Debt Payoff Order: The Math, and What Actually Finishes | HL Hunt
Small Balances: The Accounts Where the Cost Exceeds the Recovery
An agent call takes roughly the same time and costs roughly the same whether the balance is $40 or $4,000. Which means below a certain figure the treatment costs more than the account can ever return, and every additional contact makes the position worse rather than better. Most operations have never calculated where that line sits, work small balances the same way they work large ones, and report gross recovery that looks positive while the segment loses money on every account. This is the fixed-cost problem this desk keeps finding elsewhere, appearing inside the collections operation itself.
What you'll learn
The arithmetic
Work it with stylized figures to show the shape.
A fully loaded agent contact — wages, systems, telephony, management, and unsuccessful attempts — at $7.40 per attempted contact. A standard sequence of five attempts, three letters at $0.90 each, and some manual handling: call it $42 of treatment cost across the sequence.
| Balance | Recovery at 25% | Cost | Net |
|---|---|---|---|
| $40 | $10 | $42 | −$32 |
| $120 | $30 | $42 | −$12 |
| $260 | $65 | $42 | +$23 |
| $900 | $225 | $42 | +$183 |
| $4,000 | $1,000 | $42 | +$958 |
The top two rows lose money on every account, deterministically. Not on average, not sometimes — the expected value is negative before the first call is placed.
This is the mechanism from our selection analysis and verification analysis: a fixed cost applied to a small transaction consumes it. Those reports found it made small loans unviable to originate. Here it makes them unviable to collect — which means the same balance is uneconomic at both ends of its life.
Finding your break-even
The calculation, which most operations have never run:
Break-even balance = treatment cost ÷ probability of collecting
To use it you need two things, and the first is where operations go wrong.
Fully loaded cost per contact. Not the agent's hourly rate divided by calls. Include:
- Wages and benefits.
- Unsuccessful attempts — the ones that didn't reach anyone, which are most of them.
- Telephony, systems, and data costs.
- Supervision and quality review, per our monitoring analysis.
- Compliance and complaint handling attributable to the volume.
- Training and turnover.
Operations that compute this properly are routinely surprised, because the unsuccessful-attempt line dominates and nobody counts it.
Realistic collection probability, from your own data by balance band — and it should be net of what our cure analysis calls natural resolution. Accounts that would have paid anyway shouldn't be credited to the treatment, which means the honest probability is the incremental lift, not the gross rate.
That correction matters enormously here. A segment showing 25% gross recovery with 18% natural resolution has a 7% incremental rate — which raises the break-even balance by more than three times.
Why gross recovery misleads
The measurement failure specific to this segment, and it's the reason the problem persists.
Gross recovery on small balances is always positive, because some people pay. So the activity appears to work. An operation reporting "we recovered $340,000 from small balances" has said nothing about whether the segment contributed anything, because the cost isn't in the sentence.
What to report instead:
- Net recovery by balance band — gross recovery minus fully loaded treatment cost.
- Incremental recovery, from a holdout, rather than gross.
- Cost per dollar recovered, by band.
- Contacts per resolution, by band.
This is the monotonic-metric problem from our measurement analysis in one of its purest forms. Gross recovery rises with every additional contact, so it can never indicate that you've done too much — and an operation optimizing it will expand effort on accounts that lose money on each attempt.
The holdout is cheap here and unusually informative. Withhold treatment from a random sample of small balances and compare; if the difference is small, the treatment was buying almost nothing at a cost you can now quantify.
Changing the cost, not the effort
The productive response, because it moves the term that's actually binding.
The break-even balance is a function of treatment cost, so lowering the cost lowers the threshold. Channel costs differ by orders of magnitude:
| Channel | Relative cost | Suits |
|---|---|---|
| Agent call | Highest by far | Large balances, complex situations |
| Physical letter | Moderate, per item | Where required or where digital fails |
| Near zero | Small balances | |
| Text | Very low | Small balances |
| Self-service payment | Near zero per resolution | Anyone willing to pay |
Rework the earlier case with a digital sequence at $1.10 total instead of $42:
- $40 balance, 15% resolution: $6.00 recovered, $1.10 cost, net +$4.90
The same account moves from a $32 loss to a $4.90 gain — not by collecting more, but by spending less. A lower resolution rate at a fraction of the cost beats a higher one at full cost, which is the counterintuitive result that makes this segment work.
The design that follows: digital first for small balances, always, and escalation to agent contact only where the balance justifies it. Most operations do the reverse by default, because treatment paths were designed around large balances and small ones inherited them.
Removing every obstacle to paying
Where the strategy succeeds or fails, because a self-service path that doesn't work imports the cost you were avoiding.
What has to be true:
- A direct link to payment, not to a homepage.
- No account creation required — a login wall on a $40 balance loses most of the people who were going to pay.
- Minimal identification consistent with your obligations.
- The balance shown without the customer having to find it.
