Designing Fairness Under Disruption: Compensation for Travel Delays

Summary

Flight delays are one of the most consequential moments in the travel experience, and one of the few where airlines can directly shape how disruption is judged. Yet compensation—one of the primary levers available—must balance cost, operational feasibility, and policy constraints while still improving customer perception. I led a large-scale research effort to understand how travelers evaluate delays and how compensation influences those judgments. Using a survey of recently delayed travelers combined with randomized delay scenarios varying timing, format, delivery, and value, the research showed that compensation expands the scope of evaluation: once offered, the airline’s response becomes the object of judgment—not just the delay itself. When compensation is late, effortful, or not usable during the disruption, it lowers perceived fairness, even relative to receiving nothing. Fairness instead peaks with early, immediately useful gestures and shows diminishing returns as delays lengthen. These findings informed time-banded compensation strategies that prioritize in-the-moment usefulness and low-friction delivery, aligning customer perception with operational and financial constraints.

Key words: IRROPs, day‑of‑travel disruption, perceived fairness, compensation design, timing, usefulness, accessibility, proportionality, diminishing returns, NPS, acknowledgment, delay windows, operational feasibility, cost at scale

* Note: specifics details of this project have been omitted for proprietary reasons.

Problem Space

Flight delays are among the most consequential moments in air travel. They test not only operational resilience, but something more fragile: a traveler’s confidence that the system is still paying attention to them.

Across day-of-travel research, delays consistently emerged alongside security as one of the strongest drivers of dissatisfaction. Industry data and internal benchmarks show that delays reliably depress Net Promoter Scores, with declines steepening as disruptions lengthen. Unlike security, which concentrates uncertainty into a defined window between leaving for the airport and getting through security, delays can make uncertainty feel indefinite. As delays extend, even consistent updates can become harder for travelers to interpret and trust. Plans become increasingly difficult to maintain, and travelers begin to evaluate not just the disruption itself, but how the airline appears to be managing it.

Improving the experience of travel delays has become a strategic priority at United Airlines; yet it has also proven to be one of the most difficult to address. Doing so requires coordinated efforts across operations, communications, digital experience, and ultimately finance, working together to reduce uncertainty and support travelers during disruptions.

Compensation plays a particularly complex role within this effort. It is one of the primary tools an airline can use to influence how a delay experience is ultimately evaluated. But it is also one of the most difficult to deploy effectively. Compensation introduces a distinct set of constraints that the airline must balance during and after a flight disruption:

• Cost: it carries substantial financial impact when deployed at scale.

• Operational and policy feasibility: it must be executable within operational realities (systems, airport/vendor constraints, staffing) and consistent with policy and regulatory requirements.

• Context-fit: it must work across a wide range of disruption situations (short, mid, long, overnight) and match what travelers expect in each—particularly around timing, usability in the moment, and proportionality.

Shaping how delays are experienced—and ultimately judged—requires balancing cost, operational feasibility, and psychological impact. Recovery gestures must signal acknowledgment and intent without introducing new sources of dissatisfaction.

The task of research was to translate these constraints into actionable rules for compensation that are both effective for travelers and feasible for the business at scale.

Research Design

To translate these constraints into rules the business could actually use, the research needed to answer three questions:

1. Does compensation change how delays are evaluated?
   How do travelers assess current industry-wide compensation norms and outcomes?

2. What drives judgments of fairness in delay-compensation scenarios?
   How do timing, type, delivery, and value shape those judgment, and what tradeoffs do travelers make between them?

3. Where do diminishing returns begin?
   At what point does increasing compensation value stop improving fairness judgments and brand impression?

A survey was chosen as the primary method because it could capture how travelers evaluated a recent real delay experience while also eliciting their evaluation of different compensation responses under standardized delay scenarios. By varying compensation timing, format (e.g., snack, voucher, points, credit), delivery method (e.g., automatic vs. choice), and stepped value levels, the study could separate the bundled elements of real delays and identify which combinations best supported customer’s judgments regarding fairness, utility, acknowledgement as a customer, and their impression of the airline .

Key sample characteristics included:

• N = 1121 responses from travelers who had experienced a flight delay within the previous 14 days.

• A intendedly broad carrier mix across United, American, Delta, Southwest Airlines so that expectations reflected common industry experience rather than a single airline’s policy environment.

• Most delays fell in the 1–3 hour range, and most disruptions in that window occurred at the airport.

• 639 travelers (57%) reported receiving no compensation, while 482 (43%) reported receiving some form of compensation (most commonly credits/points or meal vouchers), enabling direct comparison of compensated vs. uncompensated experiences.

The survey was structured in two parts:

Part A: Context and Real-World Baseline:
Participant’s trip and delay context and their evaluation of the airline’s delay response, including whether compensation was received or not.

Part B: Randomized modules
Each participant completed one of five randomly assigned modules, each focused on a different compensation dimension: 1) policy expectations/tipping points, 2) meal vouchers, 3) credits/miles/gift cards, 4) choice and proactive bundles, and 5) personalization/recognition.

