand ironclad data hygiene, perfect audit trails, and rigorous compliance regarding collections. Compliance with all applicable laws is essential for BNPL providers seeking a bank charter.

The Debt Catalyst approach: Our team describes the Autonomous Compliance Engine as a monitoring tool; I am not representing complete regulatory coverage or a “perfect” audit trail. Regulation F prescribes federal rules for FDCPA debt collectors. It does not itself create a blanket rule for original creditors, all BNPL providers, vendors, or state-law obligations; role- and jurisdiction-specific compliance requires qualified counsel’s review. CFPB, “12 CFR Part 1006 - Fair Debt Collection Practices Act (Regulation F)” (most recently amended Apr. 19, 2023; reviewed Oct. 8, 2026); CFPB, “What laws limit what debt collectors can say or do?” (last reviewed Dec. 9, 2024; reviewed Oct. 8, 2026)

4. The Scale of Data

Zip’s FY25 results reported U.S. total transaction volume growth of 41.6% in U.S. dollars to US$6.0 billion for the 12 months ended June 30, 2025. That is a different period and metric from the article’s unverified 51% claim. Zip Co, “FY25 Results Update” (Aug. 22, 2025; reviewed Oct. 8, 2026)

The Risk: Portfolio growth can increase operational workload, but an automation or collection decision should be tested against account-level economics, consumer-protection requirements, and applicable reporting practices. I am not making a general claim that a collection referral will affect a borrower’s credit score.

The Debt Catalyst approach: Our team describes the Settlement Sensitivity Score (SSS) as a portfolio-segmentation concept. I am not presenting independently verified evidence that it predicts settlement acceptance, determines the appropriate communication channel, or replaces human staff; a lender should validate results, controls, and compliance for its own portfolio.

The Verdict

Growth is great. Profitable growth is better.

As the market matures, the winners won't just be the apps with the best checkout flow. They will be the lenders who understand their borrower's Financial Health better than the competition.

I built Debt Catalyst for lenders evaluating high-volume risk and recovery workflows. I am not claiming independently verified revenue retention, underwriting accuracy, or outcome improvement; results depend on the portfolio, model, implementation, and applicable legal requirements.

Are you a lender looking to optimize your recovery strategy?

Related Strategic Briefings

• CRE Debt Portfolios: Refinancing Risk & Asset-Level Review — Review why commercial judgment portfolios can lose recovery value and how asset tracing may support collection efforts.
• Credit Scores and Cash-Flow Data: What the Evidence Supports — Learn why traditional credit scores may miss changing repayment capacity and how cash-flow signals can inform default analysis.
• Specialty Debt Solutions for Underserved Sellers — Explore portfolio-sale approaches for credit unions, BNPL lenders, auto finance, and other specialized debt sellers.
• Behavioral Debt Selling: Beyond Traditional Asset Classes — Explore how behavior-based segments can add detail beyond broad debt categories when assessing portfolios.

Discuss Portfolio Valuation & Liquidity

Looking to structure a portfolio sale or optimize recovery yield? Schedule a strategic call with Jeffery Hartman or test our Portfolio Recovery Valuation Tool.