Tools
Last Updated on August 9, 2026
Last reviewed and updated: August 2026.
Buy now, pay later has stopped being a novelty and become plumbing. Your customer splits a purchase into four interest-free installments, you get paid up front minus a fee, and the provider carries the collection and default risk.
The scale is now well documented. Adobe Analytics recorded $20 billion in BNPL-funded online spend across the 2025 US holiday season (November 1 to December 31), up 9.8% year over year, with Cyber Monday becoming the first single day to cross $1 billion. A February 2026 economic brief from the Federal Reserve Bank of Richmond puts total 2025 US BNPL transaction value near $70 billion — roughly 1.1% of US credit-card purchase volume, growing about 20% a year.
So: real, mainstream, still small relative to cards. That framing matters, because it tells you BNPL is a conversion lever, not a payments strategy on its own.
Key takeaways
- BNPL reliably lifts conversion and average order value — but the honest, independently grounded AOV lift is closer to +17% than the +87% figures providers market.
- You are paid up front, so cash flow improves; you pay for it in fees that typically run 2–3x card processing.
- The regulatory picture reversed. The CFPB withdrew its 2024 rule treating BNPL like credit cards in May 2025 and confirmed it will not reissue it. Federal oversight is lighter; state-level scrutiny is not.
- Credit reporting is fragmenting: Affirm now furnishes data to Experian and TransUnion, while Klarna and Afterpay hold back their US pay-in-four data.
- Consumer stress is rising. Self-reported late payments climbed from 34% of users in 2024 to roughly 47% in 2026 — a reputational and returns risk you inherit at the counter.
How BNPL actually works at your checkout
The core US product is a closed-end, pay-in-four loan repaid over six to eight weeks with no interest to the shopper. Klarna, Afterpay, Affirm, and PayPal handle instant approval, collection, and default risk. You receive settlement quickly — usually within a few days — minus the provider’s fee.
The mechanism is simple psychology: you remove the sticker total as a barrier. A $240 cart reads as four payments of $60, and the hesitation that would have produced an abandoned cart produces an order instead.
Two structural details are worth knowing before you evaluate providers:
- Mobile is the channel. Adobe found 82.2% of holiday BNPL purchases happened on smartphones. If your mobile checkout is slow, the option won’t perform.
- The average ticket is modest. US BNPL transactions average around $135, which tells you where in your catalog the option belongs.
BNPL is one expression of a broader shift in which non-financial brands offer financial products directly. If you’re weighing pay-later against wallets, branded cards, or lending, our guide to embedded finance maps the wider landscape.
The upside: what the numbers actually support
Be careful with vendor benchmarks. Published AOV-lift claims for BNPL range from a conservative +17% to marketed figures of 45%, 91%, and merchant case studies claiming triple-digit gains. Only the low end is independently grounded; the rest is selection bias dressed as evidence — high-intent shoppers who would have converted anyway simply choose the pay-later button.
Use this as your planning range instead of provider marketing:
- Conversion lift: model a modest single-digit to low-double-digit improvement, concentrated on higher-ticket carts.
- AOV lift: plan on roughly +15–20%, and treat anything above that as upside you have to prove in your own data.
- New customers: the more defensible benefit. BNPL reaches shoppers who lack or avoid credit cards, which genuinely expands reach rather than reshuffling existing demand.
- Cash flow: unambiguously positive. You are funded up front while the provider carries the receivable.
Measure incrementality, not attribution
The single most common mistake is reading the BNPL dashboard as proof of value. That dashboard shows orders that used BNPL — not orders that only happened because of BNPL. Run a holdout: show the option to a traffic segment, withhold it from a matched segment, and compare total revenue per session across both.
To connect that experiment to real shopper behavior, pair it with behavioral analytics so you can see which segments respond and where the option merely cannibalizes card payments.
The trade-offs: fees, returns, and who owns the customer
Fees compress margin faster than you expect
BNPL acceptance fees commonly run about 1.5% to 7% per transaction, versus roughly 1.5% to 3.5% for card payments. On a 40% gross margin, an extra four points of payment cost eats a tenth of your margin on every BNPL order — which is why the AOV lift has to be real, not assumed.
| Factor | Cards | BNPL (pay-in-four) |
|---|---|---|
| Typical merchant fee | ~1.5–3.5% | ~1.5–7% |
| Settlement to you | 1–3 days | Up to a few days, in full |
| Default risk | Issuer | Provider |
| Dispute mechanics | Mature chargeback rails | Provider-specific, less standardized |
| Best-fit ticket size | Any | Roughly $50–$500 |
Returns get more expensive and more confusing
Reverse logistics is already a serious cost line in retail, and installment payments add a second layer: the refund has to unwind against a payment schedule that may be partly complete. A customer who has made two of four payments and returns the item expects a clean resolution, and they will call you first — regardless of who holds the loan.
Customers contact the merchant first for questions about installments, late fees, and refunds, even when the provider handles collections entirely.