Outcomes were assessed using 7-point scales capturing fairness, usefulness, acknowledgment, and brand impression, alongside structured context and open-ended questions.

Research Insights

1. Compensation expands what travelers evaluate during delays.

An immediately notable result was that, on average, travelers who received compensation rated the fairness of the airline’s response significantly lower (M = 2.83) than travelers who received no compensation at all (M = 4.48). How could this be? One explanation is that compensation changed the scope of what people were evaluating.

Before compensation, travelers assessed the situation of the delay itself—how disruptive it was, whether it felt unavoidable, and whether the airline appeared to be managing the situation reasonably. After compensation, travelers additionally evaluated the airline’s response. In particular, they evaluated in terms of fairness (whether the compensation was proportionate to the disruption), timeliness (arrived at the right time), accessibility (was easy to access), utility (was actually useful during the delay), and overall acknowledgement (whether the airline acknowledged the discomfort the customer had to endure.

When these elements are misaligned with the moment of disruption, compensation exposes the airline’s judgment in a way that, on average, reduce average perceived fairness and related metrics.

Further data corroborated this thesis that, by expanding the scope of customers’ evaluation, compensation exposes the airline’s judgment to scrutiny, often to its detriment, when the airline’s compensation response misaligns with customer expectations around 1) timing, 2) proportionality relative to delay length, 3) ease of access, 4) immediate utility, and 5) acknowledgment at the moment of disruption.

Among travelers who received compensation:

• 57% received it after the disruption had already ended (lack of timeliness).

• 52% reported that it did not help during the delay itself (lack of utility).

• 61% said they had to ask for it or actively search for it (inaccessibility).

• 49% received digital-only or post-travel value, such as credits or points (lack of immediate utility).

When compensation a) arrived after the disruption had largely resolved, b) required active effort to obtain it, or c) could not be used during the delay, it failed to align with expectations around timing, accessibility, and usefulness. This did not simply limit the benefit of compensation; it made the airline’s response itself the main object of judgment.

Perceived unfairness developed not simply or even primarily from the amount offered, but from the response delivery as a whole. Compensation that arrived late, required effort, or could not be used during the delay, became abstract rather than tangible, misaligning with the traveler’s situation and contributing to lower fairness ratings—even relative to travelers who received no compensation at all.

2. Perceived fairness peaks with early, immediately useful compensation responses.

Using travelers’ evaluations of their most recent delays as a reference point, we then examined how they judged compensation in controlled delay scenarios. Plotting fairness, usefulness, and brand impression across these scenarios revealed a consistent pattern: ratings rose sharply with early, tangible gestures—such as snacks or drinks provided around the first hour of delay—then declined quickly and sharply, flattening as delay length increased.

Across the controlled scenarios, perceived fairness peaks early and begins to level off by the 3+ hour range. Beyond this point, increasing monetary value produces diminishing returns: more compensation does not meaningfully improve fairness or brand impression, particularly when delivered in abstract forms (miles or credits) and/or can’t be used during the delay itself.

For short delays, immediate acknowledgment mattered most. Small, tangible gestures (i.e., a snack and/or drink) delivered early had the strongest impact, outweighing monetary value.

For moderate delays, uncertainty became the dominant factor. Compensation was most effective when it was delivered automatically and required no effort to access. How the gesture was delivered mattered more than increasing its value.

For longer delays, travelers shifted from noticing individual gestures to judging whether the response as a whole matched the scale of the disruption.

For extended or overnight disruptions, compensation alone had limited influence. Fairness perceptions stabilized over value increases. In this window, perceived effort or logistical support drove participants’ assessments more than incrementally higher compensation amounts.

Strategic Impact

This work reframed how delay recovery was understood internally: compensation is not a simple make‑good, and escalating value by default is often the wrong lever. The research showed that fairness outcomes were governed by measurable variables—usefulness, timing, effort, and value—and that the same spend could produce very different results depending on how they factored in the experience. In the process, the work also informed a move toward time‑banded compensation logic: different delay windows require different recovery responses.

Just as importantly, the findings clarified what the next phase of research needed to address:

• Extended and overnight disruptions: the data suggests a ceiling where compensation stops being the main driver of fairness. The open question is what combination of logistical support (rebooking clarity, hotel/meal support where relevant, assistance access) and communication can move fairness in that window, and what “good” looks like when outcomes are constrained.

• Accessible delivery at scale: “accessibility” emerged as a major determinant of fairness, but the operational question remains: what delivery models reliably reduce effort (automatic issuance, app delivery, gate/agent workflows, vendor integration) without creating new failure points.

• Communication as a multiplier: the results consistently pointed to timing and acknowledgment, but more work is needed to specify which communications—cadence, specificity, tone, and personalization—most effectively supplement monetary value, especially in the mid and long delay windows.

In short, the research did not only identify which compensation gestures perform best; it clarified where compensation is sufficient, where it is not, and what additional system design questions must be solved to improve fairness under real operational constraints.