An airport lounge can look comfortably sized across a whole day and still turn away travellers before the morning departure wave. Daily visits and simultaneous occupancy answer different questions.
Build the daily estimate first. Then ask when those visits happen and how long people stay. That second step is where the estimate becomes useful for an operator deciding whether to expand.
Start with the departures the lounge can serve
Suppose an airport terminal has 20,000 departing passengers per day. We are estimating visits to one lounge accessible to these passengers. Assume each lounge user makes one visit on that departure.
Use departures from the relevant terminal rather than the airport’s headline traffic figure. ACI’s airport rankings define total passenger traffic using arrivals and departures, with a specified treatment of transit passengers. A headline annual passenger number therefore needs interpretation before it becomes a lounge-demand input.
For this fictional exercise, the departure base is already supplied. There is no need to halve it again or add arriving passengers without an arrivals-lounge service in scope.
Separate customers without counting the same person twice
We will use two non-overlapping groups:
| Group | Share of departures | Probability of visiting this lounge |
|---|---|---|
| Premium-ticket passengers | 10% | 50% |
| Non-premium passengers with an eligible membership | 8% | 40% |
The second group explicitly excludes premium-ticket passengers. A business-class traveller with a qualifying card belongs in the first row only.
This matters because access rights overlap in real life. Adding all premium passengers to all cardholders without removing overlap would inflate the customer base. You can avoid the problem through mutually exclusive groups or subtract a known overlap, but you need to choose one approach.
Quick Guesstimate Drill
A rail operator is sizing a station lounge open to first-class ticket holders and to standard-class travellers with a current rail-club membership.
Place each traveller in exactly one group so nobody is counted twice.
Convert access into actual visits
Premium-ticket passengers generate:
20,000 × 10% × 50% = 1,000 visits per day.
Eligible non-premium members generate:
20,000 × 8% × 40% = 640 visits per day.
Total demand is 1,640 lounge visits per day.
Access does not imply use. A traveller might arrive too late, prefer another lounge or skip the visit. Here the usage probabilities already refer to this particular lounge, so do not apply another market-share adjustment unless the question gives a reason.
These are visits, not unique people over a year. A frequent traveller can create many visits. If you annualise the figure, label the result annual visits and account for any assumed seasonal variation.
Make a first capacity check
Suppose the operator supplies a planning assumption of three daily uses per seat and has 500 seats. That implies:
500 × 3 = 1,500 visits per day of planning capacity.
The daily estimate is 140 visits above capacity. At the same turnover assumption, accommodating 1,640 visits requires 1,640 ÷ 3 = 546.7, or at least 547 seats. That is 47 more seats than today.
Three uses per seat is a supplied operating assumption for this example. It should not be presented as a universal lounge benchmark. Opening hours, dwell time and demand concentration determine whether it is achievable.
I would treat the 47-seat result as an initial sizing answer and investigate the busiest period before committing to a floor plan.
Test the departure wave
Now suppose the busiest two-hour window attracts 600 visits. Visitors spend an average of 1.5 hours in the lounge during that window.
For a steady-flow approximation, the arrival rate is 600 ÷ 2 = 300 visitors per hour. Average occupied seats are then:
300 visitors per hour × 1.5 hours = 450 seats.
At a two-hour average stay instead, occupancy becomes 600 seats. The same number of daily visitors can therefore create a very different capacity requirement.
People in the lounge depend on the arrival rate and time spent there. For a departure wave, examine arrivals and departures over shorter intervals, including visitors already present. Use the average as a starting point for that schedule.
Quick Guesstimate Drill
A ski resort's mountain restaurant looks comfortably sized across the whole day, but skiers arrive in a rush at lunchtime.
- Lunch-wave diners
- 720 between 12:00 and 14:00
- Average stay
- 45 minutes per diner
- Current seats
- 240
Using a steady-flow estimate, how many seats short is the restaurant at lunch?
Check how much more demand the current seats can absorb
At 1.5 hours per visit, 500 seats support a steady arrival rate of approximately:
500 ÷ 1.5 = 333 visitors per hour.
That is about 667 visits over two hours under the same steady conditions. The current 600-visit window sits below that arithmetic limit, but there is little room for a sharp surge or a longer stay.
An operator might add seats, change access arrangements, manage reservations or open overflow space during the wave. The right choice depends on what causes crowding. Adding capacity used for only a short period has different economics from expanding a lounge that is busy all day.
Decide whether expansion earns its cost
Suppose an expansion costs £120,000 and can recover all 140 daily visits currently above the simplified capacity estimate. Each recovered visit contributes £8 after incremental service costs, and the lounge has 350 comparable operating days a year.
Annual added contribution is:
140 × £8 × 350 = £392,000.
If only 40 of those visits are genuinely recoverable, it falls to £112,000. The important research question is how much denied or diverted demand the expansion actually brings back, not merely how many eligible passengers pass through the terminal.
