Settlement Offers: When to Discount, How Much, and How to Document It

Settlement is the most misused tool in recovery, in both directions. Creditors settle accounts that would have paid in full, and refuse to settle accounts that will never pay anything — because both decisions get made against face value, which is the one number that is definitely not relevant. The balance on the account is what you're owed. The comparison that matters is what you would actually collect if you didn't settle, and against that benchmark a discount is frequently the higher-value outcome and occasionally a pure giveaway. This guide covers how to size an offer, why broad settlement campaigns can destroy more than they recover, and the documentation that prevents a settled account from becoming a dispute.

The comparison that matters

A settlement decision is an expected value calculation, and stating it explicitly resolves most of the confusion:

Settle: the offered amount, received with high probability, soon, at low cost.
Don't settle: your realistic recovery rate on this account × the balance, received later, minus the cost of pursuing it.

Everything follows from the recovery rate, which means a creditor without its own recovery-by-age data is guessing at every settlement decision. The measurement discipline in our metrics guide produces exactly this number, and it's the input that converts settlement from negotiation instinct into arithmetic.

Two consequences that surprise people:

Deep discounts are frequently correct. On an aged account where expected recovery is genuinely low, accepting a small fraction of face value beats the alternative — which is a small fraction of face value arriving later, less certainly, after more expense, or not at all.

Modest discounts are frequently wrong. Settling a recent account at a small discount, where the customer had capacity and would have completed a payment plan, is giving away margin to someone who would have paid. That's the more common error and the less visible one, because it looks like a successful recovery.

Why the plan comes first

The sequencing rule that protects the most value: offer a payment plan before offering a discount.

The reasoning is straightforward. A meaningful share of customers who can't pay today can pay in full over several months. Present those customers with a settlement offer and some will take it — which converts a full recovery into a partial one on an account that was never at risk. Present them with an affordable installment option and many pay everything, which is the outcome our payment plan guide is built around.

Settlement is the right tool where capacity genuinely won't reach the full balance in a reasonable period, or where the account is aged enough that the realistic alternative is little or no recovery. Those are identifiable conditions, not defaults.

The practical implementation: settlement authority should sit behind plan authority. A front-line agent should be able to offer installments freely and settlement only within defined parameters or with approval — which prevents the natural tendency to reach for the discount because it closes the conversation faster.

Sizing the offer

Uniform settlement percentages are the standard approach and they perform poorly, because they apply the same discount to accounts with completely different expected recovery.

The variables that should move the number:

Factor	Direction
Account age	Older accounts justify deeper discounts, because expected recovery has decayed
Reason for non-payment	A cash-constrained customer with intent may complete a plan; a disputed account needs resolution, not a discount
Contact and engagement	An engaged customer is more collectible than an unreachable one, which argues against discounting
Demonstrated capacity	Evidence of income and other payments argues for a plan rather than a settlement
Balance size	Larger balances justify more effort before discounting; small balances may be uneconomic to pursue at all
Prior arrangements	A customer who has broken plans repeatedly has a lower realistic recovery than the age alone suggests
Relationship value	A customer worth retaining may justify terms that pure recovery arithmetic wouldn't

Two design principles. Set authority tiers rather than a single number — defined discount bands available at different approval levels, keyed to account characteristics rather than to how persuasive the customer was. And don't negotiate against yourself. An opening offer followed by a better one when refused teaches the customer that refusing improves the terms, which is a lesson that spreads.

The campaign problem

Broad settlement campaigns — a discount offer sent across a portfolio — produce visible short-term recovery and two invisible costs that frequently exceed it.

Cannibalization. The offer reaches customers who were going to pay in full. Some take it. Every one of those is a pure loss, and none of them appear in the campaign's reported results — the campaign shows recovery, not the recovery it displaced.

Behavioral learning. Customers who observe that non-payment produces a discount adjust accordingly. This affects the accounts in the campaign and, more importantly, future obligations — a customer base that has learned waiting is rewarded pays more slowly on everything afterward. The effect is diffuse, delayed, and genuinely hard to measure, which is exactly why it gets ignored.

The version that works: targeted offers to accounts with genuinely low expected recovery, identified by age, prior contact history, and demonstrated inability rather than by portfolio-wide criteria. The targeting is what prevents both failure modes, and it requires the segmentation our collections framework describes rather than a blanket send.

If a campaign is run anyway — and there are legitimate reasons, including portfolio cleanup before a sale or year-end — measure the cannibalization by holding out a control group and comparing full-payment rates. That's the only way to know whether the campaign created value or moved it.

Structuring the settlement

The most common settlement failure is that the customer can't fund it. A lump sum they agreed to under pressure and cannot actually produce results in a broken settlement, an account back where it started, and a customer who has now learned that agreements don't need to be honored.

