Singapore has a familiar little ritual.
You walk into a clinic, service centre or food stall.
You look around.
Then you ask the question that sounds almost too small to matter:
“Who is next?”
That question contains an entire civilisation problem.
There are more people who want something than can receive it at the same instant.
So somebody has to wait.
The interesting part is not the waiting.
The interesting part is how strangers agree whose turn comes when.
Quick Read
Singapore works partly because scarce capacity is often converted into visible allocation rules.
A queue number is not merely a ticket. It is a temporary claim on future service. An appointment is not merely a time printed on a screen. It is an attempt to reserve capacity before the person arrives. A triage rule is not a broken queue. It is a different allocation rule for situations where urgency matters more than arrival order.
Modern Singapore uses many versions of this idea. HealthHub lets patients book appointments, pre-register and obtain queue numbers at participating public healthcare institutions. SingHealth’s Q Track lets patients monitor how many patients are ahead, while explicitly noting that registering earlier does not make a scheduled patient see the doctor first: the scheduled appointment order still matters. ICA’s Appointment & Queue System lets customers book, receive notifications and check in digitally. IRAS offers appointment booking, check-in and queue-status functions. ServiceSG still operates physical service centres, including queue-ticket cut-off times, for people who need in-person assistance.
These are not identical systems.
That is the point.
The deeper capability is not “Singapore has queues.”
It is:
scarcity → allocation rule → visible position → expected turn → service.
And when the situation changes, the rule must change too.
Wait, What? A Queue Is a Technology?
A queue contains no engine.
No microchip is required.
Sometimes there is not even a ticket.
People simply stand one behind another.
Yet it is still a technology in the broader human sense: a repeatable method for solving a recurring problem.
The problem is simultaneous demand.
Five people want one cashier.
Twenty patients want three consultation rooms.
Hundreds of applicants want a service that can process only a certain number at once.
If everybody pushes forward independently, the scarce resource is allocated through force, luck, social dominance, confusion or whoever can attract attention.
A queue replaces that contest with a rule.
The simplest rule is:
first arrival → earlier turn.
That rule is not always fair.
But once it is visible and accepted, it can reduce conflict enormously.
The Queue Converts Conflict into Sequence
Imagine a hawker stall with ten hungry customers and no queue.
The cook can prepare only one plate at a time.
Scarcity already exists.
Removing the queue does not remove scarcity.
It removes the allocation mechanism.
Now every customer must negotiate individually.
“I was here first.”
“No, I ordered before you.”
“Mine is takeaway.”
“I only need one drink.”
The service provider now spends cognitive effort resolving social conflict instead of producing service.
A queue creates a shared sequence.
The demand remains.
The waiting remains.
But uncertainty about sequence falls.
That is the first hidden function of the queue:
it does not eliminate scarcity; it civilises the order in which scarcity is experienced.
A Queue Number Is a Portable Place in Line
A physical queue ties your body to your position.
If you are fifth, you stand fifth.
A queue number separates the position from the body.
Now you can sit.
You can look at your phone.
You can move within the waiting area.
Your claim on the sequence travels with the token rather than with your exact physical spot.
Digital queue systems push this further. HealthHub can display an assigned queue number and updated queue status. SingHealth’s Q Track can show how many patients are ahead. ICA’s Appointment & Queue System allows digital check-in before arrival in eligible workflows.
This creates a subtle improvement:
the user can wait without devoting full attention to waiting.
That matters because waiting has an opportunity cost. A person who must physically guard a place in line cannot use that time elsewhere.
A better queue does not only reduce minutes.
Sometimes it reduces the amount of life that must be organised around those minutes.
The Appointment Is a Queue in the Future
An appointment looks different from a queue because nobody is standing behind anyone else.
But structurally, an appointment often solves the same scarcity problem earlier in time.
Instead of arriving first and then claiming capacity, the user reserves a future slot.
This moves part of the queue from physical space into a scheduling system.
The system asks:
- How much service capacity exists?
- At what times?
- How long does a typical service take?
- How much variation should be expected?
- How many no-shows occur?
- How much urgent demand must be absorbed?
- How much buffer is needed when reality does not follow the timetable?
