The Embedded Credit Opportunity
Embedded credit describes any lending product that is integrated into a non-financial platform's customer journey rather than offered as a standalone financial product. The e-commerce checkout that offers "buy now, pay later" is embedded credit. The accounting software that surfaces a working capital loan based on receivables data is embedded credit. The payroll platform that allows employees to access earned wages before payday is embedded credit. The marketplace that provides seller financing based on GMV data is embedded credit.
The appeal for platforms is clear: credit deepens the customer relationship, increases platform stickiness, creates a new revenue stream, and leverages the data advantage that platforms have over traditional lenders. A B2B marketplace that sees all transaction data for its sellers knows the creditworthiness of those sellers better than any bank. A payroll platform that sees monthly payroll runs can price an employee advance with near-zero default risk. The data moat creates a genuine pricing advantage over traditional lenders.
The challenges are equally real. Credit is regulated. Running a loan book requires capital. Managing credit losses requires expertise. Funding a growing loan book requires relationships with warehouse lenders that most platforms do not have. And the accounting for a credit portfolio under IFRS 9 is materially more complex than the revenue recognition for a SaaS subscription. None of these challenges are insurmountable, but they require serious preparation from the finance function.
The Three Embedded Credit Models
The choice of model determines both the regulatory requirements and the financial economics. There is no universally correct answer; the right model depends on the platform's capital position, risk appetite, and the importance of credit as a core product versus an ancillary feature.
Model 1: Balance Sheet Lending
The platform lends directly from its own capital. Every loan originated appears as an asset on the platform's balance sheet. The platform bears the full credit risk. In exchange, the platform earns the full net interest margin: the spread between the yield on the loan book and the cost of funding that book. This is the highest-margin model and the most capital-intensive. It requires regulatory authorisation and creates material financial statement complexity.
Model 2: Marketplace / Facilitation Model
The platform introduces borrowers to a licensed lender but does not itself take credit risk. The origination engine, the application journey, and the customer relationship sit with the platform; the actual lending sits with the lender. The platform earns a referral fee or an origination fee (typically 1 to 3% of the loan amount) and, in some arrangements, an ongoing trail commission. No credit risk sits on the platform's balance sheet. The regulatory requirements for this model are lower but not zero: facilitation of consumer credit may still require FCA authorisation depending on the specific activities carried out.
Model 3: White-Label Partnership
A licensed lender operates a credit product under the platform's brand. The platform provides the distribution channel and potentially some data for credit decisioning; the lender provides the balance sheet, the regulatory permissions, and the credit expertise. Revenue sharing between the two parties is negotiated commercially. This model allows a platform to launch a credit product quickly without building the regulatory and operational infrastructure of a lender, but it creates dependency on the lending partner and typically generates lower margin than balance-sheet lending.
Regulatory Requirements by Model
The regulatory framework for consumer credit in the UK sits primarily under the Financial Services and Markets Act 2000 (FSMA 2000) and the Consumer Credit Act 1974 (CCA 1974). The FCA's CONC Sourcebook contains the conduct rules for consumer credit firms. Understanding which authorisation permissions are required is the most critical early step for any platform considering embedded credit.
For balance-sheet lending to consumers, a full consumer credit licence under Part 4A of FSMA is required. The relevant regulated activities include: credit broking, operating an electronic system in relation to lending, and providing credit. The authorisation process involves demonstrating adequate capital, fit and proper management, a compliant credit risk framework, and appropriate affordability assessment processes. The Senior Managers and Certification Regime (SMCR) applies, meaning key individuals must be individually approved by the FCA.
For the marketplace model, even though the platform does not take credit risk, credit broking activity — introducing consumers to lenders — is a regulated activity under FSMA. A platform that introduces consumers to lenders in the course of business will typically require authorisation for credit broking. There are limited appointed representative structures available, but these require an existing authorised firm to act as principal.
The Financial Model for Balance-Sheet Lending
The economics of a balance-sheet lending book are best modelled using a contribution margin framework that captures each stage of value creation and destruction. The key variables are origination volume, approval rate, average loan size, yield on the book, cost of funds, credit loss rate, and operational cost per loan.
