Recoverable vs. preventable leakage: what decides ROI

Executive Summary

A finding is not finished work until it is sorted into recoverable or preventable, because the two require different actions from different people on different timelines. Recoverable leakage needs a credit memo request, a rebate claim, or a formal dispute against a specific invoice, and it has a shelf life set by the vendor's own dispute window. Preventable leakage needs a rule change: a corrected rate card, a three-way match rule, an approval threshold, something that stops the same error from recurring on the next invoice.

The mechanism that causes underperforming audits is treating recovery as the whole engagement. A one-time credit clears the invoices already paid but leaves the contract's ambiguity, the missing rate table, or the unenforced cap exactly as it was. The next billing cycle reproduces the same drift, and the client pays for the same finding twice: once in the invoice, once in the audit fee to find it again.

What changes the return is scoping both halves from the start. A diagnostic that closes with a recovery total and a separate prevention roadmap gives the reader a number they can bank and a number that protects next year's spend. Only the first number is a check. The second is the one that compounds.

1. What separates recoverable leakage from preventable leakage?

Recoverable leakage is a dollar amount tied to specific invoices already paid: overbilled, duplicated, or under-credited. It is fixed by a claim against those invoices. Preventable leakage is a gap in a control, a missing rate table, an unchecked cap, an ambiguous clause, that will keep producing new overcharges on every invoice going forward until the rule itself is corrected.

One is a transaction problem. The other is a process problem.

The test is simple: can this dollar be claimed back from a vendor, or does claiming it back only fix the past while the same error keeps happening? A duplicate payment is recoverable: the vendor was paid twice for one shipment, and a claim against that specific transaction gets the money back. The absence of a check that would have caught the duplicate is preventable: nothing about issuing the credit stops the next duplicate from slipping through.

Most findings in a diagnostic carry both halves at once. A missed credit memo is recoverable on the invoices where the rebate was earned and not applied, and it is preventable on the process that failed to route the rebate clause to AP in the first place. Treating it as only the recoverable half means writing a good recovery memo and leaving the routing gap untouched.

This is why the diagnostic's roadmap output is split into two lists rather than one total, per the stated deliverable of a prioritized recovery and prevention roadmap in 2 to 4 weeks. Each finding lands on one list, the other, or both, and the split is what a client actually acts on.

2. Why does the split change what the audit is worth?

A recovery-only engagement returns a check and stops. The vendor relationship, the contract, and the AP process that let the drift happen are unchanged, so the same category of error accrues again on the next invoice cycle. Naming the preventable half turns a one-time number into an avoided cost every period after the engagement ends, which is the part a purely retrospective audit never touches.

A contingency-fee recovery firm is paid on the recovered dollar, so its incentive stops at the claim. Fixing the underlying control does not add to its fee, and in many engagements is outside its scope entirely. That is a legitimate business, and for a client who only wants cash back on old invoices, it is the right and cheaper answer.

The gap shows up in the following period. If the rate card was never corrected, the surcharge schedule never updated, or the approval threshold never tightened, the invoices that arrive after the audit closes repeat the pattern the audit just found. The client is now paying to have the same category re-discovered.

Naming both halves at the point of diagnosis is what closes that loop. The recovery total answers what comes back. The prevention roadmap answers what stops this from happening again, and it is scoped against the specific gap the recoverable finding exposed, not a generic control checklist.

3. How does a single finding get split into both categories?

Every finding is checked against two questions in sequence. First: is there a specific invoice, credit memo, or rebate clause a claim can be filed against right now. If yes, that dollar amount goes on the recovery list.

Second: what rule, if enforced going forward, would have stopped this from billing incorrectly in the first place. That answer, regardless of the first, goes on the prevention list.

A. Recoverable-side questions

Is the overcharge tied to an invoice still inside the vendor's dispute window. Is there a signed contract clause or rate schedule the vendor is answerable to. Has the credit or rebate already been earned under the contract's own terms, meaning the claim is a matter of enforcement rather than negotiation.

Each yes narrows the finding to a specific claim amount and a specific vendor contact to file it with.

B. Preventable-side questions

Would a rate card check at invoice entry have caught this. Would a not-to-exceed cap, correctly configured, have blocked the line. Was the contract term itself ambiguous enough that two reasonable readings produce two different bills. Each answer points to a specific control, a rate card, an approval rule, or a contract redline, rather than a generic recommendation to improve oversight.

4. Which categories tend to carry more of each type?

Categories differ in mechanism, not in which one leaks more, since that is not a comparison the diagnostic dataset supports. A duplicate payment is almost entirely recoverable because it names one transaction. A rate card gap is almost entirely preventable because there may be no single overcharged invoice to claim, just a table that was never corrected.

Most categories carry a mix and need both treatments.

Sorting a finding into either bucket depends on whether a specific invoice exists to dispute, not on the category's label alone. The categories below illustrate the pattern rather than rank it.

