How to estimate queue length and waiting time at event entry

Your entrance can handle everyone coming tonight and still leave a queue outside when the first act starts. The missing detail is when people arrive. Capacity you didn't use half an hour ago can't admit the group getting off a coach now.

To estimate an event entry queue, divide the arrival period into short intervals. For each interval, add new arrivals to anyone already waiting, then subtract the people your entrance can serve. Carry the remaining queue into the next interval. Estimate a new arrival's wait from the people ahead and the capacity available while that person waits.

Those calculations answer different questions. How many people will be outside? How long will someone joining now wait? When will the entrance finally catch up? Confusing them can make a difficult opening look comfortable on paper. Here's how to build a forecast you can check against what actually happens at the door.

Decide where your queue starts and ends

Start by naming the section of the arrival experience you're measuring. For ordinary ticket checking, the queue might begin where attendees join the line and end when a check-in position starts serving them. Someone receiving help at that position is in service, rather than still waiting in the queue.

That distinction sounds fussy until you compare reports. A person can spend eight minutes waiting, another minute getting their ticket checked, then several more minutes waiting for a badge. Calling all of that an eight minute entrance experience hides the second queue and understates the time required to reach the event.

MIT's definitions of queue waiting and total time in a system make the distinction explicit. Waiting ends when service begins. Total time includes the service itself. Use the same definitions in your forecast, your observations and the information staff give attendees.

For this guide, queue length means people waiting to start the entrance task. Waiting time means the period before that task starts. If you want to predict when someone reaches their seat, add the time needed for the remaining tasks and movement through the venue. Don't quietly change the meaning halfway through your spreadsheet.

Keep physically separate queues separate. A busy west entrance and an empty east entrance don't produce a meaningful experience when you average them together. You can combine their totals for reporting, but the entrance lead needs to see the conditions at each location and whether attendees can actually use the other route.

Count arrivals before you count admissions

A ticket scan records a point in the admissions process. It doesn't tell you when the attendee joined the queue. If people arrive faster than the entrance can handle them, completed scans can stay almost level while the line outside gets longer. Reading that flat scan history as arrival demand would reproduce the same mistake next time.

Choose an arrival counting point that captures people before they wait. If the queue can extend beyond that point, move your observation upstream or use another counting arrangement. Otherwise the counter records people after they've already spent time waiting, and the apparently measured arrival pattern becomes another record of restricted flow.

The Federal Highway Administration's guidance on measuring queues from arrivals and departures describes this counting principle for transport queues. Its vehicle procedures aren't event operating standards. The useful principle here is conservation of people across a clearly defined boundary, with arrivals counted before the queue restricts them.

Count people rather than bookings if people are the unit your entrance serves. A group booking can contain several attendees. A family can arrive together while one person holds every ticket. Agree how your counters handle those cases so your forecast doesn't compare booking counts against a service rate measured in people.

You also need to record anyone who leaves the queue, changes entrance or moves to a different service. Add those movements to your reconciliation. If arrivals minus service starts predicts a longer line than observers can find, unexplained departures might be part of the difference. Don't automatically blame slow scanning or an inaccurate counter.

Check whether people return to the same queue after another task. Someone sent away to collect a document might rejoin later, while someone directed to a help position might finish there. Record the route your team intends to operate. Counting both as fresh ordinary arrivals would inflate demand, while ignoring both would miss work. For a return that takes priority, also include its effect on the people already waiting. The forecast needs to represent those movements consistently, even if you keep the numbers in a separate column. Ask the entrance lead to walk through one example before accepting the totals, especially when several suppliers share responsibility for the same arrival process.

Build an arrival forecast in short intervals

The first thing you want to do is map the reasons people will arrive together. Look at the opening programme, transport timetable, parking arrangements and the arrival instructions you plan to send. A headline performance at a fixed time gives attendees a shared deadline even if your doors opened much earlier.