This is why scheduling is not merely a calendar problem.
It is capacity allocation under uncertainty.
Why “First Come, First Served” Is Not Always Fair
Queues are often associated with fairness because the rule seems neutral.
You arrived first.
You go first.
Simple.
But fairness depends on the objective.
Imagine an emergency department.
Patient A arrived twenty minutes earlier with a minor injury.
Patient B arrives now with a life-threatening condition.
Should A be treated first because A arrived first?
No.
Healthcare uses triage precisely because urgency can outrank arrival order.
Now consider a scheduled outpatient clinic. SingHealth explains that mobile-registering earlier does not make a patient see the doctor earlier; patients are seen according to the order of scheduled appointments.
The visible queue number is therefore not the full allocation rule.
This is a very useful lesson:
fairness is not always equal treatment; sometimes fairness requires relevant differences to change priority.
The Queue Has a Constitution
Every functioning queue has rules, even when nobody writes them down.
Who may enter?
What establishes position?
Can you leave and return?
What happens if your number is called and you are absent?
Who gets priority?
Can someone join on behalf of another person?
What happens near closing time?
ServiceSG’s physical centres, for example, publish opening hours and a final time for queue tickets. That rule prevents a service point from pretending it can accept unlimited new work right up to closing.
The queue therefore has something like a tiny constitution: entry rules, priority rules, exit rules and exceptional cases.
The clearer those rules are, the less social negotiation must happen at the counter.
Queueing Theory: Why Waiting Explodes Near Capacity
There is a mathematical reason queues can suddenly become awful.
Suppose a service point can handle roughly sixty customers an hour and fifty arrive.
There is spare capacity.
Now suppose fifty-nine arrive.
On paper, the system still has enough average capacity.
But real arrivals are not perfectly smooth and service times are not identical. Several customers can arrive together. One case can take unusually long. A staff member can be diverted. A machine can fail.
As utilisation approaches the system’s practical limit, small variations create disproportionately large waiting times.
This is one reason “maximum utilisation” can be a dangerous management goal.
A system with no spare capacity looks efficient until variability arrives.
Then the queue becomes the buffer.
The Human Becomes the Buffer
This is one of the most important moral insights hidden inside queue design.
When a system has insufficient capacity, somebody absorbs the mismatch.
Sometimes machines buffer the load.
Sometimes inventory does.
Sometimes schedules do.
And sometimes the human being waits.
The queue is therefore not just a measurement of operational performance.
It is where system insufficiency becomes lived time.
That matters because an hour does not cost everyone equally.
A salaried professional with flexible work may experience an hour differently from a shift worker who loses pay, a caregiver arranging childcare, an elderly person in pain or a parent managing three children.
So a queue can be orderly and still impose unequal burdens.
Good service design asks not only:
How long is the queue?
It also asks:
Who is carrying the cost of the queue?
Visibility Changes the Psychology of Waiting
Imagine two waits of exactly thirty minutes.
In the first, nobody tells you anything.
You do not know whether the system remembers you.
You do not know how many people are ahead.
You do not know whether the counter has stopped.
In the second, your position is visible and progressing.
The clock time is identical.
The experience is not.
Uncertain waiting consumes more attention because the brain keeps checking whether action is required.
Queue-status systems therefore create value even when they do not physically accelerate the server.
They reduce informational uncertainty.
This is an excellent example of how information can improve a system without changing its raw productive capacity.
A Digital Queue Can Move Waiting Out of the Building
HealthHub’s mobile registration illustrates a broader design shift.
At participating institutions, a patient can pre-register before the appointment, obtain a queue number and check queue status. ICA similarly describes its digital Appointment & Queue System as allowing customers with appointments to check in online, with the aim of reducing waiting time on-site.
Notice what this does.
The system does not necessarily make the underlying consultation, document check or service magically instantaneous.
It changes where the waiting happens and how much attention it requires.
That can release physical space.
It can reduce crowding.
It can let the user time arrival more intelligently.
It can make the queue more legible.
That is a good reminder that optimisation has many dimensions.
But Digital-First Cannot Mean Digital-Only
Now the counterargument.
