Designing for Uncertainty:
Security Wait Times and
Upstream Travel Behavior
* Specific details have been omitted or modified for proprietary reasons.
Role:
Partners:
Scope:
Research lead
across all phases
Mobile,
Operations,
Airport Experience,
CX leadership
Multi-year, multi-phase research program spanning generative research, concept and content testing, and live pilot evaluation.
Presented to senior leadership across multiple functions accelerated as a top priority from the highest leadership level following heightened security disruption.
Informed development of United's Security Wait Times feature, now live in the app.
Summary
Few aspects of the day-of-travel experience are as consequential as airport security — and few sit so far outside the airline's direct control. Even though screening is run by the TSA, security is one of the strongest drivers of how customers judge their travel day overall, shaping when they leave for the airport and how they experience everything that follows. I led a multi-year, multi-phase research program to understand how uncertainty around security shapes traveler behavior and whether pre-arrival wait-time information in the app could change it. The research established a behavioral model: when travelers can't form credible expectations, they compensate by arriving early, which increases dwell time, sensitivity to delay, and frustration — a self-reinforcing cycle. In live pilot evaluation, surfacing wait-time information through the app shifted observed arrival times later. The work demonstrated that a targeted product intervention can reduce precaution-driven congestion without added capacity — and now underpins a feature that's live in United's app.
Problem Space
Travelers' experiences of airport security — and the terminal experience leading up to it — are poised to shift significantly over the next decade. Rising passenger volumes and structural constraints at major hubs are reshaping how security and pre-security spaces operate. As major carriers exert greater influence over airport expansion and how security operations integrate into the terminal, airlines increasingly shape how travelers prepare for screening, including before they ever arrive at the airport.
From the customer perspective, security poses a distinct challenge for airline experience design. Alongside delays, it's one of the strongest drivers of day-of-travel satisfaction. It shapes when travelers decide to leave for the airport and how they experience arrival. Yet it's the one stage of the journey airlines don't control directly. It sits between parts of the experience the airline does influence — check-in, bag drop, boarding — but travelers don't separate those boundaries when evaluating their experience. Stress at security produces a halo effect that the airline ultimately absorbs.
The research and design challenge was clear. If airlines are to increase customer confidence and reduce operational pressure without relying solely on infrastructure change, they need ways to influence customer behavior upstream, before airport arrival. That requires understanding how travelers plan around security, and whether providing credible wait-time information before arrival could change that behavior — encouraging later, more distributed arrivals while making the security experience feel more predictable.
Approach
The work unfolded as a multi-phase research program over two years, each phase addressing a distinct uncertainty about how security information could shape traveler behavior and experience.
Phase 1: Generative research. Foundational work examining how uncertainty around security shapes traveler emotions, planning behavior, and perceptions of fairness. The phase combined traveler interviews, in-airport intercept conversations, journey walkthroughs, an attitudinal survey, and academic and industry research on the psychology of waiting — perceived wait time, perceived fairness, and perceived justice. The objective was to understand how travelers decide when to leave for the airport, how they build buffers, and how security outcomes influence downstream satisfaction with the airline despite security not being airline-controlled.
Phase 2: Concept and content testing. Concept testing evaluated whether providing security wait-time information through the airline app could credibly reduce uncertainty and influence planning decisions. Content testing then refined the language, ranges, timestamps, and presentation needed for travelers to trust the information enough to act on it — rather than treating it as generic signage they'd seen before. The phase examined how phrasing influenced perceived accuracy, how variability could be communicated credibly, and how recency cues should appear.
Phase 3: Live pilot evaluation. Following design refinement, the feature was evaluated in live conditions on selected pilot flights to assess real-world impact on satisfaction, arrival behavior, and use of the app during the pre-arrival window. The phase combined post-security SMS surveys, intercept interviews, and geofencing-based arrival timing — paired together so reported behavior could be evaluated against observed behavior, not just self-report.
The design choice that mattered most was insisting that the work span all three phases under a single research program rather than being broken into separate engagements. Generative research alone would have surfaced the behavioral pattern but couldn't have validated the intervention. Concept testing alone would have produced a usable feature but couldn't have established the underlying behavioral logic — meaning the team wouldn't have known why the intervention worked and would have struggled to extend the principle to adjacent problems. Live pilot evaluation alone would have measured impact but couldn't have explained the mechanism behind it. Holding the three phases together produced something more durable: a behavioral framework, a product intervention, and observed evidence of behavior change, each one validating the others.
