Why Do Customers Leave a Queue?

Across 1,673,247 emergency visits in New South Wales, how busy the department already was when someone arrived predicted whether they stayed. People arriving when the department was in its busiest fifth were “3.5 times more likely not to wait”. That is against those arriving at the quietest (Aust Health Rev, 2025).

That is a condition of the place, measured at the moment of arrival rather than after any waiting. Some of the reasons customers leave a queue get settled at the door, before anyone joins the line at all.

The rest get settled inside the queue, and they are not all about the wait. Three hundred of the people who walked out of one hospital were asked afterwards why they went. Their answers split between the queue in front of them and the day they still had to get through (Emerg Med Int, 2025).

So there is no single reason and no single moment. How many people leave, and how long they last first, is a different question: how long customers wait before leaving.

Key Takeaways

• Asked why they left, 240 of 300 people who walked out of a Riyadh emergency department said they were tired of waiting. 151 of the same 300 said the place looked very busy (Emerg Med Int, 2025). Each factor was asked separately, so those are two answers, not a ranking.
• How busy a place already is when someone arrives predicts whether they stay, before the wait itself has begun. Across 1,673,247 emergency presentations in New South Wales, people arriving in the busiest fifth were “3.5 times more likely not to wait” than those arriving at the quietest (Aust Health Rev, 2025).
• Leaving also depends on what it costs to leave. Where waiting was free to walk away from, delay pushed people out. Where they had already travelled, the same delay did not (M&SOM, 2025).

When customers leave a queue, what reason do they give?

636 people left one emergency department in Riyadh between January and May 2023 without seeing a doctor. Researchers reached 300 of them by phone within a week and asked why. The paper excluded the other 336 under criteria covering under-14s, pregnancies past 20 weeks and purely obstetric or gynaecological cases. Unanswered calls are a separate limitation the authors note. 240 said they were tired of waiting. 151 said the department appeared very busy (Emerg Med Int, 2025). So these are answers people recalled afterwards, not anything anyone watched.

Being tired of waiting is about the wait. A room that looks very busy is about what someone saw when they walked in.

Nobody in the Riyadh study was asked to pick a reason. Six statements were put to every patient, one at a time, with yes, no and don’t know recorded for each. So each row below adds up to 300, and the six rows add up to nothing.

Statement put to all 300 people who had left	Yes	No	Don’t know
I was tired of waiting	240 (80.0%)	52 (17.3%)	8 (2.6%)
I felt my problem could wait	73 (24.3%)	215 (71.7%)	12 (4.0%)
The ED appeared to be very busy	151 (50.3%)	134 (44.7%)	15 (5.0%)
I was unhappy with the hospital staff	79 (26.3%)	180 (60.0%)	41 (13.6%)
I was unhappy with the other patients	30 (10.0%)	227 (75.7%)	43 (14.3%)
I had to go back to home/work/school	71 (23.7%)	210 (70.0%)	19 (6.4%)

Statements worded as the study worded them. Source: Emerg Med Int, 2025.

215 of the 300 said their problem could not wait. 134 said the department did not look busy to them (Emerg Med Int, 2025). Asked directly, they ruled the reason out.

Two of the six statements are about other people. 79 said they were unhappy with the hospital staff, and 30 said they were unhappy with the other patients (Emerg Med Int, 2025). Both carry the most don’t knows in the table, 41 and 43, more than double any other row.

A second study puts unfairness on the list, in people’s own words. A 2024 mixed-methods study in Alberta, Canada, drew on interviews, a focus group and sharing circles. In it, 64 participants discussed why First Nations patients leave emergency care (CMAJ, 2024). Racism, stereotyping, communication problems, transport barriers, long waits, and “being made to wait longer than others” all came up.

Unfairness turns up as a stated reason for leaving, named by patients and by the staff who watch them go. Why it lands so hard is a separate question: being overtaken in a queue.

Abuguyan and colleagues set their own limits plainly: the data came from one hospital, and the factors behind leaving there need not hold anywhere else.

Does a crowded room make people leave before they join?

