In emerging markets, credit providers usually build products separately from one another. For instance, a telco starts with airtime advances, then adds nano-lending, then installment credit. Each product tracks its own data and scores risk on its own terms. From the start, no one inside the organisation has a complete picture of any single customer.
That gap gets wider as a customer uses more products. Borrowers in certain emerging markets borrow multiple times from multiple providers. For example, borrowers in Ghana and Kenya utilise two to five different types of credit products on average.
This means a customer who borrows across three of the same provider’s products shows up three separate times, in three separate models. The provider is lending to its own customers without knowing how much it has already lent them.
A borrower can be dangerously overextended across three lenders and still look creditworthy to each one. No single lender can see what the others have lent.
Both problems come from the same place: credit decisions are made at the product level, not the customer level. Fixing that requires a different approach entirely.
What a customer-level approach means
A customer-level risk model looks at the full picture: how much a borrower owes, how reliably they repay, and what they can realistically afford. It does this across every product they use, all at once.
In practice, this means one platform that treats nano-lending, line of credit, and instalment products as parts of the same customer relationship, not separate transactions. One model. One view of the customer.
A product-level model asks: can this customer repay this product? A customer-level model asks a harder, more useful question: what is this customer’s actual credit capacity, and how much of it is already being used?
The customer-level approach replaces the product-level model entirely. It gets more accurate over time, because every product added to the platform generates more data about the same customer. Customers will also have better choices of what products they want to consume and the exposure they have for each.
Telcos and digital wallet providers: Turn customer visibility into risk and revenue advantage
Telcos that offer financial services carry risks that standalone lenders do not. When credit goes wrong at scale, across millions of subscribers, the damage affects more than the loan book. Customers leave. Regulators take notice. The telco’s core business takes the hit.
A customer-level risk model reduces that exposure. When a telco can see a customer’s total borrowing across all its products, it stops lending more than that customer can handle. That protects the loan book and the subscriber relationship at the same time.
That same visibility opens revenue opportunities that a product-level model skips over. A telco can see which customers are creditworthy but underserved: borrowers whose actual risk profile justifies more credit, but who are capped because the model only sees part of the picture. It can also time offers to when customers are in a good repayment position, rather than sending them at fixed intervals.
There is a growth benefit too. A telco with a single customer model can roll out new financial products without starting from scratch each time. Each new product is fit for purpose according to the needs of a customer, rather than creating a new one.
For a digital wallet provider, the same logic applies: the customer already has a relationship with the wallet. A shared risk model across credit products turns that relationship into something that compounds in value over time. Without it, every transaction stands alone.
Banks and fintechs: Gain an edge in emerging market lending
Some banks and fintechs have been slow to adopt micro-lending in emerging markets, treating it as a higher risk than it is worth. This might cause a loss of opportunities.
Research from the Global Emerging Markets Risk Database Consortium shows that the actual risk of lending in emerging markets is lower than most institutions assume. The opportunity is real. The main obstacle has been the lack of tools to assess it properly.
A customer-level model acts as an intra-credit bureau: it gives a bank, money-transfer operator, or fintech a complete picture of a borrower’s credit exposure, built from within their own platform. Most external credit bureaus in these markets cannot provide that level of detail. An institution that builds it internally gains a real underwriting edge.
For a bank, that changes the conversation about emerging market lending from one about managing downside risk to one about seizing the opportunity. For a fintech already active in payments or savings, it provides the foundation for moving into credit in a way that makes sense: one risk model, one customer, one platform.
The Ezra model: one platform, compounding intelligence
Ezra builds its platform around the customer-level model. A single platform runs nano-lending, line of credit, and airtime credit services for the same customer base, treating them as one credit relationship, not three separate products.
Every product added to the platform makes the model more accurate. Having more products means obtaining and utilising more data about the same customers, which sharpens risk assessment across the board. Adding a product to a siloed architecture adds exposure. Adding a product to Ezra’s platform adds intelligence.
For a partner institution, the platform becomes more valuable as it grows. For example, a client that starts with nano-lending and adds a line of credit gains both an additional product and a sharper model for both.
A crucial platform decision
Choosing a credit partner is a platform decision. Because customers use multiple products, a partner that can only see one product at a time will always have an incomplete picture of customer risk.
A partner that sees the full customer relationship can actually manage it and can offer lending services for customers effectively and sustainably.
To learn how Ezra provides a single view of customer risk, reach out to our team.
