One Customer, Four Accounts: Collecting Without Colliding | HL Hunt
One Customer, Four Accounts: Collecting Without Colliding
A customer falls behind on a card, a loan, an overdraft, and a store account, all with the same institution. Four queues pick them up, four collectors call, and each one asks what they can afford. The customer, being honest, gives a similar answer to each, and agrees four arrangements that together exceed what they have. Within two months most of them break. The operation now has four broken plans where one sustainable plan was available, and a customer who has learned that engaging produced nothing except more calls.
What you'll learn
What account-level treatment produces
| Account-level | Customer-level |
|---|---|
| Four collectors, four calls | One conversation |
| Affordability asked four times | Assessed once |
| Arrangements that sum to more than income | One total the customer can keep |
| Hardship disclosed on one account | Recognized across all |
| Each team measures its own recovery | Recovery measured per customer |
| Customer disengages | Customer stays in contact |
Row three is the mechanical failure. Per our payment plan analysis, arrangements set above what a household can sustain break at predictable rates, and four arrangements each sized to the full affordable amount are guaranteed to do so.
Row six is the expensive one. Per our disclosure analysis, a customer who engaged in good faith and then found the pressure unchanged is much harder to reach the second time, and the operation has spent its most valuable asset, which is a customer willing to talk.
Linking accounts to a person
The prerequisite, and it is the same problem our exposure analysis describes on the lending side.
Matching fails in both directions and the failures are not symmetric:
- Failing to link leaves the collision in place, which is the problem this article is about.
- Over-linking attributes someone else's debt to a customer, which is among the most serious errors an operation can make, and one that is difficult for the person affected to correct.
What makes it hard in practice: name variations, address changes, accounts opened in different systems or acquired through purchase, joint accounts where one party is linked and the other is not, and business accounts held by the same individual.
Treat the matching rule as a governed input. Document it, measure both error rates rather than only the one that costs money, and review it, per our records analysis. Where a link is uncertain, the conservative action is to apply protective statuses across the possible match while not consolidating balances, which captures most of the conduct benefit without risking a wrong attribution.
Contact limits belong to the customer
The conduct point, and it is the one most likely to be missed by an otherwise careful operation.
A frequency limit applied per account multiplies by the number of accounts. An operation that allows a defined number of attempts per account per week, applied to a customer with four accounts, is contacting that household four times the intended rate while every individual limit is respected.
What to do:
- Apply caps at customer level, with account-level limits inside them.
- Count across channels, since calls, texts, emails, and letters are all experienced by the same person.
- Include agencies, per our agency analysis, because a placement elsewhere does not reduce what the customer receives.
- Suppress the rest when one account reaches the person, since the conversation should cover everything.
- Report contact volume per customer, which is the number that matches their experience.
Per our conduct analysis, requirements in this area differ by jurisdiction and by channel, and several regimes address contact frequency directly. Confirm with compliance how limits apply where one person holds several accounts, because the answer is not always what the account-level configuration assumes.
Assessing affordability once
Per our affordability analysis, capacity is a household property. It does not divide by the number of accounts.
A single assessment should establish:
- Income, at a reliable level, per our irregular income analysis where it varies.
- Essential outgoings, including obligations to other creditors.
- What is genuinely available for all debts with this institution.
- Whether the situation is temporary or lasting, which determines whether the answer is a short pause or a longer restructure.
Then say the number back to the customer as a single figure. One of the strongest reasons customers disengage is that they cannot tell what the institution is actually asking for in total, and a single figure with a clear split is the answer to that.
The same assessment should also surface whether the customer has obligations elsewhere that make even the agreed figure unrealistic. An arrangement that ignores the rest of their creditors is a plan built to break, and pointing them toward free debt advice is not a lost recovery, it is a more durable one.
Allocating one payment
Once there is a single affordable total, something has to divide it, and that something should be a rule rather than whoever called first.
| Approach | Effect |
|---|---|
| Pro rata by balance | Simple, defensible, easy to explain |
| Worst consequence first | Protects the customer from the most damaging outcome |
| Smallest balances first | Reduces the number of live accounts quickly |
| Highest cost first | Reduces what the customer pays overall |
Row two deserves priority where it applies. Per our triage logic, an obligation with severe consequences should be protected before one with mild consequences, and an institution that holds both is in a position to see that where the customer may not.