A digital queue can reduce friction for one group while creating a new barrier for another.
What if someone does not own a suitable phone?
What if they are uncomfortable with apps?
What if they cannot read the interface easily?
What if authentication fails?
What if the person’s problem does not fit the digital workflow?
ICA has explicitly described its approach as “Digital-First, but not Digital Only,” and its newer service-centre design includes in-person assistance for customers who face difficulty with digital transactions. ServiceSG likewise operates physical centres to help citizens access services across agencies.
This is not an accidental side issue.
It is a general design principle:
an efficiency improvement is incomplete if it simply moves the difficulty onto the people least able to absorb it.
The Best Queue May Be the Queue That Never Forms
Suppose a government service can be completed entirely online.
The person no longer needs to travel, take a number, sit in a building and wait for a counter.
Has the queue disappeared?
Maybe.
Or perhaps it has been transformed into processing time inside a digital workflow.
This is why serious systems thinking avoids confusing visible queue with all delay.
A form submitted online can still wait in a back-office queue.
An application can still be awaiting review.
A computer job can wait for processing capacity.
A parcel can wait in a sorting hub.
The absence of people standing in a line does not prove the absence of queueing.
Sometimes the queue has simply moved somewhere the customer cannot see.
Hidden Queues Are Harder to Trust
Physical queues have one strange advantage.
You can see them.
If twenty people stand ahead of you, you understand roughly what is happening.
A digital or administrative queue can be invisible.
Your application says “processing.”
For how long?
How many cases are ahead?
Has anyone looked at it?
Did something fail?
This is why status information matters. A good digital service often needs to reconstruct some of the visibility that the physical queue provided naturally.
Progress states, estimated timelines, notifications and escalation routes can all make invisible work more legible.
The Queue Is a Trust Machine
Why do strangers tolerate waiting?
Partly because they expect the rule to continue applying.
If you believe people can simply cut ahead whenever they are louder, richer, better connected or more aggressive, waiting becomes psychologically harder.
Your position is no longer a credible claim.
The queue depends on procedural trust:
I accept waiting now because I believe the rule will eventually deliver my turn.
That is a powerful form of cooperation among strangers.
It appears trivial precisely because the social rule is so familiar.
Queue-Cutting Reveals What the Queue Was Doing
Sometimes we understand a system most clearly when someone violates it.
One person cuts the queue.
Why does everyone react?
Because the person has not merely moved forward physically.
They have broken the shared allocation rule.
Every person behind them now pays a small cost.
The violation is therefore collective.
This explains why queue norms can be so emotionally charged. A functioning queue converts private patience into a shared expectation of procedural fairness.
Priority Lanes Are Not Automatically Queue-Cutting
But now we need a distinction.
A priority lane can look like queue-cutting from the outside.
It may not be.
If the rule is publicly defined and tied to a relevant need, it belongs to the allocation system itself.
Examples might include urgent medical triage, accessibility assistance or service categories with different processing routes.
The key difference is legitimacy.
Was the priority rule known?
Is there a defensible reason?
Is it applied consistently?
Can people understand why the sequence changed?
This is a more sophisticated idea of fairness than “everyone must always be treated identically.”
Queues Reveal Bottlenecks
A queue is not only a problem.
It is also a sensor.
Repeated queues tell us where demand is accumulating faster than the next stage can absorb it.
If a restaurant kitchen is fast but payment is slow, the queue may form at the cashier.
If registration is fast but consultation rooms are constrained, the queue forms before consultation.
If immigration clearance is fast but baggage retrieval is slow, passengers may simply reach the next bottleneck sooner.
This leads to an important operational lesson:
speeding up one stage can move the queue without improving the whole journey.
Good systems therefore need end-to-end observation, not isolated performance trophies.
The Singapore Habit of Booking Is Really Demand Shaping
Appointments do more than organise individual convenience.
They shape demand across time.
Without appointments, many people may arrive at the same popular hour.
The service point experiences a spike.
Staff become overloaded.
Users wait.
Later, the same service point may sit partly idle.
A scheduling system tries to redistribute arrivals across available capacity.
This is demand smoothing.