Write the scripts before you launch, not after the first escalation.
Operational overhead is real but bounded
Expect sandbox testing, accreditation, and reconciliation work across a new settlement stream. Choosing a provider with a native plugin for your platform removes most of this. Custom gateway work does not.
Is BNPL a fit for your price points and customers?
The sweet spot for most merchants sits between roughly $50 and $500. Below that, the fee overwhelms the benefit and the installment framing is meaningless. Far above it, approval rates fall and default risk pushes providers to decline.
Adoption still skews young, and the categories are consistent: Adobe’s November 2025 consumer survey found shoppers most likely to reach for BNPL on electronics, apparel, toys, and furniture.
Practical filters before you switch it on:
- Set a minimum order threshold so low-ticket items never route to a high-fee method.
- Exclude categories with structurally high return rates.
- Show installment amounts on product pages, not just at checkout — that’s where the hesitation happens.
- Track conversion, returns, and repeat purchase rate by price band, not in aggregate.
If most of your revenue is recurring rather than transactional, the calculus is different — the friction you’re fighting is subscription fatigue, not sticker shock, and installments won’t solve it.
What changed in 2025–2026: regulation and credit reporting
This is the section most BNPL articles still get wrong, because the direction of travel reversed.
The CFPB rule was withdrawn
In May 2024, the CFPB issued an interpretive rule treating BNPL providers with “digital user accounts” as credit card issuers under the Truth in Lending Act and Regulation Z — bringing billing statements, dispute rights, and refund-crediting requirements with it.
That rule is gone. The CFPB withdrew it on May 12, 2025 as part of a broader rescission of guidance documents, and in June 2025 told a federal court it does not intend to reissue it, describing the original as procedurally defective and a poor fit for what are structurally closed-end loans.
What this means for you: the compliance burden you may have budgeted for in 2024 did not arrive. But federal restraint has pushed attention to state regulators and consumer advocates, so expect a patchwork rather than a settled field. Ask providers directly how they handle disputes and refund crediting, because there is no longer a federal baseline forcing consistency.
BNPL is entering credit files — unevenly
The more consequential shift is on the data side:
- Affirm began furnishing all pay-over-time products to Experian on April 1, 2025 and TransUnion on May 1, 2025.
- Klarna and Afterpay have declined to furnish US pay-in-four data, arguing that legacy scoring models would unfairly penalize their customers.
- FICO announced FICO Score 10 BNPL and 10 T BNPL on June 23, 2025, originally expected that fall. As of mid-2026 they are not yet in market at scale, and bureaus generally tag BNPL tradelines and exclude them from the scores most lenders read.
The practical read: the “BNPL now affects your credit score” headlines ran ahead of reality, but the infrastructure is being built. When it lands, loan stacking becomes visible to underwriters, approval rates may tighten, and your conversion lift could compress. Build that scenario into any multi-year forecast.
Watch the consumer-stress signal
Charge-off rates remain low at roughly 1.8–2%, but self-reported late payments have climbed sharply: about 34% of users in 2024, 41% in 2025, and roughly 47% in 2026. Most of those are only days late. Still, a payment method associated with financial strain carries brand risk, and placement matters — nudging a struggling customer into a fifth concurrent plan is not a growth strategy.
Implementing BNPL without breaking your margins
Run the breakeven math first
Before a sitewide rollout, build a simple model: incremental orders × incremental gross margin, minus the fee delta versus cards, minus expected additional return costs. If the answer only works at a +45% AOV assumption, you do not have a business case — you have a provider’s pitch deck.
If the fee delta is the problem, the fix is usually upstream in your pricing strategy framework rather than in the payment method itself.
Pilot narrowly, then expand
- Start with two or three categories in the $50–$500 band.
- Run for a full return cycle, not a fortnight — return costs surface late.
- Track support ticket volume as a first-class metric, not an afterthought.
- Compare at least two providers on approval rate for your audience; the spread between them is often larger than the fee spread.
Train the team before launch
Align finance and support on settlement timing, partial refunds, and reconciliation across payment methods. Your agents need one clear answer to “I returned it but Klarna still charged me” — and they need it on day one.
Review provider performance monthly: approval rates, effective fee rate, dispute volume, and satisfaction. Renegotiate as volume grows; fee tiers are far more negotiable than published rates suggest.
Conclusion
Treat BNPL as one lever among several, sized to your margin structure. The upside is real but smaller than the marketing implies, the fees are certain, and the regulatory environment that looked restrictive in 2024 has loosened federally while credit-reporting infrastructure quietly reshapes the risk picture.
Pilot it, measure incrementality with a holdout rather than a vendor dashboard, and set thresholds that keep low-margin items away from high-fee rails. Done that way, pay-later earns its place at your checkout. Bolted on because competitors have it, it mostly moves margin from you to a provider.
For the wider context on how payment and ownership models are reshaping retail, see our analysis of the recommerce trend.