Consider a worked illustration: a B2B marketplace that originates £500,000 of SME loans per month. With an average loan size of £10,000, this represents 50 new loans per month or approximately 600 per year. If the annualised yield on the book is 18% and the cost of funds through a warehouse facility is SONIA plus 400 basis points (approximately 9% at mid-2024 rates), the gross net interest margin is approximately 9%. Against a credit loss rate of 3% (expressed as a percentage of the average book balance) and operational costs of £200 per loan (credit assessment, servicing, collections), the net contribution per loan per year is approximately £520 on a £10,000 loan.
This illustration makes the economics look attractive. The complexity lies in the assumptions: credit losses on SME books are highly cyclical and can spike materially in a downturn; the cost of funds through a warehouse facility increases if the portfolio quality deteriorates; operational costs per loan escalate as collections become more complex; and the model requires significant upfront capital to fund the book to the scale at which it becomes economic.
"A loan book is a fundamentally different kind of asset from a SaaS subscription book. It earns yield every month, but it also deteriorates every month as credit losses are realised. Getting the provisioning wrong — in either direction — distorts every performance metric the business reports."
Funding the Loan Book
A balance-sheet lending business needs external funding to grow beyond what equity alone can support. The standard institutional funding structure for a growing fintech lender is a warehouse facility: a revolving credit line from a bank or specialist credit fund, secured on the loan portfolio.
Warehouse facility terms in 2024 typically involve an advance rate of 80 to 90% of the eligible loan portfolio value, interest at SONIA plus 300 to 500 basis points depending on portfolio quality and lender type, a minimum portfolio quality covenant (maximum default rate, weighted average loan term, concentration limits), and a borrowing base certificate requirement (typically monthly or more frequent). The facility draws down as loans are originated and repays as borrowers repay.
For larger loan books (typically above £20 to 30 million), securitisation becomes an option. A securitisation structure issues notes backed by the loan pool to capital market investors, typically at a lower cost of funds than a bilateral warehouse facility and without the concentration risk of relying on a single funder. Securitisation requires substantially more legal and administrative infrastructure, is typically only economic above £50 million of portfolio size, and requires an established track record.
IFRS 9 for the Loan Book
IFRS 9 Financial Instruments replaced IAS 39 and introduced the expected credit loss (ECL) model for loan book provisioning. For any business carrying a loan portfolio, understanding and implementing IFRS 9 is non-negotiable — it directly affects reported profit and, via regulatory capital requirements, the business's ability to grow.
Under IFRS 9, credit loss provisions must be recognised from the moment a loan is originated, not only when a loss event actually occurs. This is the key departure from the incurred loss model under IAS 39. On initial recognition, a 12-month ECL provision is required for performing loans (Stage 1). If the credit risk of a loan increases significantly relative to origination, the loan moves to Stage 2 and a lifetime ECL provision is required. Loans that are credit-impaired move to Stage 3.
The practical implication for a growing loan book is that provisioning is a front-loaded cost. As the book grows, the Stage 1 provision charge grows with it. In the early months of a lending business, the provision charge can be significant relative to the interest income earned on a still-small book, creating an initial period of reported losses even on a genuinely sound portfolio. This is not a sign of business failure; it is a mechanical consequence of IFRS 9's forward-looking provisioning approach, and it must be explained clearly in board reporting and investor communications.
Key Takeaways
- Embedded credit takes three primary forms: balance-sheet lending (full credit risk, highest margin), marketplace facilitation (no credit risk, origination fee), and white-label partnership (revenue share, no balance sheet). The right choice depends on capital, risk appetite, and strategic importance of credit.
- All three models may require FCA authorisation. Credit broking is a regulated activity even where no credit risk is taken. Get regulatory advice before launch, not after.
- CONC applies to lending to consumers, sole traders, and small partnerships. Lending to limited companies is regulated differently, which makes B2B embedded credit materially less complex from a conduct perspective.
- The financial model for balance-sheet lending hinges on net interest margin minus credit losses minus operational cost. All three variables are harder to predict than they appear at the modelling stage.
- Warehouse facilities fund loan growth, typically at 80 to 90% advance rate and SONIA plus 300 to 500 basis points. The facility terms include quality covenants that limit operational flexibility.
- IFRS 9 requires Day 1 expected credit loss provisioning. Growing loan books generate front-loaded provision charges that depress reported profit even on sound portfolios. This must be explained clearly to boards and investors.
- Embedded credit is not a simple feature addition. It requires regulatory infrastructure, credit risk expertise, funding relationships, and a finance function capable of managing a complex loan book. Budget accordingly.