It appears everywhere in modern life: restaurant reservations, medical appointments, airport slots, school timetables, maintenance windows and delivery time bands.
The same civilisation principle keeps returning:
when capacity is finite, timing becomes part of allocation.
Why Buffers Matter
A timetable that assumes every task will take exactly the average duration is brittle.
People are not average cases.
One patient needs five minutes.
Another needs twenty.
One application is straightforward.
Another has missing documents.
One customer knows exactly what to do.
Another needs explanation.
Variation is not an exception.
Variation is normal.
So a robust system needs enough slack, triage, cross-training, overflow paths or schedule flexibility to absorb variation without collapsing.
This is one reason a system that looks slightly under-utilised can outperform one operated permanently at the edge.
A Queue Can Fail in Several Different Ways
“The queue is long” is only one failure mode.
A queue can also be:
- unfair — priority is arbitrary or opaque;
- uncertain — nobody knows their position or expected wait;
- inaccessible — the process excludes people who cannot use the required channel;
- misrouted — users wait in the wrong line because the front-end classification failed;
- fragile — one staff absence or technical fault causes cascading delay;
- invisible — users cannot tell whether their request is progressing;
- misleading — the visible queue is short while a hidden back-office queue is enormous;
- unsafe — urgent cases are forced through ordinary order;
- wasteful — people are made to be physically present when remote processing would suffice.
This gives us a much richer way to evaluate service quality.
Do not ask only, “How many minutes?”
Ask, “What kind of waiting is this?”
No-Show Is a Systems Problem Too
Appointments create another difficulty.
People do not always appear.
If a system reserves scarce capacity for a person who does not attend, that slot may be wasted.
So operators send reminders, allow cancellation, permit rescheduling and sometimes design controlled overbooking based on expected behaviour.
But overbooking introduces its own risk. If everyone appears, the queue returns.
This is a beautiful example of a broader idea:
systems are often optimised against probabilities, not certainties.
That means management is partly the art of deciding which errors are tolerable.
What Happens When Demand Surges?
Ordinary queues are designed for ordinary demand.
Then something unusual happens.
A new policy launches.
A popular event opens booking.
A disease outbreak drives healthcare demand.
A transport disruption forces passengers into alternatives.
Suddenly arrivals exceed assumptions.
A resilient system needs surge logic.
Can extra counters open?
Can demand be redirected online?
Can lower-priority work be deferred?
Can staff be redeployed?
Can users be told when to return rather than being trapped in an indefinite wait?
The queue during a surge is a test of system adaptability.
A Queue Is Also an Information Problem
People often join the wrong queue because they do not know which service they need.
This is not really a capacity problem.
It is a classification problem.
If a citizen reaches the front after forty minutes and is told, “Wrong counter,” the queue has consumed time without moving the underlying task toward completion.
Good front doors therefore narrow the problem early.
What are you trying to do?
Which documents do you need?
Can the task be completed online?
Does it require an appointment?
Which specialist route applies?
This reduces misrouting before scarce service capacity is consumed.
The Student Version: “I Don’t Know What to Do First”
Students have queues too.
They just call them homework.
English composition.
Science revision.
Math correction.
CCA preparation.
Project work.
Everything wants attention.
The learner has finite cognitive capacity and finite time.
So a prioritisation rule is needed.
“Do whatever I feel like first” is one rule.
It may be a bad one.
A better rule considers deadline, importance, prerequisite structure, energy and task duration.
The queue therefore becomes a model of executive function:
many demands → limited capacity → priority rule → sequence → completion.
The Mathematics Version: Little’s Law
For older students, queueing opens a door into applied mathematics.
One famous relationship is Little’s Law. In a stable system, the average number of items in the system is related to the average arrival or throughput rate and the average time an item spends in the system.
In compact form:
L = λW.
You do not need advanced mathematics to understand the intuition.
If many jobs are present and throughput is fixed, time in the system tends to be larger.
If you reduce time without changing demand, either throughput or work-in-process must change.
This is why queue length, service speed and arrival rate cannot be discussed independently.
Mathematics gives us language for what everyone has felt while staring at a slow line.
The Economics Version: Time Has a Price
Queues often appear when prices are not being used to allocate scarce capacity.
