Smart collectionsUpdated 6 min read
How can a credit union detect early delinquency sooner without losing closeness with its members?
How a credit union can spot early delinquency with data it already has and reach each member on time, with the right tone and without losing closeness.
Certia team

A credit union detects early delinquency sooner when it reads, beyond days past due, the signals that show up weeks before the due date, such as payments that arrive closer and closer to the deadline or deposits that become less frequent. With those signals organized, outreach goes out on time and in the tone that fits each member, the way a loan officer who has known them for years would do it.
What is early delinquency and why is it hard to see in time?
Early delinquency is the delay in the first days after a due date, when the installment can still be settled with a call or a simple arrangement. At that stage, the delay usually has a specific and temporary cause, such as an unexpected expense or a change in the date the member gets paid. If nobody talks with them in time, the debt ages and the relationship gets tense.
It is hard to see in time for a practical reason: the information lives in different places. The core banking system records the missed installment, the branch knows the member's history and the collections team works from a report that arrives the next day, or at the end of the week. By the time someone puts the pieces together, the member has already piled up several days of delay and the conversation starts at a disadvantage.
Which signals anticipate a delay before the due date?
The data your credit union already has says a lot. Before an installment comes due, it pays to watch for:
- Payments that arrive closer to the deadline every month, or that are split into partial payments.
- Changes in the savings account where the member receives income, such as deposits that arrive late or drop in amount.
- Members with several obligations at the credit union, when one of them is already behind.
- Channels that stopped responding: unread messages, unanswered calls, an address that is no longer valid.
- Local events that affect a whole group, such as the end of a harvest season or an employer that is late with payroll.
None of these signals is enough on its own. Together, and compared with the same member's behavior in previous months, they let you estimate a probability of payment and prioritize the portfolio before the delay shows up in the report.
Imagine a credit union with 40,000 members in the Sierra, Ecuador's highlands, with a large share of its portfolio in agricultural microloans. If the system crosses each area's harvest calendar with members' deposits, it can anticipate that a group of farmers will be short on cash for a few weeks. With that reading, the loan officer calls before the due date and offers the alternatives the credit policy allows, instead of finding out about the problem once the installment is already past due.
How do you organize the portfolio so each member gets the contact they need?
A useful segmentation answers two questions: how likely the member is to pay without help, and how delicate the relationship is. That produces groups with different treatments:
- Members who pay after a reminder. A friendly message a few days before the due date, through the channel they use most, is enough.
- Members with a one-time difficulty. They need a conversation and, if the policy allows it, a payment alternative.
- Members with signs of sustained deterioration. The officer who manages the relationship should hear about it before the collections team does.
- Cases that call for special care, such as a death in the family, an illness or possible identity impersonation, where an automated message would do harm.
Segmentation changes every week. A member can move from one group to another, and the system should reflect that without waiting for month-end close. That is why portfolio indicators are best read on a single dashboard, with the same source for management, risk, collections and the branches. This is what Data and decisions provides: a consolidated reading that arrives in time to act.
What tone and timing work with a credit union member?
Members expect a close relationship. They often know their loan officer by name and value that the credit union remembers their history. A cold, generic message can damage that trust, even when its content is correct. A few practices help protect it:
- Before the due date, inform. Amount, date and payment channels, in a short message and in the member's language; in some areas of the Sierra, that may include Kichwa.
- In the first days of delay, ask before you demand. A line like "can we help you find a date that works for you?" leaves the door open for the member to explain what happened.
- Respect the hours and limits set by applicable regulations and your own policy, and watch the frequency. Several messages on the same day feel like pressure, even if each one is correct on its own.
- Keep every payment commitment visible: what the member committed to pay and when, confirmed in writing and recorded for their officer.
- Escalate gradually, from the reminder to the call, from the call to a visit or a conversation at the branch, and only after that to formal collections.
Automation lets this happen with thousands of members at once. Sensitive decisions, such as offering a restructuring or sending a case to an external agency, stay with a person who knows the case.
What does your credit union need to get started?
A narrow scope and data that already exists are enough for the first step:
- Bring together the payment, savings and contact data that today is spread across the core system, the branches and the collections spreadsheets.
- Define with your risk team which signals you will watch and which action goes with each one.
- Test with one branch or one product for a full payment cycle and compare the results with your usual process.
- Record every contact and every commitment, so the officer and the collections team see the same thing and internal audit can review it later.
There are two precautions worth taking from the start. First, members' data is used with a defined legal basis and purpose, as Ecuador's Organic Law on Personal Data Protection (LOPDP) requires, and if a model takes part in decisions that affect a member, that member has the right to be informed and to ask for an explanation. Second, the strategy has to fit the regulations of the Superintendency of Popular and Solidarity Economy (SEPS) that your credit union already applies to portfolio management, so review them with your legal team before you change the process.
If you want to see how this approach would look with your own portfolio, our Smart collections service starts with a diagnostic on your data, as described in our operating model. Write to us with the size of your portfolio and the products that worry you most, and we will schedule a first review with your team.
Ideas for better decisions, in your inbox.
Up to two emails a month. No third-party advertising, and one-click unsubscribe.