Two behaviours hide inside the word leaving. Balking is the decision not to join. It happens before a person has waited a minute, on the scene in front of them. Reneging is giving up after joining. It happens later, on how the queue behaves. A 2025 paper modelling electric vehicle charging queues uses both terms in that sense (Sci Rep, 2025).

A large recent count of what crowding does comes from New South Wales. Across 1,673,247 emergency visits, 155,425 people were recorded as not having waited, or 9.29% (Aust Health Rev, 2025). Those arriving when the department was in its busiest fifth were “3.5 times more likely not to wait” than those arriving at the quietest.

That figure is adjusted for age, triage category and time of arrival. It comes from linked hospital records at level 5 and level 6 hospitals, between April 2022 and March 2023. The count includes both those who turned round on arrival and those who gave up later.

What it fixes is how full the department was at the moment someone walked in, not the wait that person went on to have. Whether an arriving person could see any of that is a separate question, and this study did not ask it.

Fiorentino and colleagues describe a gradient rather than a threshold. They report a “direct dose–response relationship” between how busy the department was and how many people did not wait (Aust Health Rev, 2025).

A 2026 systematic review of crowding included 226 studies. It states that “Longer wait times and crowding have been associated with increased rates of LWBS and left without triage” (Clin Exp Emerg Med, 2026). Those are two of the categories a hospital uses: left without being seen, and left without triage. The same review rates 62% of its studies low quality and only 4% high. Both halves are the finding.

What a person meets on arrival belongs to the place and the day, not to the customer. Across eight UK airports in 2025, 41% of passengers at Birmingham reported no more than five minutes queuing at security. At Luton it was 80% (DfT, 2026).

According to the department, those are passengers’ perceptions of time queued, not measured waits, and people can overestimate. That gap is the subject of why waiting feels longer than it is.

Is it the wait, or what the queue looks like?

A queue shows a waiting person several things at once: how many people are ahead, how fast the front is moving, and where they stand in it. Which of those is a person actually responding to?

What is ahead, not what has passed

A 2026 study asked people what it would take to compensate them for seeing a queue through. It ran laboratory and online experiments with 1,163 unique subjects across 31 queue variations (M&SOM, 2026). Rational theory says that amount should follow the wait still to come. It did not.

The compensation people asked for was built from two visible things, added together. The number ahead of them, and the speed the queue was moving. The authors describe payment set on “what is ahead instead of what has already occurred” (M&SOM, 2026). That holds in the setting they built, where the queue was visible and the remaining wait was certain. Time already spent did little work. Neither did the remaining wait, taken as a single quantity.

A fast-long queue is a long one moving quickly. The paper finds that “people may react more negatively to a fast-long queue than predicted by rational theory” (M&SOM, 2026).

Amounts were elicited from participants rather than watched in a shop, and the country of the subject pool is not stated.

Being last in the line

In 2021 Ryan Buell reported a study of last-place aversion in queues, combining observation of grocery-store queues with four online studies (Manag Sci, 2021). Waiting in last place lowered how satisfied people were with the wait. It also raised the chance they switched queues or left.

Buell puts part of that down to a missing comparison. With nobody behind you, less reassures you that staying is worth it. The effect disappears when people cannot see that they are last.

Older still, and taken in a real shop. Lu and colleagues joined video of a retail store queue to till records. They reported in 2013 that customers seemed to weigh the queue’s length heavily, without adjusting enough for the speed it moved at (Manag Sci, 2013).

The 2013 field study and the 2026 experiment are compatible rather than identical. The older one says people underweight speed. The newer one says they do use speed, but add it to length instead of folding both into a wait. People do not work out the wait they have left. They respond to the parts of the queue they can see.

How much of a queue a person sees, and from where, follows from the layout of the counter. That choice is weighed up in single queue or multiple lines.

Does being noticed change anything?

Something other than the clock moves these numbers. At one large academic hospital in the Midwestern United States, staff called guest services ambassadors worked the waiting room. Ambassador presence was associated with a fall in the rate of leaving without being seen, from 3.4% to 2.0% (Acad Emerg Med, 2025). The adjusted odds ratio was 0.65, with a 95% confidence interval of 0.49 to 0.85.