- Multiple payment methods, per our method analysis.
- Mobile-first design, since that's where the message was read.
- An instalment option available without a call, per our plan guide.
- Immediate confirmation.
Every step between the message and the payment loses a share of the people willing to pay — and on small balances there's no margin to absorb that loss. The economics only work if the path is genuinely frictionless.
The diagnostic worth running: attempt to pay one of your own small balances from a phone, and count the taps. Operations discover logins, forms, and dead ends they'd have said didn't exist.
When writing off is correct
The decision operations avoid because it feels like giving up, and which is frequently the right commercial call.
Below your break-even, after cheap channels have been tried, further treatment is a loss you are choosing to take. Writing off isn't failure — it's declining to spend money on something that cannot repay the spending.
What a threshold policy needs:
- A defined balance figure, derived from your break-even.
- A point in the sequence — after digital treatment, not before it.
- Consistent application, because an unwritten threshold becomes an inconsistent one and inconsistent treatment of similar accounts is a fairness problem.
- Correct reporting treatment, per our furnisher guide.
- Documentation of the rationale, since a write-off policy will be asked about.
- Periodic review, since the threshold moves as costs change.
And the comparison worth making explicit: selling small balances is usually not the alternative it appears to be. Per our debt sale analysis, buyers face the same fixed-cost arithmetic, so small-balance paper prices accordingly — and the documentation and representation obligations don't shrink with the balance. The economics that make the account uneconomic for you make it uneconomic for a buyer.
The risks in the other direction
Stated honestly, because the argument shouldn't be read as licence to ignore obligations.
- Consistency matters. Treating similar accounts differently without a defined rule creates a fairness problem, which is why the threshold has to be a policy rather than a practice.
- Obligations don't scale with balance. Dispute handling, per our dispute analysis, and accurate reporting apply identically to a $40 account and a $4,000 one. Cheap treatment cannot mean absent controls.
- Digital contact has its own rules on consent, frequency, and content — a low-cost channel isn't an unregulated one.
- Frequency still matters. Cheap contact makes over-contact affordable, which is a new failure mode. Per our complaint analysis, frequency complaints are usually configuration failures, and a cheap channel removes the cost constraint that was accidentally limiting them.
- Reputation doesn't scale down. A poor experience on a small balance is a poor experience.
The fourth is the one to watch as an operation shifts digital. The saving is meant to fund a lower threshold, not more messages — and an operation that spends the saving on volume has recreated the conduct problem at lower cost per unit.
Designing the segment
- Calculate fully loaded cost per contact, including unsuccessful attempts.
- Measure incremental resolution by balance band, using a holdout.
- Compute the break-even balance.
- Route everything below it to digital only.
- Rebuild the self-service path and count the taps.
- Set a write-off threshold and document it.
- Report net recovery by band, never gross.
- Cap contact frequency on cheap channels deliberately.
- Keep dispute and reporting controls identical across bands.
- Review the threshold as costs and channels change.
Steps one and seven are what make the rest possible. An operation that can't state its cost per contact and reports gross recovery cannot see this problem at all, which is why small-balance treatment persists in a form that loses money quietly and consistently.
Change the cost, not the effort
HL Hunt AI Debt Collection handles small balances through coordinated email and text with direct self-service payment and instalment setup requiring no call — so the break-even balance falls far enough that accounts which used to lose money contribute.
Frequently asked questions
Contact cost is largely fixed regardless of the amount, so below a certain balance the expected recovery is less than the cost of the attempts required.
Fully loaded treatment cost divided by realistic collection probability — using incremental rather than gross resolution, which usually raises the threshold substantially.
After cheap digital treatment, yes, below a defined threshold. Writing off after low-cost treatment recovers more than working them expensively.
No — gross is always positive because some people pay. Only net of fully loaded cost reveals whether the segment contributes anything.
Key takeaways
- Contact cost is fixed while balance varies, so below a threshold the expected value is negative before the first call.
- Compute break-even as treatment cost over incremental resolution — using gross resolution understates the threshold by several times.
- Gross recovery on small balances is always positive and therefore says nothing; report net by balance band.
- Digital channels move a $32 loss to a small gain by spending less, not collecting more.
- A self-service path with a login wall imports the cost you were avoiding — count the taps yourself.
- Cheap channels remove the cost constraint that was accidentally limiting contact frequency, which is a new conduct risk to manage deliberately.
Every balance treated at a cost it can support
Get started with HL Hunt AI Debt Collection for balance-band routing, frictionless self-service payment, contact frequency caps, and net-of-cost recovery reporting by segment.
This guide is educational and does not constitute legal, accounting, or compliance advice. Worked figures are stylized illustrations. Collection conduct requirements, consent and frequency rules for electronic contact, dispute handling obligations, credit reporting treatment, and the accounting and tax treatment of write-offs apply regardless of balance size and vary by jurisdiction. Consult qualified counsel and your accountant.