That does not mean prices should always replace queues.
It means the scarcity cost has to appear somewhere.
If access is free or fixed-price while demand exceeds immediate supply, waiting time may become part of the effective cost.
This creates difficult questions.
Should a scarce service go to whoever pays most?
Whoever waits longest?
Whoever needs it most?
Whoever has a prior entitlement?
Different institutions answer differently because different goods carry different ethical meanings.
The allocation of concert tickets is not the same moral problem as the allocation of emergency care.
The Political Version: Procedure Makes Power More Predictable
At a larger scale, the queue points toward rule-based institutions.
People tolerate institutions more easily when they can understand how decisions are made and what happens next.
That does not require every decision to be identical.
It requires the relevant rule to be sufficiently knowable, defensible and consistently applied.
The queue is therefore a tiny model of procedural order.
It says:
your outcome should depend more on the declared allocation rule than on your ability to dominate the room.
That is not the whole of good government.
But it is a recognisable civic principle.
Why Small Singapore Can Still Have Big Queues
Singapore’s compact size does not abolish queueing.
In fact, density can concentrate demand.
A popular clinic, food stall, checkpoint, attraction or service centre can receive large numbers of people because many users are geographically close enough to access it.
Digital connectivity can concentrate demand even further. When everyone can click at the same second, a scarce appointment book can become a virtual queue instantly.
So “small country” does not mean “no scarcity.”
It changes the topology of scarcity.
The Queue Teaches the Difference Between Equality and Equity
Give every person exactly the same rule and you have formal equality.
But what if one person faces a relevant disadvantage?
What if standing for forty minutes is a mild inconvenience for one person and physically impossible for another?
What if one case is routine and another is clinically urgent?
Then identical treatment may produce substantively unequal outcomes.
This is where accessibility, priority seating, triage and assisted service routes enter.
The queue becomes a practical classroom for a difficult social idea:
fair procedure sometimes requires recognising relevant difference without turning every difference into privilege.
The Queue Teaches Patience — But That Is Not Its Main Job
Adults sometimes tell children that queues teach patience.
They can.
But that explanation is too moralistic if it stops there.
The queue exists because the system cannot serve everyone simultaneously.
Patience is the human adaptation to that constraint.
A badly designed institution should not excuse needless delay by praising citizens for patience.
Good service design respects people’s time.
That means reducing avoidable queues, making unavoidable queues more predictable and ensuring priority rules match legitimate needs.
A Thought Experiment: Singapore Without Queues
At midnight, every queueing rule disappears.
No numbers.
No appointments.
No priority rules.
No booking slots.
No digital check-in.
Every scarce service is allocated at the instant of demand through improvisation.
At first, people invent local solutions.
“You go first.”
“I was here before him.”
Someone writes names on paper.
Someone hands out handwritten numbers.
Someone starts taking bookings.
Within hours, society begins rebuilding queues.
Why?
Because the underlying scarcity never vanished.
That is the key insight:
when demand cannot be satisfied simultaneously, allocation rules reappear because civilisation needs a way to decide “next.”
Primary-School Lens: Why Can’t Everyone Go First?
For a younger child, the best question is wonderfully simple:
Why can’t everyone go first?
Because some things can only happen one or a few at a time.
One slide.
One cashier.
One teacher answering one complicated question.
One doctor examining one patient.
This introduces scarcity without needing an economics textbook.
Then ask a second question:
How should we decide who goes next?
Now the child has entered ethics, operations and social cooperation.
Secondary-School Lens: What Makes a Rule Fair?
For Secondary students, compare four allocation rules.
- First come, first served.
- Most urgent first.
- Random lottery.
- Highest price first.
Then give four situations:
- concert tickets;
- emergency medical treatment;
- a school competition with too many applicants;
- a limited number of subsidised support places.
Which rule fits which situation?
There is no single answer because fairness depends on what the institution is trying to achieve and which differences are morally relevant.
This turns a queue into a serious civics lesson.
JC Lens: Capacity, Incentives and Public Value
At JC level, queueing can connect economics, mathematics and public policy.
If a service is underpriced relative to demand, queues may ration access through time.
