A lender running a personal loan book decides to add a revolving credit facility. The ledger they use was built for fixed-term amortising loans: equal instalments, a known maturity date, a single interest rate. The revolving product does not have those properties. Balances move up and down. Draws and repayments happen at any time. Interest accrues daily on a changing principal. The platform bends, then breaks, then someone builds a parallel spreadsheet.
The problem is not that credit products are complicated. The problem is that each product type has a distinct financial mechanic, and a ledger that conflates them produces errors in accrual, arrears and reporting that compound silently until a regulator or an auditor finds them.
The main credit product types and what makes each one different
Each product type has a different balance mechanic, a different accrual rhythm and different regulatory obligations. A platform that treats them as variations on one template will misstate interest, misclassify arrears and produce incorrect Consumer Credit Act disclosures. The five types a consumer lender is most likely to run are these.
- Fixed-term instalment loan. A single advance repaid in equal scheduled payments over a defined term. Interest accrues on a reducing balance. The total charge for credit is known at origination, which is why the Annual Percentage Rate disclosure is straightforward. CONC 6.7 and the Consumer Credit Act 1974 set out how variation is treated if the rate is variable.
- Revolving credit facility. A credit limit that the borrower can draw from and repay repeatedly. Principal at any moment is the outstanding drawn balance. Interest accrues daily on that balance, not on the limit. Minimum payment rules, credit limit increase governance and persistent debt obligations under CONC 6.7.27R all apply.
- Buy now, pay later. Typically a deferred settlement product: no interest if the balance clears within the promotional window, a reversion rate if it does not. The ledger must track the promotional expiry date per transaction, not per account, and apply the correct accrual switch at the right moment.
- Mortgage or secured term loan. Similar instalment mechanic to an unsecured term loan but with a charge over an asset, different arrears treatment under MCOB, and a redemption calculation that must account for early repayment charges exactly as the product terms specify.
- Overdraft. A transactional credit product where the balance can move between positive and negative within a single account. FCA rules introduced in 2020 require a single annual interest rate with no fixed fees, and the arrears definition differs from that of an instalment product.
How a ledger should represent and govern multiple product types
The architectural answer is to separate the product definition from the account record. A product definition carries the rules: accrual basis, payment hierarchy, arrears definition, fee schedule and the jurisdiction rules that apply. The account record carries the state: balance, transactions, status. When a payment arrives or a rate changes, the ledger applies the product's rules to the account's state. That separation is what lets a single platform run an instalment loan and a revolving facility without conflating their mechanics. What a loan management system actually has to do sets out the broader requirements that follow from this.
- Define the product's financial mechanic precisely. Specify the accrual basis (actual/365, 30/360 or otherwise), the payment waterfall (fees, then interest, then principal, or another order), the minimum payment formula for revolving products, and the arrears trigger. These are not configuration options to be changed lightly; they are contractual terms that must match the credit agreement.
- Bind jurisdiction rules to the product, not the platform. A revolving facility offered in the UK sits under CONC. The same product offered in Germany sits under the Verbraucherkreditgesetz. The rules on persistent debt, rate caps and arrears notices differ. Encoding jurisdiction at the product level, rather than applying it as a platform-wide setting, is the only way to run a multi-market book without errors. Jurisdiction packs explains how that binding works in practice.
- Record every event, not just the current state. A payment that arrives three days late on an instalment loan changes the arrears status, the next payment due date and the total interest the borrower will pay. A ledger that overwrites state rather than appending events cannot reconstruct what the balance was at any prior moment. That matters for complaints, for Subject Access Requests and for past-due credit handling where the sequence of events is the evidence.
- Separate the accrual clock from the calendar clock. A buy now, pay later product that crosses a month-end, a rate change that takes effect mid-period, or a payment holiday all require the ledger to accrue interest against the contractual timeline, not simply against today's date. Why your loan ledger needs two clocks works through the mechanics of why this matters.
- Make the audit trail a first-class record. Every rule application, every state change and every agent action must be written immutably. Under Consumer Duty, a lender must be able to show that the product operated as the borrower was told it would. An event-sourced ledger provides that evidence without reconstruction.
A lender who can do all five for each product type in their book can answer a regulator's question about any account at any point in time, in any market they operate. That is the practical test of whether the platform is fit for the product range.
Why building each product on a shared template is not the answer
The usual workaround is to configure a general-purpose account object to approximate each product: turn off some fields, add a flag, write a side process to handle the exceptions. This works until the exceptions accumulate. A revolving facility built on an instalment loan template will misfire on minimum payment calculations, will apply the wrong arrears definition and will produce incorrect statements when the borrower makes an unscheduled repayment. The errors are not obvious until a borrower complains or an audit samples the book.
The alternative is a ledger where the product definition is a genuine first-class entity: it carries the rules, the account carries the state, and every transaction is the application of one to the other. Kontor is built on that model, which is also why migrating a live credit book onto it does not require a single cutover: product definitions migrate independently of account state, and the ledger can run both the old and new definitions in parallel during transition.
Credit products differ in their financial mechanics, their regulatory obligations and their audit requirements, and a platform that does not encode those differences precisely will produce errors that are hard to find and expensive to correct.