Whatever rule is chosen: write it down, apply it consistently, explain it to the customer, and confirm the split in writing. Per our promise-to-pay analysis, an arrangement the customer understands is one they can keep, and a split nobody explained is one they will query at the first payment.
Consider also whether consolidation is appropriate. Per our workout analysis, combining several delinquent balances into one arrangement can be simpler for everyone, though it has accounting, reporting, and legal implications that belong with your finance and compliance functions before it becomes an option a collector can offer.
Status that must propagate
The properties that belong to the person rather than the balance. Per our restrictions analysis, propagation is the recurring failure in collections systems, and multiple accounts make it worse.
- Hardship or vulnerability indicators. A disclosure made to one collector applies to every account.
- Contact restrictions and preferences. A request not to be called at work is about the person.
- Third-party authority, per our authority guide.
- Deceased notification, per our deceased accounts guide, which must suppress everything immediately.
- Notice of representation or insolvency, per our bankruptcy guide.
- Disputes, which are account-specific in substance but frequently signal something about the relationship.
The test is simple and worth running. Record a hardship disclosure on one account of a multi-account customer in a test environment, then check whether the dialer, letter production, digital channels, and any agency all reflect it the same day. Most operations discover at least one channel that does not.
The disclosure that reached one team
A customer explains a serious illness to a collector on one account and receives a demand letter about another the following week. From their side they told you. From the system's side the flag sat on one record, and the conduct failure that follows is indistinguishable from not having listened.
Why teams resist it
Worth naming, because the obstacle is usually structural rather than technical.
- Each product or portfolio has its own targets, so a customer-level arrangement that allocates less to one book looks like underperformance there.
- Collector incentives are account-based, per our incentive analysis, so time spent resolving a customer's whole position may not be credited.
- Portfolios sit in different systems, sometimes with different servicers.
- Placement with agencies fragments the picture further.
The fix has to reach the measurement, not only the process. If the operation measures recovery per portfolio and the arrangement is agreed per customer, the teams are being asked to cooperate against their own numbers. Credit the customer-level outcome to whoever resolves it, and measure total recovery per customer rather than per book.
Measuring at customer level
- Total recovery per customer, against total balances held.
- Arrangement durability at customer level, per our payment plan analysis, since that is the real test.
- Contacts per customer per period, across all channels and agencies.
- Share of multi-account customers with a single arrangement rather than several.
- Complaints per customer, which is where collision shows up first.
- Time from first contact to a whole-position resolution.
Item four is the simplest diagnostic and most operations have never produced it. Count the customers who hold more than one delinquent account and see how many are subject to more than one arrangement. That number is the size of the problem, and it can be calculated this week from data you already hold.
Per our segmentation analysis, the customer view also improves treatment selection, because a person behind on four accounts is a different situation from a person behind on one, and an account-level view cannot tell those apart.
Treat the customer, not the balance
HL Hunt AI Debt Collection links accounts to a single customer with governed matching, applies contact caps at customer level across every channel, holds one affordability assessment with rule-based allocation across accounts, and propagates hardship and restriction status to every linked record.
Frequently asked questions
The customer has one income and each account is treated as if it had its own, so the arrangements agreed cannot all be kept and most of them break.
By a rule set in advance: pro rata, worst consequence first, smallest balances first, or highest cost first. Consistency and explanation matter more than which rule.
Per customer. Account-level caps multiply by the number of accounts held, so every individual limit can be met while the household is contacted far too often.
Reliable matching of accounts to the same person, governed like any other decision input, since failing to link leaves the problem and over-linking attributes someone else's debt.
Key takeaways
- Four arrangements each sized to full affordability are guaranteed to break, and the customer is the only party who can see that.
- Frequency caps applied per account multiply by the number of accounts a person holds.
- Assess affordability once and give the customer a single total with a clear split.
- Allocate by a written rule, and protect the obligation with the worst consequence first where that applies.
- Hardship, restrictions, authority, and bereavement belong to the person and must reach every linked account the same day.
- Count multi-account customers with more than one arrangement; that number is the size of the problem.
One conversation instead of four
Get started with HL Hunt AI Debt Collection for customer-level treatment, cross-channel contact governance, and reporting on recovery and arrangement durability per customer rather than per book.
This guide is educational and does not constitute legal or compliance advice. Contact frequency rules, requirements for consumers in vulnerable circumstances, treatment of authority and bereavement notifications, and the accounting, reporting, and legal implications of consolidating balances vary by jurisdiction and product. Consult qualified counsel and your compliance function before changing treatment structures.