Ambassadors were trained to “guide patients and visitors through the check-in and triage process” (Acad Emerg Med, 2025). They also attended to people in the waiting room, answered common questions and provided updates. The last item on the paper’s own list is to “manage expectations around wait times”.

The design limits how far that travels. It was an observational cohort, not a randomised trial. The comparison was before and after, at a single hospital of roughly 60,000 visits a year, from April to December 2022. Of 50,507 visits analysed, 9,798 fell in periods when ambassadors were on duty (Acad Emerg Med, 2025). Patients counted had arrived by private vehicle and spent at least 15 minutes between arriving and being given a room.

Set that next to a line in the Riyadh table. 75% of the people who later left had asked someone, while waiting, why they had not been seen (Emerg Med Int, 2025). They asked. They went anyway. The paper prints that figure without drawing the conclusion.

So a lack of contact turns up ahead of leaving, and contact on its own is not a cure. What is worth doing about a wait is a separate question, and it belongs in how to reduce perceived wait time.

Leaving costs more in some queues than others

Some queues are cheap to leave and some are expensive. A 2025 study looked at two large internal medicine outpatient clinics. One set of patients waited online. The other waited in the building. It measured what happens when a clinic runs late. Intraday delay “increases the telemedicine service incompletion rate by 7.40%, but it does not have a significant effect on the in-person service incompletion rate” (M&SOM, 2025).

Service incompletion is a composite, defined by the authors as either a patient no-show or leaving without being seen, so it is not a walkaway rate. The published abstract does not say whether 7.40% is a relative rise or a rise in percentage points, so the number stays inside its own sentence (M&SOM, 2025).

The authors’ own summary is that “telemedicine patients may leave without being seen, whereas in-person patients are not sensitive to intraday delay” (M&SOM, 2025). The same delay moved the queue people could leave for nothing. It did not move the queue they had already travelled to.

An illustrative pair. At a pharmacy counter inside a supermarket, the customer is already in the building, and stepping away costs a walk round the aisles. At a standalone pharmacy, the same customer made a separate trip and parked. The wait can be identical. What it costs to leave it is not.

There is a counting problem inside all this. Someone who joins and then goes leaves a gap in a queue you were already watching. Someone who takes one look and keeps walking leaves nothing behind at all. That is what makes measuring customer wait times harder than it sounds.

Two reasons to leave: the queue, and everything else

Two of the six statements the Riyadh leavers answered are about the counter: they were tired of waiting, and the place looked very busy. Two are about something else entirely. Their problem could wait, and they had to get back to home, work or school (Emerg Med Int, 2025).

Some of those people were going to leave whatever the queue did. That is not a service failure. It is the rest of a person’s day.

So a walkaway count is a mixed quantity. Neither the study nor a tally at your own door can split the two groups apart. An operator who reads the whole of it as their own failure will over-correct, chasing a wait that was never the reason those particular people left.

The count is still worth having. It just is not all one thing.

Frequently Asked Questions

What is the difference between balking and reneging?

Balking is refusing to join a queue. Reneging is joining and then giving up. A 2025 paper defines both terms in that sense (Sci Rep, 2025). The practical difference is where each one happens. Balking happens outside the queue, on whatever the place shows a new arrival. Reneging happens inside it, once the person has some experience of how fast the front moves.

Do people leave because the queue is long, or because they cannot tell how long it will be?

The evidence separates the wait from the room, not the wait from uncertainty. Of 300 people interviewed after leaving a Riyadh emergency department, 240 said they were tired of waiting and 151 said the place appeared very busy (Emerg Med Int, 2025). Nobody was asked whether not knowing the wait had pushed them out, so on that half of the question this study is silent.

Does arriving at a busy time make people more likely to leave?

Yes. A New South Wales study covered 1,673,247 emergency department arrivals, from April 2022 to March 2023. People arriving in the busiest fifth were “3.5 times more likely not to wait” than those who came when the department was quietest (Aust Health Rev, 2025). The authors report the risk rising steadily across the range, not switching on at one point. Volume was grouped into fifths, so no tipping point inside a band could be tested.

Why do people leave the self-checkout queue?