If prices rise to clear the queue, access may become faster but less equitable.
If capacity expands, financial and staffing costs rise.
If appointments become strict, utilisation may improve but flexibility may fall.
If digital channels dominate, transaction cost may fall for many while digital exclusion grows for some.
Every design moves costs and benefits across groups.
The sophisticated policy question becomes:
Which mix of capacity, pricing, scheduling, priority and access produces the best public value under real constraints?
What Singapore’s Queue Systems Reveal
Across healthcare, public services and administrative transactions, the interesting pattern is not uniformity.
It is layered allocation.
Some tasks move online and avoid physical queues.
Some tasks are scheduled.
Some allow mobile check-in.
Some retain in-person assistance.
Some use urgency-based priority.
Some make status visible.
In other words, the system keeps asking:
what kind of scarcity is this, and what allocation rule fits it?
That is much more powerful than a cultural stereotype about Singaporeans liking orderly lines.
Why Singapore Works Does Not Mean Nobody Waits
A country does not prove its competence by eliminating every queue.
That would be impossible and sometimes wasteful.
If a clinic hired enough doctors to guarantee zero waiting even during the most extreme imaginable surge, much of that capacity might sit unused most of the time.
The goal is not zero waiting at any cost.
The goal is proportionate waiting inside a system that is legible, safe, fair enough for its purpose and capable of adapting when conditions change.
That is a more serious definition of “works.”
A Five-Question Test for Any Queue
Next time you wait anywhere, ask five questions.
- Capacity: What scarce resource are we waiting for?
- Rule: How is priority actually decided?
- Visibility: Can users tell where they are in the process?
- Burden: Who pays most heavily for the waiting?
- Adaptation: What happens when demand surges or a person cannot use the normal route?
Those five questions can analyse a clinic, a school canteen, an airport, a website, a government service or your own study plan.
That is the power of a good node.
The queue is small enough to see and large enough to connect.
One Queue or Five Queues? Pooling Changes the Mathematics
Imagine a service hall with five counters.
There are two ways to organise the waiting.
In the first design, each counter has its own line. You choose one and hope.
In the second, everyone joins one common queue and the next available counter takes the next person.
The number of counters is identical.
The service speed of each counter can be identical.
Yet the experience can differ sharply because the second design pools variation.
In separate lines, one unlucky customer can be trapped behind a complicated case while another line moves quickly. In a pooled line, one slow case occupies one server but does not imprison everyone who happened to choose that line.
This is a beautiful operational principle:
shared capacity can absorb variability better than isolated capacity when tasks are interchangeable.
But the final phrase matters: when tasks are interchangeable.
If one counter handles passports, another handles complex appeals and a third requires specialised staff, a single undifferentiated queue may route people badly. Pooling works only where the servers can genuinely serve the same class of demand.
So the question is not “one queue is always better.”
It is:
which demands can share capacity without losing the expertise or priority distinctions the service needs?
Average Waiting Time Can Hide the People Who Wait Far Too Long
Suppose a service centre proudly reports an average wait of ten minutes.
Sounds good.
Now imagine nine customers waited one minute each and the tenth waited ninety-one minutes.
The average is still ten minutes.
But one person experienced a completely different system.
This is why high-quality operations should care about the distribution of waits, not only the mean.
How long do most people wait?
How long do the slowest ten per cent wait?
How often does someone cross an unacceptable threshold?
Are extreme waits concentrated in one kind of case or one kind of person?
This is the difference between average performance and tail performance.
Averages compress.
Compression is useful.
Compression can also hide receivers.
A civilisation that cares about service quality therefore has to ask who lives in the tail.
Balking, Reneging and Abandonment: The People Who Disappear from the Queue
Queueing has some wonderfully precise vocabulary.
Balking means seeing the queue and deciding not to join.
Reneging means joining and then leaving before service.
In everyday service work, we might simply call both forms of abandonment.
Why does this matter?
Because the people who disappear can vanish from the performance statistics too.
If a clinic measures only the waiting time of patients who stayed until consultation, a person who gave up after three hours may not appear in the average at all.