Present the two sizing answers clearly
“I estimate 1,640 daily visits. At the supplied three visits per seat, that suggests at least 547 seats. Before expanding, I would check the departure-wave occupancy and recoverable lost visits: dwell time and demand timing could change both the layout and the economics.”
Try Hospital-Campus Food Truck next. It gives you another setting where eligible customers, actual purchases and the busiest service period need separate treatment. For more on time in a system, see sizing a reusable container fleet.
Market sizing drill
Could you defend your estimate to an MBB interviewer?
Hospital-Campus Food Truck
Build an estimate aloud from your own assumptions. Get feedback on your breakdown, numbers and final sanity check.
Frequently asked questions
Building the daily visit estimate
Should I start from the airport's total passenger figure?
No. Start from departures through the terminal that the lounge serves. Headline airport traffic, such as the totals in ACI's rankings, counts arrivals and departures, with a specified treatment of transit passengers. Arriving passengers do not use a departure lounge, so the headline figure would overstate the base. Here the base is 20,000 departing passengers per day.
How do I avoid counting a business-class traveller with a lounge card twice?
Build groups that do not overlap. The premium-ticket group covers 10% of departures. The membership group covers 8% of departures and includes non-premium passengers only. A business-class traveller with a qualifying card belongs in the premium row. They are counted once, however many ways they could gain access.
How many visits a day should the lounge expect?
About 1,640. Premium passengers give 20,000 × 10% × 50%, or 1,000 visits. Eligible non-premium members give 20,000 × 8% × 40%, or 640 visits. The estimate assumes each lounge user makes one visit on their departure, so visits and visitors are the same here.
Does having lounge access mean a traveller will use it?
No. That is why each group carries a usage probability: 50% for premium passengers and 40% for eligible members. Eligible travellers may arrive too late, prefer another lounge or simply skip the visit. Treating every eligible passenger as a visitor would inflate demand and distort the capacity conclusions.
Should I also discount for travellers who choose a different lounge?
Not in this case, because the usage probabilities already refer to this particular lounge. A further discount for competing lounges would reduce the estimate twice for the same behaviour. If your probabilities described use of any lounge, you would need a separate step for the share choosing this one.
From daily visits to peak seating
Are 500 seats enough for 1,640 daily visits?
Not on the supplied planning assumption of three daily uses per seat. That assumption gives capacity for 1,500 visits, 140 short of demand. Meeting 1,640 visits at the same turnover needs 1,640 ÷ 3, or at least 547 seats. That is 47 more than today, and the answer depends heavily on the turnover assumption.
Why isn't a daily capacity check enough?
Daily visits and simultaneous occupancy answer different questions. A lounge can look adequately sized across the day yet still turn travellers away during the morning departure wave. Once you have the daily total, ask when the visits happen and how long people stay. Those two factors decide how many seats are occupied at once.
How do I estimate how many seats are occupied at the peak?
Multiply the arrival rate by the average stay. If the busiest two-hour window attracts 600 visits, the steady-flow arrival rate is 300 visitors per hour. At an average stay of 1.5 hours, about 450 seats are occupied on average. This is an approximation, so treat it as a planning guide rather than an exact peak.
What happens if travellers stay longer during the wave?
Occupancy rises in proportion to the stay. At a two-hour average stay instead of 1.5 hours, the same 600-visit window fills about 600 seats on average. That is more than the 500 available. Dwell time therefore matters as much as visit numbers and could change the layout decision as well as the economics.
How much extra peak demand can the current seats absorb?
Not much. At 1.5 hours per visit, 500 seats support roughly 333 arrivals an hour, or about 667 visits over two hours in steady conditions. The current 600-visit window fits below that limit. A sharp surge or longer stays could exceed it, so an operator might add seats or change access arrangements.
Expansion economics and presenting the answer
Is a £120,000 expansion worth it?
It depends on how many lost visits are genuinely recoverable. If all 140 daily visits above capacity are recovered, at £8 contribution each over 350 operating days, annual added contribution is £392,000. If only 40 are recoverable, it falls to £112,000, which would not cover the £120,000 cost within a year.
Why do the daily check and the peak check seem to disagree?
They rest on different assumptions. The daily check uses a supplied turnover of three uses per seat and suggests a shortfall. The wave test, at a 1.5-hour stay, puts average occupancy at 450 of 500 seats. Report both results, then check dwell time and demand timing, since either could change which constraint really binds.
How should I present the sizing answer to the operator?
Give the daily estimate, its capacity implication and what you would check next. For example: you estimate 1,640 daily visits, which at three visits per seat suggests at least 547 seats. Before expanding, you would test departure-wave occupancy and recoverable lost visits, because dwell time and demand timing affect both layout and economics.
What should I practise after this lounge case?
The article suggests the Hospital-Campus Food Truck guesstimate drill next. It applies the same discipline in a different setting. You separate eligible customers from actual purchases and then check the busiest service period. This case does the same by separating lounge access from visits and daily totals from peak occupancy.