Use records from a comparable edition where you have them. Preserve the differences that matter. A changed start time, different venue approach or new coach service can change the arrival pattern even when ticket sales are similar. Write those changes beside the forecast so someone reviewing it can challenge the actual assumptions.

For a first edition, make the uncertainty visible. You might estimate arrivals from ticket holders' travel arrangements, timed bookings and venue observations of comparable events. Those are inputs to a forecast, rather than measured behaviour from your event. Keep a note of which intervals depend most heavily on judgment.

Five minute intervals are a workable starting choice for the fictional example below. They aren't a universal recommendation. If a coach can release a large group within a minute, five minute totals can hide the queue forming inside that interval. Use a shorter interval around that arrival or model the group as a separate burst.

Keep the forecast totals connected to your attendance assumptions. If you've forecast more arrivals than the relevant ticket holders, check for double counting. If some guests are expected after the headline act starts, keep them in the later rows. Removing inconvenient late arrivals doesn't make the entrance recover any faster.

Where you control booking windows, our guide to choosing timed-entry intervals explains the scheduling decision. For this calculation, use the arrival pattern you have a credible reason to expect, including early and late arrivals, rather than assuming everyone follows their assigned time perfectly.

Use capacity the entrance can actually deliver

Your capacity input is the number of people who can start the defined service in each interval when a queue is waiting. Measure or rehearse the complete operating process. If the next attendee can't approach until the current attendee has received a wristband, that handover affects how often a new service can start.

A quiet period tells you how much work arrived, rather than the maximum work the entrance could handle. If only twelve people arrive each minute, twelve completed interactions don't establish that your entrance is limited to twelve. Look for a representative busy period or use a rehearsal that keeps the working positions supplied with realistic tasks.

Record changes over time. Positions might open after doors, close for a supply problem or temporarily handle another job. Enter that change in the affected rows. Applying one capacity number to the whole opening period would assume help was available before it arrived and equipment kept working while it was stopped.

Use the capacity of the task you're forecasting. A ticket check can operate smoothly while mandatory security screening creates the actual entrance delay. Either model the stages separately or use a carefully defined combined process. Adding the maximum output of different stages together would count capacity that cannot admit additional people through the whole route.

This guide assumes you already have a proposed entrance operation. The purpose is to test the queue it creates over time. Building the wider staffing roster is a separate decision, covered in our event staff planning guide. Keep the queue forecast specific enough to show which part of that plan needs attention.

Carry each interval's queue into the next

Create one row for each interval. The opening queue is the previous row's closing queue. Add the arrivals expected during the interval. Subtract the number the entrance can serve during that interval, then stop the result at zero. A negative queue would pretend that spare capacity can serve people who haven't arrived.

Written as a spreadsheet calculation, the closing queue is MAX of zero and opening queue plus arrivals minus capacity. In this simplified model, capacity means service opportunities across the interval. The model treats flow as continuous and smooth within each row. It doesn't reproduce the order or duration of every individual interaction.

That last assumption matters. If everyone arrives at the final second, the entrance can't use all the capacity available during the earlier minutes. Shorten the interval or record the burst explicitly when timing within the row would change your decision. A tidy formula doesn't repair an arrival assumption that puts people in the wrong minutes.

For practical spreadsheet columns, use interval start, interval end, arrivals, capacity, opening queue, people served and closing queue. People served equals the smaller of capacity and opening queue plus arrivals under the same smooth-flow assumption. The closing queue is then opening queue plus arrivals minus people served.

For example, if columns C and D hold arrivals and capacity, put zero in E2 for the initial queue. Enter =MIN(D2,E2+C2) in F2 for people served and =MAX(0,E2+C2-D2) in G2 for the closing queue. Set E3 equal to G2 and copy the calculations down into the next spreadsheet row.

Add a notes column beside these calculations. Record a transport arrival, reduced capacity or an assumption you changed. When the forecast produces a surprising queue, those notes let you trace it to a specific condition. Without them, someone reopening the spreadsheet later has to guess why the numbers move.

