I Tried to Compare Ecommerce Lending Data. It Broke My Brain.
Measuring Brand Stress
Shopify Capital, Amazon Lending, Wayflyer, BNPL, billions in credit flowing through commerce platforms, and almost none of it is measured the same way.
I started with what felt like a reasonable question.
Ecommerce platforms are lending serious money to merchants and consumers. Shopify Capital pushed $4.2 billion out the door in 2025 alone. Wayflyer has deployed over $6 billion since 2020. Buy Now Pay Later hit an estimated $70 billion in transaction value last year. Amazon was lending over a billion a year before they quietly killed the programme.
So: what does the credit quality look like? Are defaults rising? Is this a growing systemic risk, or a well-managed feature of modern commerce? Surely someone has pulled this data together.
They haven’t. And when you try, you discover why.
The Hypothesis
The idea was simple. Take the major ecommerce-adjacent lending programmes, Shopify Capital (merchant lending), Amazon Lending (seller financing), Wayflyer (revenue-based financing for DTC brands), and the BNPL providers (consumer credit at checkout), and chart the trend lines over the last decade. Origination volumes, delinquency rates, defaults. Put them side by side. See what patterns emerge.
What I found instead is that these datasets exist in entirely different universes. Different borrower types, different disclosure regimes, different metrics, different regulatory oversight. Comparing them directly would be like comparing a mortgage arrears rate with a credit card charge-off rate and a trade finance loss, technically they’re all “credit”, but the comparison tells you almost nothing.
That, in itself, is the story.
What the Data Actually Shows
Shopify Capital: The Best-Disclosed, Fastest-Growing
Shopify Capital is the most transparent of the lot, because Shopify is SEC-reporting and discloses detailed loan ageing data in its 10-K filings.
Launched in April 2016 as a small cash advance product for US merchants, it’s grown into a genuine lending operation across 8 countries. The trajectory is remarkable:
The 2022 dip reflects a model transition, Shopify shifted from primarily merchant cash advances to a mix of MCAs and term loans originated by bank partners and purchased by Shopify. From 2023, originations accelerated dramatically.
Here’s where it gets interesting. Shopify actually publishes delinquency ageing, which almost nobody else does:
By end-2025, the proportion of deals considered “current” across the full portfolio had dropped further to 91.9%, down from 93.7% a year earlier.
So: originations are surging. Credit quality is softening at the margins. The allowance ratio is holding steady, meaning Shopify’s risk team is provisioning in line with what they’re seeing. But the trajectory of 90+ day delinquencies, from 3.9% to 5.3% in a single year, is worth watching.
The question is whether this is a normalisation (bigger portfolio = more tail risk) or a signal that the quality of merchants being funded is declining as the programme scales.
Amazon Lending: The One That Died
Amazon’s lending programme is a fascinating case study in platform economics, and in how little a tech giant needs to disclose about financial services when it’s a rounding error on the P&L.
Launched in 2011, Amazon Lending was invitation-only, offering term loans of $1,000 to $750,000 to FBA sellers. Repayments were automatically deducted from seller disbursements, a structurally elegant collection mechanism.
The programme grew steadily, then crashed during COVID when Amazon pulled back hard, cutting receivables from $863M to $381M. It rebuilt post-pandemic but never returned to aggressive growth. By 2023, receivables were flat at around $1.2-1.3 billion.
Then, in March 2024, Amazon killed the programme entirely, shifting to third-party providers (Parafin, Lendistry, SellersFi).
Amazon never published delinquency data. The only credit quality figure that leaked was a 1.34% expected loss rate for 2023, referenced in internal documents obtained by Business Insider. That’s remarkably low, but impossible to verify independently, and the programme no longer exists.
The interesting question: why did Amazon exit? At $1.3B in receivables against $514B in net sales, it wasn’t a capital allocation issue. More likely, the regulatory overhead of being a direct lender wasn’t worth it when third parties would happily take the risk in exchange for access to Amazon’s data. Amazon now brokers the relationships without the balance sheet exposure. The seller data stays with Amazon. The credit risk moves to someone else. Classic platform move.
Wayflyer: The Irish Unicorn and the Concentration Risk Problem
Wayflyer is the most opaque of the four, and also the most instructive about the risks specific to revenue-based ecommerce financing.