No published measurement was found for self-checkout queues specifically, so anything precise about them would be guesswork. What transfers is the finding that people price a queue from its visible parts rather than from the wait still to come. In a 2026 experiment with 1,163 subjects across 31 queue variations, the number of people ahead and the speed of service each counted separately (M&SOM, 2026).

Is it rude to leave a queue?

Not by any rule anyone enforces. A queue is a claim on a place in line, not a promise to stay in it. Leaving costs the people behind you nothing, because it moves them forward. The exception is a queue where going wastes something already done for you, such as an order made up or a table held open.

Do customers come back after leaving a queue?

In one study, most said they would. Asked whether they would use the same emergency department again, 213 of the 300 people who had walked out said yes, and 87 said no (Emerg Med Int, 2025). The scope is narrow. One hospital in Riyadh, a service people may have little choice about, and answers about intention rather than a record of what anyone did.

Sources

• Academic Emergency Medicine, Wing, Barclay-Buchanan, Arneson and colleagues, “Reduced left-without-being-seen rates and impact on disparities after guest services ambassadors implementation”, published online 5 February 2025, retrieved 2026-08-28, https://doi.org/10.1111/acem.15100
• Australian Health Review, Fiorentino, Dinh, Seimon and Bein, “Relation between emergency department patient volume at time of patient arrival and likelihood of patient to ‘wait’ for clinical care. A state-wide data linkage analysis from New South Wales, Australia”, volume 49 issue 5, published online 26 August 2025, retrieved 2026-08-28, https://doi.org/10.1071/AH24318
• Canadian Medical Association Journal, McLane, Bill, Healy and colleagues, “Leaving emergency departments without completing treatment among First Nations and non–First Nations patients in Alberta: a mixed-methods study”, published 21 April 2024, retrieved 2026-08-28, https://pmc.ncbi.nlm.nih.gov/articles/PMC11045230/
• Clinical and Experimental Emergency Medicine, Oskvarek, Leubitz, Rahman, Sure and Pines, “Emergency department crowding in the modern era: a systematic review (2018–2025)”, volume 13 issue 2, pages 119 to 139, published online 14 January 2026, retrieved 2026-08-28, https://doi.org/10.15441/ceem.25.172
• Department for Transport, “Air passenger experience of security screening: 2025”, published 26 August 2026, retrieved 2026-08-28, https://www.gov.uk/government/statistics/air-passenger-experience-of-security-screening-2025/air-passenger-experience-of-security-screening-2025
• Emergency Medicine International, Abuguyan, Alhusainy, Alsuliman, Alqahtani and Alrajhi, “Reasons and Features of Patients Who Leave the Emergency Department Without Being Seen”, published online 10 January 2025, retrieved 2026-08-28, https://pmc.ncbi.nlm.nih.gov/articles/PMC11742075/
• Management Science, Buell, “Last-Place Aversion in Queues”, volume 67 issue 3, pages 1430 to 1452, published March 2021, retrieved 2026-08-28, https://doi.org/10.1287/mnsc.2020.3619
• Management Science, Lu, Musalem, Olivares and Schilkrut, “Measuring the Effect of Queues on Customer Purchases”, volume 59 issue 8, pages 1743 to 1763, published August 2013, retrieved 2026-08-28, https://doi.org/10.1287/mnsc.1120.1686
• Manufacturing & Service Operations Management, Luo, Valdés and Linardi, “Experienced and Prospective Wait in Queues: A Behavioral Investigation”, volume 28 issue 2, pages 459 to 478, published March 2026, retrieved 2026-08-28, https://doi.org/10.1287/msom.2022.0352
• Manufacturing & Service Operations Management, Qin, Chan, Dong, Homma and Ye, “Waiting Online vs. In Person: An Empirical Study on Outpatient Clinic Visit Incompletion”, volume 27 issue 5, pages 1377 to 1395, published September 2025, retrieved 2026-08-28, https://doi.org/10.1287/msom.2023.0365
• Scientific Reports, Varshney, Panda, Shah and colleagues, “Novel control strategies for electric vehicle charging stations using stochastic modeling and queueing analysis”, published 1 July 2025, retrieved 2026-08-28, https://doi.org/10.1038/s41598-025-04725-7