Follow a fictional opening from doors to recovery

Imagine a paid evening event opening at 18:00. For illustration, assume the entrance can start service for 24 people per minute throughout the next forty minutes. That gives 120 service opportunities per five minute interval. These are invented planning inputs, not measured performance, an industry benchmark or a promise about any system.

Assume arrivals flow evenly within each interval. The five minute arrival counts are 60, 180, 240, 180, 60, 60, 60 and 60. Nobody is waiting at 18:00. There are no queue departures, priority admissions or interruptions. The continuous-flow calculation ignores the small number currently receiving service.

During the first five minutes, the entrance has capacity for 120 people and only 60 arrive. Everyone can start service without a modelled queue. The unused 60 opportunities disappear. During the next interval, 180 arrive and 120 can start service, so 60 people are waiting at 18:10.

Another 240 arrive between 18:10 and 18:15. Add them to the 60 already waiting, subtract 120 served and the queue reaches 180. The following interval adds 180 arrivals and serves 120, producing the peak of 240 waiting at 18:20. That is the number the entrance plan needs to address.

Arrivals then fall to 60 per interval while capacity stays at 120. The queue falls by 60 each time, reaching 180 at 18:25, 120 at 18:30 and 60 at 18:35. It reaches zero at 18:40. A queue created during fifteen busy minutes takes another twenty minutes to disappear.

The chart comes entirely from those fictional arrival and capacity assumptions. Its points show the queue at interval boundaries. The straight lines between points reflect the assumed smooth arrivals and constant service rate. They don't claim that a real queue grows in a perfectly straight line or peaks exactly on a five minute boundary.

Across the whole period, 900 people arrive and the entrance has 960 possible service opportunities. That overall comparison looks comfortable. Yet the peak queue reaches 240 because the spare capacity occurs before the surge. Moving the same total attendance around the clock changes the entrance experience without changing your ticket count.

Estimate the wait for someone joining now

For a single shared queue served in arrival order, a useful rough estimate is the number of people ahead divided by the combined rate at which service can start. At the fictional peak, 240 people ahead divided by 24 service starts per minute gives about ten minutes before a newly joining person starts service.

This is an approximation that smooths several working positions into one combined rate. The next available position might open sooner or later than that smooth rate suggests. People already receiving service also affect the first available opportunity. For exact individual predictions, you'd need to model those occupied positions and the sequence of interactions.

The calculation also assumes people who arrive later join behind the attendee you're following. New ordinary arrivals lengthen the queue behind them, rather than automatically increasing the work ahead. If priority guests or returning help cases enter ahead, their workload changes the estimate. Separate routes need separate assumptions.

What if capacity changes while the attendee waits? Follow the future rows instead of dividing by the current rate. Count how many people ahead can start service in each upcoming interval until you've accounted for the queue ahead. The last partial interval tells you approximately when that attendee reaches service.

Suppose, as a separate fictional sensitivity test, capacity falls to 12 people per minute immediately after the peak. Serving the 240 people ahead would then take about twenty minutes if that reduced rate persists. Keeping the old ten minute estimate would assume capacity the operation no longer has. Update the wait estimate when the working conditions change.

Staff communicating waits should use observations as well as the forecast. If the line has stopped, explain that the estimate is being checked rather than repeating an old number. Make clear whether the estimate covers ticket checking alone or the remaining route inside. Attendees need to know whether they can reach the activity they're paying to attend.

Calculate clearance separately from individual waiting

At 18:20 in the main example, an attendee joining the back waits about ten minutes to start service. The overall queue takes twenty minutes to clear. Both answers are correct because they describe different things. The person moves forward while more people join behind them.

To calculate clearance under constant conditions, divide the current backlog by service capacity minus the ongoing arrival rate. After the peak, the entrance serves 24 people per minute and twelve arrive. It removes twelve waiting people per minute. Dividing 240 by twelve gives twenty minutes to eliminate the backlog.