Founded in Dublin in 2020, Wayflyer provides unsecured working capital to DTC and consumer brands, funding inventory and marketing spend, with repayment tied to revenue. They hit unicorn status in February 2022 with a $1.6B valuation, deployed $6 billion cumulatively by early 2026, and recently secured a $250M credit facility from ATLAS SP Partners (Apollo-backed).
Because Wayflyer is privately held and Irish-registered, the data comes from CRO (Companies Registration Office) annual accounts. These are abbreviated compared to SEC filings, you get aggregate financials, not portfolio-level breakdowns.
—
$250M ATLAS facility
That 2024 line is the one that jumps out. A €41.2 million credit impairment from what Wayflyer described as “an isolated customer default.” That single loss was roughly half of Wayflyer’s net assets at the time.
This is the concentration risk inherent in revenue-based financing for larger DTC brands. When you’re writing bigger tickets to fewer, higher-growth businesses, rather than small MCAs to thousands of Shopify merchants, one blow-up can be material. Wayflyer said they “strengthened a number of controls” as a result, and claimed that excluding this event, impairments were in line with forecasts.
But we don’t know: who the borrower was, what sector they were in, whether the default was fraud or simply a business that collapsed, or what Wayflyer’s overall portfolio loss rate looks like. The CRO accounts don’t require that level of disclosure.
BNPL: The Consumer Side of the Equation
Buy Now Pay Later is the best-documented segment, thanks to the CFPB’s market monitoring of six large lenders (Affirm, Klarna, Afterpay/Block, PayPal, Zip, and others).
The growth story is extraordinary, from $2.2B in originations in 2019 to an estimated $70B market by 2025. But this is consumer credit, not merchant working capital, which makes it fundamentally different from Shopify/Amazon/Wayflyer.
The charge-off data is counterintuitive. Despite the massive growth, charge-off rates actually declined from 2.63% in 2022 to 1.83% in 2023, well below the 4.19% credit card charge-off rate. The CFPB attributes this to tighter underwriting and a strategic shift toward deepening usage among existing customers rather than acquiring new (and riskier) ones.
But survey data tells a different story at the consumer level. LendingTree’s 2025 survey found that 41% of BNPL users made at least one late payment in the past year, up from 34% a year earlier. There’s a divergence opening up between aggregate provider-level charge-offs (improving) and individual consumer payment behaviour (worsening).
Provider-specific data adds nuance:
Klarna: Credit loss rate of 0.54% of payment volumes in Q1 2025, up from 0.51%. But their “Fair Financing” receivables (longer-term products) show an expected credit loss allowance of 5.9%, up from 4.2%, a meaningful deterioration in the book that looks most like traditional lending.
Affirm: Provision for credit losses of $171M in the six months to December 2022, up 47% YoY.
Meanwhile, broader consumer credit is showing stress. US credit card accounts 90+ days delinquent hit 12.70% by Q4 2025, well above the long-term average of 9.16%. BNPL may be performing better than credit cards, but it doesn’t exist in a vacuum.
Why This Data Doesn’t Compare
This is where the initial hypothesis fell apart. Here’s the disclosure matrix:
You’re looking at four fundamentally different credit products, aimed at different borrowers, operating under different regulatory frameworks, and measured with different yardsticks.
The borrower problem. Shopify Capital and Amazon lend to their own platform merchants. Wayflyer lends to platform-agnostic DTC brands. BNPL lends to consumers. A Shopify merchant defaulting on a $50K cash advance is a completely different credit event from a consumer missing a $131 BNPL payment, which is different again from a DTC brand collapsing on €41M of Wayflyer financing.
The disclosure problem. Shopify is SEC-reporting and gives you ageing tables. Wayflyer files abbreviated Irish CRO accounts with aggregate impairment only. Amazon never disclosed credit quality. BNPL data comes from a CFPB aggregate of six firms, useful, but you can’t see individual provider portfolios in comparable detail.
The structural problem. These products have entirely different collection mechanisms. Shopify deducts repayments from future sales through its platform. Amazon did the same through seller disbursements. These are structurally secured, the platform controls the cash flow. Wayflyer ties repayment to revenue but doesn’t control the payment rail. BNPL charges the consumer’s debit or credit card. The structural protection levels are wildly different, which means the same “delinquency rate” means different things in each case.