If a website is too difficult and users abandon an application halfway through, the completed applications can look efficient while the front door is failing.
If a citizen cannot obtain a booking slot and stops trying, the queue may appear short because unmet demand has gone invisible.
This produces an important measurement rule:
do not evaluate a queue only by the people who survived it.
Backlog Has an Age, Not Just a Size
Suppose two agencies each have one thousand cases waiting.
Are their backlogs equally serious?
Not necessarily.
In Agency A, most cases arrived yesterday.
In Agency B, hundreds have been waiting for months.
Same queue length.
Different queue health.
Backlog therefore has an age distribution.
Old cases are especially informative because they may indicate complexity, missing information, repeated handoffs, misclassification or cases that fell between owners.
A good system does not merely ask how many items are waiting.
It asks:
which items have been waiting abnormally long, and why have they not moved?
This turns the oldest part of the queue into a diagnostic sensor.
Priority Can Create Starvation
Earlier we saw why urgent cases sometimes need to jump ahead.
Now consider the opposite failure.
If urgent cases arrive continuously, a low-priority case can keep being postponed.
In computer science and operations, indefinite postponement is sometimes described as starvation.
The priority rule is individually defensible each time.
Yet the accumulated outcome can become unfair.
This is why priority systems sometimes need ageing rules, escalation thresholds or reserved capacity so that ordinary cases cannot be postponed forever.
The lesson is subtle:
a locally sensible priority rule can produce a globally bad queue if nobody watches what happens across time.
The Flash Queue: When Everyone Arrives at the Same Millisecond
Physical queues have friction.
You need to travel.
You need to reach the door.
Digital systems remove much of that friction.
That is usually good.
It creates a new problem when scarce capacity opens at a known time.
Thousands of people can press “Book” almost simultaneously.
The physical queue has become a flash queue.
Now tiny differences in internet latency, device speed, automation tools or refresh timing may determine who gets scarce slots.
First-click-first-served can look neutral while rewarding technical advantage rather than meaningful priority.
Alternative mechanisms include randomised lotteries, timed windows, eligibility tiers, rate limits or controlled virtual waiting rooms.
Each changes the fairness model.
The queue therefore follows civilisation into cyberspace, carrying the same old question in a new form:
when demand arrives faster than capacity, what should determine who gets through?
Goodhart’s Queue: When the Metric Improves but the Waiting Does Not
Suppose managers are told that “front-counter waiting time” must fall below ten minutes.
They respond by moving people from the front counter into a second internal holding stage.
The measured front-counter queue improves beautifully.
The customer still spends the same total time in the system.
This is a classic measurement failure. Once a metric becomes a target, people can optimise the visible number rather than the actual purpose.
Queue design is particularly vulnerable because waiting can be moved between stages, channels and definitions.
Check-in time can fall while back-office processing rises.
Appointment availability can look good if difficult users stop trying.
A helpline can answer quickly and then keep the caller on hold elsewhere.
The defence is to measure the receiver’s full journey.
Did the person’s real task reach completion, and how much total burden did the route impose?
One Patient Can Pass Through Many Queues
A hospital visit makes the point concrete.
A patient may wait for an appointment slot.
Then wait to register.
Then wait for consultation.
Then wait for a laboratory test.
Then wait for results.
Then perhaps return to another clinician.
Then wait for medication or follow-up scheduling.
Every individual queue can meet its local target while the total journey remains exhausting.
This is why end-to-end service design matters.
The user does not experience the organisation chart.
The user experiences the accumulated route.
Batching: Sometimes the Fastest Individual Service Is Not the Fastest System
Imagine a laboratory machine that can process samples more efficiently in groups of twenty.
Running every sample immediately may reduce one person’s wait but waste capacity overall.
Waiting briefly to form a batch may increase throughput enough to reduce total delay for everyone.
The same logic appears in transport, logistics, computing and administrative work.
But batching also creates a new cost: the first item waits while the batch fills.
So even “process faster” is not a simple instruction.
The system must choose between immediate responsiveness and efficient aggregation.
Good operations is full of these trade-offs because the goal is rarely to optimise one person, one counter or one minute in isolation.