Dividing 240 by 24 would answer how long the current backlog alone needs, under the simplified assumptions. It ignores the people who keep arriving. That is why a supervisor can correctly estimate roughly ten minutes for someone joining now while also expecting to manage a queue for another twenty minutes.

If arrivals equal service capacity, the existing backlog doesn't shrink in this model. If arrivals exceed capacity, it grows. Clearance needs some future period with enough spare capacity to remove what has accumulated. When the rates change between intervals, continue the row calculation until the closing queue returns to zero.

Check recovery against the programme. A line that clears at 18:40 is a different operating result if the performance begins at 18:30. Some people may have entered, while others are still waiting or completing later entrance tasks. Decide how many attendees must finish the whole arrival route by the relevant programme deadline.

Use averages without hiding the worst period

An average wait can help compare versions of a plan, but it doesn't describe the experience of someone arriving at the peak. MIT's explanation of Little's relationship between average queue size and average waiting connects the quantities when the measurements and conditions match. It isn't a shortcut for predicting every attendee's wait from a single queue snapshot.

For the fictional example, the area under the queue line is 4,200 person minutes. You calculate that by taking the average of each interval's opening and closing queue, multiplying by five minutes and adding the results. This works here because the queue changes linearly within each interval and the full queue clears before measurement ends.

Divide those 4,200 waiting minutes by the 900 arrivals and the average modelled wait is about 4.7 minutes. That includes the people who encountered no queue. The estimate for someone arriving at the busiest point remains about ten minutes. Reporting only the average would hide the specific experience you most need to improve.

For observations, report the sample size and arrival period alongside any average or percentile. A percentile describes the waiting time below which a stated share of observed waits fall. It requires individual wait observations or a suitable detailed model. The interval totals alone don't establish a trustworthy percentile for your event.

Test the conditions that could change your decision

Start with one forecast you can explain, then alter the assumptions that might reasonably be wrong. Move a coach arrival into the busiest interval. Reduce capacity during an equipment interruption. Move some early arrivals closer to the opening act. Keep the total attendance unchanged when you're testing timing alone.

Change one assumption at a time first. You'll see whether the peak queue comes mainly from concentrated arrivals, slower service or a delayed opening. Then combine plausible difficulties where that combination matters. Document why you chose each case instead of presenting an arbitrary uplift as a measured confidence level.

For every case, compare peak people waiting, approximate wait for a peak arrival and the time the queue clears. Also check the point where the queue meets a physical or operational limit agreed with the venue. An earlier response may be necessary even if the eventual clearance time still looks acceptable.

A forecast in people doesn't determine the safe length of a queue in metres. The layout, available routes, access needs and crowd conditions matter. The HSE's guidance on crowd controls and queuing arrangements addresses protected routes and safe ways to leave queues. Have the responsible venue and crowd management people assess the arrangement.

Check what your five minute rows can hide

Return to the first interval of the fictional example. Sixty people arriving evenly over five minutes produce no queue because the entrance can serve twice that rate. Now suppose all sixty arrive together at 18:04. The interval total hasn't changed, but the queue forms immediately when the group reaches the waiting area.

Under the same smooth service assumption, that group needs two and a half minutes to start service. Only 24 can start before 18:05, leaving 36 still waiting at the boundary. The original row predicted zero. The difference comes from arrival timing inside the row, rather than a change in total demand or entrance performance.

You don't need one second rows across the whole evening to investigate this. Split the affected period into shorter intervals, or enter the group at its expected arrival time and calculate forwards. Keep the rest of the forecast as simple as the decision allows. Detail helps when it changes what you will do.

The same issue applies when capacity arrives late. A position opening at 18:04 doesn't supply five full minutes of work before 18:05. Enter the available portion of its capacity. If it can begin serving twelve people per minute, that final minute adds twelve opportunities rather than sixty. Check opening times against the actual setup schedule.

