Freeze the population before analyzing it
Define the entity, business unit, date range, carriers, modes, currencies, invoice population, and exclusions before measuring results. Silent population changes can create artificial savings, hide false positives, and make later settlement reconciliation impossible.
For a Second-Look review, the freeze should also capture incumbent-known findings, automatic credits, existing disputes, and already-open claims where those records are available.
Establish economic identity before comparing dollars
Match the invoice to the shipment or freight obligation using the strongest identifiers available: invoice number, PRO, tracking number, BOL, load ID, purchase/order reference, shipment date, carrier, lane, or other source identity. Corrected invoices, rebills, split bills, multi-leg moves, and duplicate-looking records need special care because several documents can represent one economic event.
Recalculate the expected charge independently
Use shipment facts and the controlling commercial source to reconstruct what should have been billed. Preserve the calculation inputs, units, effective dates, and rounding assumptions so another reviewer can reproduce the result.
Route ambiguity to review
Corrected invoices, credit-and-rebill sequences, multi-leg shipments, contradictory timestamps, unclear accessorial support, and uncertain identity should not be forced into pass/fail automation. Ambiguity stays visible until resolved.
Separate new findings from value that already existed
A freight audit should not take credit for an automatic carrier adjustment, a customer-known issue, an incumbent auditor's existing finding, an already-open claim, or the same economic error repeated across several records. Second-Look work is therefore measured as a delta: what supportable value remains after existing controls receive credit for their own results?
Separate a finding from actual recovered funds
A billing discrepancy is not automatically an approved claim, and an approved claim is not recovered cash. Credits, refunds, remittances, usable account adjustments, offsets, reversals, and counter-events are reconciled separately to the validated economic finding.
That state separation is why RETALLY distinguishes candidate difference, validated finding value, approved claim value, gross settlement, reversals, net actual recovered funds, and fee-eligible recovery.
Preserve reproducibility and reviewability
Another qualified reviewer should be able to understand why the finding exists without reverse-engineering a black-box score. Preserve the source identity, controlling authority, effective date, shipment facts, calculation inputs, reviewer disposition, authorization, claim state, settlement reference, and reversal treatment.
Automation can help assemble and prioritize evidence, but it does not replace the commercial source that governs the billed charge.
Why the method matters
Freight audit credibility depends less on how many exceptions a system can flag than on whether the evidence chain remains intact when finance, procurement, or the carrier challenges a finding. That is the standard the workflow is designed around.
The named standard
Read RETALLY Evidence Standard v1.0 for the ten buyer-visible rules, including Second-Look delta attribution and precise recovery language.
Apply the method to a bounded historical population
Start Free Recovery Audit715 Yorktowne Road, Pottsville, PA 17901.