A Field Exercise: Audit a Queue Without Complaining About It
The next time you encounter a queue, turn irritation into observation.
- What is the scarce server: person, machine, room, decision authority or information?
- Is there one queue or several?
- Does everybody require the same service time?
- Can demand be classified earlier?
- Who leaves before service and therefore disappears from the statistics?
- Is the visible queue the real bottleneck or only the most visible stage?
- Does priority create a group that waits indefinitely?
- Would pooling capacity help, or would it destroy useful specialisation?
- Which metric would a manager be tempted to improve even if the user’s total journey stayed bad?
- What would the receiver call “finished”?
This turns the queue into a live systems laboratory.
The Deeper Law: Waiting Is Stored Mismatch
At maximum resolution, a queue is stored mismatch.
Demand arrived.
Capacity could not absorb it yet.
So the mismatch was stored somewhere.
In bodies standing behind one another.
In numbers on a screen.
In appointments weeks from now.
In unresolved files inside a workflow.
In computer jobs waiting for a processor.
The sophistication of a civilisation is not measured by whether mismatches ever occur.
They always will.
It is measured partly by whether the mismatch is visible, whether the allocation rule is legitimate, whether the burden is proportionate, whether the oldest and most vulnerable cases can be found, and whether the system learns enough to prevent the queue from becoming permanent.
A queue is not merely people waiting. It is unfinished demand asking the system whether it still remembers whose turn must eventually arrive.
Frequently Asked Questions
Does getting a healthcare queue number earlier mean I will see the doctor earlier?
Not necessarily. HealthHub allows mobile registration and queue-number assignment at participating institutions, but SingHealth explicitly states that registering earlier through its Q Track service does not make a scheduled patient see the doctor first. Patients are seen according to the order of scheduled appointments, with clinical workflows and urgency also capable of affecting service sequence.
Why use appointments if appointments can still run late?
Appointments shape demand and reserve expected capacity across time. They cannot remove variation: some cases take longer, urgent work appears and operational disruptions occur. A good appointment system reduces uncontrolled arrival peaks without pretending reality will follow the schedule perfectly.
Are digital queues always better?
No. They can reduce physical waiting, crowding and uncertainty, but they can create barriers for people who lack devices, confidence, accessibility support or successful authentication. Singapore’s public-service design increasingly combines digital channels with assisted and physical routes rather than assuming one channel fits everyone.
Why do queues become much worse when a system is nearly full?
Because real arrivals and service times vary. When very little spare capacity exists, even small bursts of arrivals or longer-than-usual cases create backlogs that the system struggles to clear. High utilisation can therefore produce disproportionately long waits.
Is first come, first served the fairest rule?
It is fair for some purposes because it is simple and predictable. It is inappropriate where urgency, vulnerability, prior booking or another relevant criterion matters. Fairness depends on whether the allocation rule matches the legitimate purpose of the service.
What should students learn from queues?
Scarcity requires prioritisation. A good rule makes sequence visible, protects urgent cases, reduces conflict and adapts to constraints. The same reasoning applies to homework, revision, computing tasks, project management and public policy.
Sources and Further Reading
- HealthHub — Book and Manage Appointments Easily.
- HealthHub — Pre-register and obtain a queue number.
- SingHealth — Register and Track Queue FAQs.
- ICA — Appointment & Queue System and Digital-First, but not Digital Only.
- IRAS — Appointment System.
- Public Service Division — ServiceSG.
Final Thought: “Next” Is a Civilisation Word
Return to the hawker stall.
Ten people are hungry.
One cook is working.
The food still takes time.
The queue does not defeat physics.
It does something more human.
It tells a group of strangers that the scarce thing will be distributed according to a sequence they can understand.
Then civilisation keeps improving the sequence.
Numbers make places portable.
Appointments move claims into the future.
Status screens reduce uncertainty.
Triage changes priority when lives matter.
Digital systems move some waiting out of buildings.
Assisted routes keep the system open to people who cannot use the fastest channel.
None of this removes scarcity.
It makes scarcity more governable.
That is why Singapore works, in another small but consequential way:
when everyone cannot go first, the system keeps trying to make “who goes next?” a rule rather than a fight.