A delayed checked bag can create several customer service cases at once. The passenger needs the bag located and delivered. Essential purchases may need review. A paid baggage fee may become refundable. Each path uses overlapping itinerary and baggage evidence, but it can have a different owner and completion point. This protocol evaluates whether those paths stay connected.

The protocol is an observational method for support records. It does not calculate compensation for an individual passenger, interpret an international treaty, or rate any airline. It tests whether a sampled case contains enough evidence to determine what was reported, how long the bag was delayed, which fee was paid, what actions were taken, and what the passenger was told.

Define one baggage journey as the unit

Use one checked bag on one linked itinerary as the primary unit. Store a study key for the passenger, bag tag, ticket, flight segments, and mishandled-baggage report. A party traveling together may have multiple bags with different delivery times, so avoid treating the reservation as one indivisible outcome.

The event begins when the passenger arrives at the ticketed destination without the bag or when the carrier records a mishandled bag. It ends when the passenger receives the bag at the agreed location, the bag is declared lost, the passenger withdraws the report, or the study cutoff is reached. Preserve missed connections, rebooking, and itinerary changes that affect which arrival starts the timing analysis.

Classify the outcome as delayed then delivered, declared lost, damaged, located but awaiting delivery, returned to origin, customer-collected, or unresolved. "Closed" in a tracing system is not an outcome unless the delivery or loss evidence supports it.

Build the timeline from operational events

Join scheduled and actual flight arrival, bag acceptance, screening and transfer events where available, tracing records, courier dispatch, attempted delivery, agreed delivery location, proof of receipt, and support conversations. Keep native timestamps and timezones. Record the exact source of the arrival and delivery times used in calculations.

DOT guidance states that, for the federal baggage-fee refund rule, significant delay is assessed using stated time thresholds and that the clock ends when the passenger picks up the bag or it reaches an agreed location (Lost, Delayed, or Damaged Baggage). The guidance distinguishes domestic and international timing. The codebook should preserve that classification rather than applying one threshold to every itinerary.

Do not let a courier's "out for delivery" event stop the clock. Likewise, a delivery photo without a matching agreed location should remain a review item. If evidence conflicts, retain both values and identify which owner must reconcile them.

Identify whether a baggage fee was paid, by whom, in what amount, for which bag, and through which sales channel. Separate base checked-bag fees from overweight, oversize, sporting-equipment, or optional-service charges. Record waivers, loyalty benefits, card benefits, and fees collected by another carrier.

The bag tag and payment record often live in different systems. The study should test the join rule and report unmatched records. A fee refund cannot be evaluated reliably if the analysis assumes every checked bag produced a fee or assigns one reservation-level payment to all bags.

14 CFR Part 260 contains DOT's current airline-refund framework, including provisions related to ancillary service fees (14 CFR Part 260). Use the current rule and applicable guidance to define the sampled eligibility logic. Legal or regulatory specialists should review ambiguous itinerary classes rather than expecting a frontline agent to improvise.

Keep the refund and bag-recovery workflows distinct

Map ownership for baggage tracing, delivery coordination, fee verification, automated refund issuance, manual refund review, interim-expense review, and passenger communication. A bag can be delivered while the fee refund remains pending. A fee can be refunded while the bag remains missing. Report both outcomes.

For every handoff, record the reason, destination, acceptance event, context sent, and next deadline. Look for orphaned actions, such as a tracing case that closes on delivery without sending the qualifying event to the payment system. Also examine duplicate actions created by automatic and manual refunds.

Measure time to mishandled-baggage report, first location update, delivery arrangement, actual receipt, fee eligibility event, refund authorization, refund settlement, and final notification. Do not compress these into one resolution-time metric.

Analyze interim expenses separately

Passengers may submit receipts for reasonable expenses incurred during a delay. Keep that review separate from the baggage-fee refund. Record the stated expense policy, documents requested, submission channel, decision authority, amount requested, amount approved, and reason for any difference.

Support copy should avoid implying that a baggage-fee refund resolves every expense or that delivery of the bag cancels an already valid expense review. The study should code whether agents described the separate paths accurately and whether repeated requests for the same receipt delayed completion.

Sensitive documents should be minimized. Receipts may expose location, payment details, or personal items. Restrict raw documents and use study keys in the analytical dataset.

Audit communication against current state

Extract the first acknowledgment, location updates, delivery estimate, refund statement, expense-review statement, and final notice. Compare each message with the system state at send time. "Your bag will arrive tonight" is a stronger claim than "the courier has accepted the bag for delivery," and the distinction should be coded.

Review whether the passenger received a report reference, a usable update channel, the agreed delivery location, and a clear explanation of separate refund and expense processes. If an automated refund was expected, determine whether the message explained timing and what to do if it did not arrive.

The DOT Airline Customer Service Dashboard provides public information about airline commitments (Airline Customer Service Dashboard). The DOT also provides a consumer complaint channel (File a Consumer Complaint). These sources can inform accurate escalation copy. They should not be used to threaten, discourage, or replace the carrier's own unresolved workflow.

Sample edge cases and test reviewer agreement

Stratify the sample by domestic and international itinerary, direct and connecting travel, single and multiple bags, paid and waived fees, delivered and lost outcomes, and manual versus automated refunds. Include cases with changed destinations, customer pickup, courier delivery, and incomplete data.

Two reviewers should independently code delay classification, start and end times, fee linkage, refund state, handoff completion, and message accuracy. Calculate agreement by field. Adjudicate disagreements with primary records and revise the codebook before publishing summary rates.

Run a sensitivity analysis for cases missing proof of receipt or payment linkage. Report unavailable evidence as a finding instead of treating it as no delay or no fee. Break results out by workflow and itinerary class.

This study can reveal where baggage events fail to trigger fee review or where passengers receive unsupported timelines. It cannot determine liability for a specific itinerary or prove that a missing operational event never happened. Travel organizations considering distributed support coverage can connect the evidence needs to travel and hospitality staffing, email and ticket support, and CS operations and reporting. The useful staffing question is whether an agent can see and join the records needed to move each workflow to its real completion point.

A repeat study should keep the same journey definitions while recording any changes to baggage systems, refund automation, carrier agreements, and customer notice templates. Without that change log, an apparent improvement could reflect a new data feed or a narrower cohort rather than better execution. Publish those limitations beside the results so leaders can distinguish observed service performance from changes in measurement.