The regulatory problem. BNPL providers are increasingly subject to consumer credit regulation. Shopify Capital operates as a purchaser of bank-originated loans. Wayflyer offers unsecured business lending. Amazon brokered to third parties. None of these face the same reporting requirements, which is why the data looks the way it does.
What to Watch From Here
If you can’t compare these directly, what can you monitor? Here are the metrics that matter going forward:
1. Shopify’s 90+ Day Delinquency Rate
This is the single best real-time indicator of merchant credit stress in ecommerce. It’s disclosed quarterly in SEC filings, it covers a large and growing portfolio ($1.4B in receivables), and it’s already showing a clear direction, from 3.9% to 5.3% in one year. If this crosses 7-8%, this is said to be a genuine signal that the merchant base is under pressure.
2. Will Capital Costs Rise?
Shopify, Wayflyer, and the BNPL providers all depend on wholesale funding, credit facilities, securitisations, and bank partnerships. If base rates stay elevated or credit spreads widen, the cost of capital flowing into these programmes increases. That cost either gets absorbed (lower margins) or passed through (higher fees to merchants/consumers). Watch for facility renegotiations and any changes to merchant pricing. This has been saving the UK retail landscape for years.
3. The Collection Mechanism Advantage
Shopify and the late Amazon programme had a structural edge: they controlled the payment rail. Repayments were deducted before the merchant saw the money. That’s why Shopify’s loss rates are relatively contained even as delinquency ticks up, the cash advance gets repaid from every sale, automatically. Wayflyer doesn’t have this luxury. Neither do BNPL providers (who rely on consumer cards). As credit conditions tighten, expect the platforms with embedded collection mechanisms to outperform.
4. Wayflyer’s Next CRO Filing
The 2025 accounts (likely filed late 2026) will reveal whether the €41.2M impairment was truly isolated or the beginning of a pattern. Revenue-based financing for DTC brands is more concentrated and more exposed to consumer discretionary spending than Shopify’s long-tail merchant base. If another large default surfaces, the model’s risk profile needs reassessing.
5. Does Anyone Actually Care?
This might be the most honest question. Ecommerce lending is growing fast, $4.2B from Shopify alone in 2025, $6B cumulative from Wayflyer, $70B in BNPL, but it’s still dwarfed by traditional bank lending and credit card debt ($1.23 trillion outstanding). The Richmond Fed explicitly noted that BNPL’s impact on financial stability “appears limited at present.”
But “limited at present” is what people said about subprime mortgages in 2005. The pattern to watch isn’t the absolute size, it’s the rate of growth combined with the opacity of disclosure. When Shopify is originating $4.2B/year with only abbreviated ageing tables, Wayflyer is deploying billions with CRO-level disclosure, and BNPL charge-offs are measured one way by providers and another way by consumer surveys, the data infrastructure hasn’t caught up with the lending infrastructure.
And that gap is where risk hides.
Sources
Shopify 10-K 2024 (SEC filing, shop-20241231)
deBanked: Shopify Capital tracking (2021-2026)
Amazon 10-K filings; deBanked: Amazon Seller Lending Receivables tracking (2016-2024)
Marketplace Pulse: “Amazon Is Lending $1 Billion a Year” (2019)
Business Insider: Amazon Lending internal documents (Jan 2023)
deBanked: “Amazon Discontinues Its In-House Business Loans” (Mar 2024)
Wayflyer CRO annual accounts (2021-2024) via RTÉ, Irish Times, Business Post, The Times
Wayflyer press releases (2025-2026)
CFPB: “The Buy Now, Pay Later Market” (Dec 2025)
Richmond Fed Economic Brief: “Buy Now, Pay Later: Recent Developments and Implications” (Feb 2026)
YCharts: US Credit Card Delinquency data
Klarna 20-F (SEC filing, 2025)
Affirm 10-K/10-Q filings
LendingTree BNPL survey (2025)
The disclosure gap between Shopify's ageing tables and Wayflyer's CRO accounts is the kind of thing that makes portfolio analysis feel like guesswork. You can see the signal, but you can't compare it to anything else.