What We Found
Uncertainty, Not Wait Time, Was Driving Behavior
What emerged from generative research was a behavioral cycle driven by wait-time and procedural unpredictability. When travelers lacked credible expectations, they compensated by arriving early. Earlier arrival increased dwell time and sensitivity to delay. As waiting stretched without explanation, frustration escalated and was interpreted as unfairness or neglect. That experience reinforced precautionary behavior the next time, restarting the cycle.
The research also showed a clear pattern around confidence and planning. Travelers who didn't feel confident about what to expect at security arrived earlier and rated the security experience more negatively — even when their reported wait times were similar to travelers who felt knowledgeable and unconcerned. Secondary research on the psychology of waiting helped explain this: predictability and explanation often matter more for perceived control than absolute duration. When travelers can see how a situation is unfolding and understand why it looks the way it does, they can judge risk, time their actions, and decide what to do next. When they can't, even short waits feel threatening, because there's no stable basis for planning.
The practical insight was straightforward. Travelers don't need certainty. They need their expectations set and updated in ways that let them act with confidence.
Wait-Time Information Served Different Functions Across the Planning Window
With the behavioral foundation in place, concept testing surfaced an important nuance: wait-time information wasn't serving a single purpose. Its value to travelers depended on when in the planning window they encountered it.
Earlier in the planning window, many participants described using the feature for reassurance — checking wait times to reduce uncertainty without necessarily changing their intended arrival. Closer to departure, responses shifted toward efficiency: wait-time information now informed concrete timing decisions about when to leave. A third pattern — proactive planning — appeared more stable across intervals, with participants engaging consistently but with heightened sensitivity to framing and precision.
Content testing then focused on whether travelers could understand and act on the estimate without overreading it. Point estimates were often interpreted as promises. Ranges signaled variability more credibly but required careful formatting to remain actionable. Timestamps had to make the estimate feel current rather than static. The objective wasn't usability in the conventional sense. It was interpretability — making sure travelers could understand what the estimate meant, what it didn't mean, and how confidently to act on it.
This is where the work moved from "surface a number" to designing how a piece of information would be read. A wait-time feature that travelers don't trust doesn't reduce precautionary arrival; it just becomes another piece of noise they screen out.
Pilot Evaluation Showed Observable Behavior Change
Live pilot evaluation evaluated the feature on selected pilot flights, combining post-security SMS surveys and intercept interviews with geofencing-based arrival timing. Results indicated clear awareness and behavioral impact. The majority of survey respondents noticed the feature, and among those, most reported using it to help decide when to head to the airport or approach security.
Geofencing analysis provided early behavioral corroboration: arrival times for customers on pilot flights shifted later on average compared to travelers on comparable flights at the same terminal and departure windows where the feature wasn't available. The pilot duration and sample size limit definitive conclusions, but the direction of the shift aligned with both survey responses and the underlying behavioral model — an unusually clean alignment of self-report and observed behavior.
The pilot didn't just demonstrate that the feature worked. It demonstrated that the underlying behavioral model held up against observed behavior — which is the kind of evidence that supports continued investment in the principle as it scales.
What Changed
The Security Wait Times feature is now live in United's app, directly informed by the multi-phase research program described here. The feature represents one of the clearest examples in the day-of-travel experience of research that moved from behavioral foundation to product intervention to measurable behavior change — and is now in customers' hands.
The work was accelerated as a top priority from the highest leadership level following a period of heightened security disruption. The existing research foundation made it possible for the organization to respond quickly: rather than starting from zero in a moment of customer anxiety, the team had a behavioral model, a tested concept, and pilot evidence ready to inform development. That's a meaningful illustration of how foundational research creates strategic readiness — research that's already done when the organization needs to move fast.
Active follow-on work focuses on product refinement and data transparency: how the wait-time estimate is communicated as conditions change, how its accuracy is signaled to travelers, and how it should evolve as the feature scales. The behavioral framework has also become a reference point for related questions — how the airline supports travelers in security-adjacent stages of the journey, how location-based services can provide consent-based timing guidance, and how AI might help travelers synthesize fragmented signals during the pre-arrival window without becoming another layer to manage.
The work was presented to senior leadership across Digital, Product, Operations, Airport Experience, and Customer Experience throughout the program's two years. It now serves as a foundational reference for how the organization thinks about influencing customer behavior upstream of the airport — a question that becomes more strategically important as passenger volumes rise, terminal investments scale, and the relationship between airline-controlled and third-party-controlled stages of the journey becomes more consequential.
The deeper takeaway, both for security and for service experiences more generally, is that variability isn't always the right thing to eliminate. Often the more useful intervention is making variability intelligible — letting customers form and revise expectations with confidence, so they rely less on precautionary behavior and more on information.