• Duplicate payment: Names a specific transaction paid twice, so the fix is a claim against that transaction.
• Missed credit memo: An earned credit sitting unclaimed is recoverable directly, but the routing failure that left it unclaimed is a separate, preventable gap.
• Rate card drift: Often has no single invoice to dispute; the fix is correcting and enforcing the rate card itself against every future invoice.
• Volume tier misapplication: Carries both: past invoices billed at the wrong tier are recoverable, and the trigger that should reset the tier automatically is preventable.
• Not-to-exceed overrun: The overrun itself may be disputable against the cap language, while the missing enforcement of the cap at invoice entry is the preventable half.

5. How do you size each bucket without a benchmark?

There is no industry table showing what share of a given company's leakage is recoverable versus preventable, so the sizing has to come from the invoice population itself. Sort every dollar found into one bucket, tally each bucket separately, and treat the recovery total as a bounded, near-term number and the prevention total as exposure avoided in future periods, not a cash figure to book today.

The arithmetic is straightforward once the findings are sorted. Sum the dollar value of every finding tied to a specific claimable invoice or credit; that total is what a recovery effort can realistically pursue, bounded by which invoices still fall inside a vendor's dispute window. Sum the annualized value of every finding whose invoice-level claim is small or absent but whose underlying rule is broken; that total describes what each future invoice cycle keeps losing if the rule is not fixed.

The two totals are not interchangeable and should never be added into one headline figure for a board update. A CFO told a single recovered figure when part of it is a projected, not-yet-realized prevention estimate has been given a number that will not appear in the bank account. Keep them on separate lines, with separate owners: AP or the vendor manager for the recovery list, and whoever owns the rate card, contract, or approval workflow for the prevention list.

6. Who should own each list once the diagnostic is delivered?

The recovery list belongs to whoever has the vendor relationship, usually AP or procurement, because filing a claim means engaging the vendor's account team directly. The prevention list belongs to whoever controls the system of record for the broken rule: the ERP admin for a rate card, the contract owner for an ambiguous clause, the controller for an approval threshold.

Splitting ownership this way avoids the common failure where one person is handed both lists and works through the easier one first, usually recovery, because it has a clear finish line: the credit posts, the case closes. Prevention work has no such finish line inside a single invoice cycle; it is a policy change that only proves itself the next time an invoice arrives and is billed correctly.

Assigning the prevention list to the system owner also forces a decision on each item: fix the rate table, tighten the approval rule, or renegotiate the ambiguous clause at the next contract cycle. Left unassigned, prevention items sit in a spreadsheet indefinitely, which is functionally the same as never having found them.

This is also where a managed services arrangement earns its keep for a client who does not want to build new internal ownership for every prevention item: it can hold the AP-side controls on an ongoing basis rather than handing a list back to a team already at capacity.

For the wider pattern this sits inside, start with the margin drift guide.

7. Frequently Asked Questions (People Also Ask)

Is recoverable leakage always worth more than preventable leakage?

Not necessarily, and the diagnostic does not assume it. A single rate card error can generate a small one-time credit but a much larger avoided cost across every invoice in the categories it touches going forward. The two totals answer different questions and should be compared on their own terms, not ranked against each other.

Can a finding be preventable but not recoverable?

Yes. If the vendor's dispute window has closed, or the contract language is genuinely ambiguous enough that no clean claim exists, the past invoices may not be recoverable. The rule that let the ambiguity through can still be fixed so the next invoice is billed correctly.

Does fixing the preventable half require new software?

Not always. Many preventable findings are closed with a corrected rate card, a tightened approval threshold, or a clarified contract clause at the next renewal, none of which require new tooling. Some do call for a system change, such as adding a check the ERP does not currently perform.

Who decides which vendors to file recovery claims against first?

Whoever owns the vendor relationship, typically AP or procurement, should prioritize by dispute window and claim size: invoices closest to losing their claimable window go first, regardless of dollar amount, because a missed window converts a recoverable dollar into an unrecoverable one.

Why does a contingency-fee firm not usually deliver a prevention list?

Its fee is earned on the recovered dollar, so its scope naturally stops at the claim. That is a legitimate, narrower service. A client who only wants cash back on past invoices may find that model sufficient; a client who also wants the drift to stop needs the prevention half scoped separately.

How is the prevention total kept from becoming a made-up number?

It is expressed as an annualized exposure tied to a specific, named control gap, not a lump projection. Each line states the mechanism it fixes, for example a rate card correction or a not-to-exceed enforcement rule, so it can be checked against next year's invoices rather than taken on faith.

Does every category split evenly between recoverable and preventable?

No, and there is no fixed ratio to apply. A duplicate payment is close to entirely recoverable because it names one transaction. A rate card gap is close to entirely preventable because there may be no single invoice to dispute. Most categories fall between those two and need both treatments.

What happens if the two totals get reported as one number?

A reader assumes the whole figure is cash, when part of it is an estimate of future exposure avoided rather than money in hand. Keeping the totals on separate lines, with separate owners, prevents that misreading and keeps each number checkable against what actually happens next.