Structures that complete:

• Installment settlements. The settled amount paid over two to six payments completes at meaningfully higher rates than a lump sum the customer can't fund. This combines the capacity accommodation of a plan with the reduced total of a settlement, and it is under-used.
• First payment at agreement. The same principle as payment plans — collecting something immediately converts intent into action and is the strongest predictor of completion.
• Automated remaining payments with clear authorization and pre-charge notices.
• Defined consequences for default, stated in the agreement — typically that the settlement voids and the full balance less payments made is restored, which should be explicit rather than assumed.
• A short window. A settlement offer open indefinitely has no urgency; one with a defined expiry that you actually honor has some. Fake deadlines that get extended teach the opposite lesson.

One caution worth naming: never structure a settlement so the customer must borrow expensively to fund it. A customer who takes a high-cost advance to settle has resolved your account and worsened their position, which is bad practice and — where the pressure came from your messaging — creates exposure.

Documentation that holds

Settlements generate disputes months later with predictable regularity, and almost all of them are documentation failures.

Put in writing, before taking any money:

• The settlement amount and the payment schedule.
• What it resolves — explicitly, whether it settles the account in full or leaves a residual balance. Ambiguity here is the single largest source of post-settlement disputes.
• What happens to any remaining balance — forgiven, or still owed.
• How it will be reported to credit bureaus if you furnish. This term belongs in the agreement, because the customer's expectation and your obligation frequently differ.
• What happens on default of the settlement terms.
• Confirmation of completion, sent in writing when the final payment clears.

Practical requirements: send the agreement before the first payment, not after; keep it retrievable for the retention period; and never represent that a settlement will remove the account from a credit report unless that is accurate and you intend to do it — the accuracy standards in our compliance guide apply, and this is a common overstatement in the heat of a closing conversation.

One further item specific to older accounts: a payment on a time-barred debt can restart a limitations period in some jurisdictions. A creditor soliciting a payment on a debt beyond the statute needs to understand the state rules, and several jurisdictions require disclosure. This is a legal question worth resolving with counsel rather than by practice.

Credit reporting and tax

Credit reporting. If you furnish data, a settled account must be reported accurately — generally reflecting that it was settled for less than the full balance where that's what happened. Two common failures: continuing to report a balance after settlement, and misreporting a settlement as paid in full when it wasn't. Both generate disputes, and both are furnisher accuracy problems rather than customer service ones. The status you report should match what your settlement agreement said you would report.

Tax. Forgiven debt above a threshold may carry an information reporting obligation, and the forgiven amount may be treated as income to the customer — with exceptions including insolvency and certain other circumstances.

Two implications:

• Operationally, you need a process that identifies reportable forgiveness and files correctly and on time, with accurate taxpayer information — which means collecting it, since chasing it later is difficult.
• For the customer, this is an unwelcome surprise the following year. Disclosing it in the settlement communication is both fair and practical: a customer who receives an unexpected tax form generates a complaint, and one who was told expects it. Note that you can state the possibility without giving tax advice, and should suggest they consult a tax professional.

When settlement is the wrong tool

• A genuine dispute exists. Settling a disputed balance buys a resolution to a problem you should be fixing — and if the customer was right, you've collected on an error. Route disputes to resolution, per our deductions guide for the commercial version.
• The customer has capacity. Offer a plan.
• The relationship has ongoing value. A settlement generally ends the relationship; a plan can preserve it.
• The account is new. Early-stage accounts have high recovery rates and should be worked, not discounted — the decay curve argument in reverse.
• You can't fund the operational side. A settlement program without documentation, reporting, and tax handling creates problems that exceed the recovery.
• The customer is in a hardship situation where the appropriate response is forbearance or a hardship arrangement rather than pressure to produce a lump sum.

And the alternative worth considering before settling at a deep discount: placement or sale. For accounts where expected internal recovery is very low, contingency placement costs nothing unless it collects, and it may produce more than a deep settlement — the comparison in our agency guide. Selling the paper is the other route, at the pricing our debt buying analysis describes, with the caveat that a sold account leaves your control entirely and the buyer's conduct attaches to your former customer's experience of you.

What to measure

• Settlement acceptance rate by offer level and account segment.
• Settlement completion rate, split by structure — lump sum versus installment. This usually reveals that installments outperform substantially.
• Net recovery on settled accounts versus comparable unsettled accounts, which is the only honest test of whether settling created value.
• Cannibalization — full-payment rates in offered versus held-out populations.
• Average discount by age band, to see whether your discounts track expected recovery or just track negotiation.
• Post-settlement disputes, which measure your documentation quality directly.
• Re-default rate on broken settlements.

The comparison that matters most and is measured least: settled accounts against a matched control that wasn't offered a settlement. Without it, a settlement program reports recovery it may simply have redirected.

Frequently asked questions

---

This guide is educational and does not constitute legal, tax, or accounting advice. Statute of limitations effects, disclosure requirements, credit reporting obligations, and cancellation of debt reporting vary by jurisdiction and circumstance; consult qualified counsel and a tax professional about your program.