Also include anyone already waiting when your spreadsheet begins. If doors open at 18:00 but attendees queue outside from 17:40, the initial queue isn't zero. Either start counting when the first person joins, with no service before opening, or put the observed opening queue into the first row. Don't lose twenty minutes of demand by starting the clock late.

Compare changes using the same arrival assumptions

Once the base forecast is working, you can compare a proposed intervention without changing everything else. Suppose the fictional entrance increases capacity from 24 to 30 people per minute throughout the forty minute period. Keep the original eight arrival counts and smooth timing assumptions so the comparison isolates the effect of that capacity change.

The closing queues become zero, 30, 120, 150, 60, zero, zero and zero. The peak falls from 240 people to 150. A person joining at that peak has about five minutes of work ahead at the new rate. During recovery, twelve people arrive per minute and thirty start service, leaving eighteen opportunities per minute to reduce the queue.

Starting with 150 waiting at 18:20, clearance takes about eight minutes and twenty seconds. That places the empty queue around 18:28:20 under the assumptions. The improvement isn't simply the difference between the two hourly capacity totals. You can now describe the result in terms an entrance lead and budget owner can assess.

The next question is what it takes to achieve thirty service starts per minute in the real arrangement. Price the actual operational change and check it with the people responsible. If it requires equipment, trained staff or space you won't have, leave that case as a comparison rather than quietly adopting it as the approved forecast.

You can test arrival changes in the same way. Move a specified number of people into an earlier interval only when you have a plausible way to influence that timing. Earlier entertainment or a clearer transport instruction may help, but the forecast shouldn't assume an instruction guarantees attendance behaviour. Keep an unchanged-arrivals case until observations support the change.

Agree when the team needs to act

A queue forecast becomes useful during the event when it connects an observed condition to an available response. Work backwards from the situation you want to avoid. If the approved response needs time to prepare, the lead needs to hear about the developing queue before it reaches that limit.

For a separate fictional illustration, assume a venue team has identified a queue condition requiring action and your prepared response takes six minutes to become available. If the queue is growing by twelve people per minute, another 72 people could join the backlog during that preparation period. That is a reason to assess the trigger earlier with the responsible lead.

Those figures don't establish a safe queue size or prescribe your event's trigger. They show why response time belongs in the calculation. Use the venue's agreed conditions, your observed growth rate and the time the team actually needs. Rehearse the communication as well as the physical change so a radio discussion doesn't consume the available response time.

Check the forecast against real waits

During opening, record arrivals and service starts in the same intervals as the forecast. Also count the people waiting at agreed times. Reconcile the counts before deciding the model is wrong. A moved counting point, an unrecorded queue departure or inconsistent clocks can produce a discrepancy even when the underlying arithmetic is sound.

Observe individual waits across the arrival period. Use an agreed method that lets staff identify when a sampled person joins and starts service without collecting unnecessary personal information. Sample during the build, the peak and recovery. Watching only people who arrive after the rush will make almost any entrance look better.

Record interruptions with their start and end times. Note which positions remained usable and whether the queue moved to another stage. A scan report may explain activity at ticket checking while missing a stopped security line. Review the full route before attributing the difference to one device, instruction or team member.

The HSE's crowd monitoring guidance includes entrances, queues and potential pinch points. Your forecast should support that observation. Agree who receives the measurements and what decisions they can make when the actual queue develops differently from the plan. Numbers collected without a response won't change the experience outside.

Afterwards, revise the assumption that failed. If arrivals concentrated later, update the arrival profile. If a task occupied positions longer, revise that service assumption. Keep the original forecast so you can compare it with the observations. Replacing it with the actual numbers would erase what the planning process needs to learn.

For the next edition, check whether those observations still describe the event you're producing. Read your attendee reminder emails against the arrival plan. Then run the forecast with the programme and entrance process you will actually operate. You should be able to explain when the queue builds, how long people wait and what gets it moving before doors open.