She has been coming to your shop for two years. She pays — not always on time, not always in full, but she pays. You trust her. The question is whether your system trusts her the same way you do.
Credit in Kenyan retail is not a gap in a system. It is a feature of the relationship between a shop and its neighbourhood. Regular customers buy on credit. They come back. They pay what they owe and buy again. The shops that refuse credit lose regulars. The shops that extend it carelessly lose money. The ones that manage it properly keep both.
Managing it properly means one thing: knowing the current balance of every credit customer at any point in time, without having to add up a column of figures in a notebook or ask your staff what they remember.
The notebook approach to credit has a specific failure mode. It works when the same person opens and closes every sale and knows every customer. It starts failing when staff rotates, when a second person starts handling sales, when a customer's account spans multiple notebooks, or when a customer disputes an amount and there is no clean record to refer to.
A credit ledger that lives in a system does not have these failure modes. Every sale on credit is recorded against the customer's name and phone number. Every partial payment is logged. The balance is always current. When a customer comes in to pay, your staff can see exactly what she owes without calling you, without searching through a notebook, without uncertainty.
The trust that makes credit work in neighbourhood retail does not need to be replaced by a system. It needs to be supported by one. A customer who always pays eventually should have a clean record that reflects exactly that — and so should the one who needs a gentle reminder.