How Many Counters Do You Need for Your Queue?

“How many counters do we need?” sounds like a facilities question. It is really a capacity question: how many people arrive, how long each visit takes, and how much waiting you are willing to tolerate. Get the number wrong and you either burn payroll on empty desks or watch a calm room turn into a complaint queue.

This guide walks through the practical maths of staffing a walk-in line—arrival rate, service time, utilisation—and applies it to clinics, council service centres, and salons. You do not need a PhD in queueing theory. You need honest measurements and a clear target wait. When you want a quick check against your own numbers, use the wait-time estimator; when you are ready to translate desks into budget, see pricing.

What a “counter” actually is

In queue software, a Counter is a staffed point of work that calls the next Ticket—Desk 1, Window B, Chair 3, Consultation room. It is not the same as a login, a role, or a location. One person can cover one Counter at a time. Two people sharing one window are still one Counter if only one Ticket can be served there.

That distinction matters for planning. Adding staff without adding a place to serve does not shrink the line. Closing a Counter for lunch without adjusting expectations does. Capacity lives at the desk, not in the org chart.

If your vendor prices by named users instead of Counters, you can still use this guide—just translate “counters” to “simultaneous servers.” The physics of the line do not care what the invoice column is called.

The three numbers that decide staffing

Almost every useful staffing conversation collapses to three inputs. If you cannot estimate them, start measuring for one busy week before you buy software or hire.

1. Arrival rate (λ)

How many people join the line per hour during the period you care about. Use the busy hour, not the daily average. A clinic that sees 60 visitors between 9:00 and 12:00 is not the same as 20 per hour all day—the midday spike is what breaks the room.

Count joins, not completed visits. Someone who leaves after twenty minutes still consumed a Ticket and staff attention. If you only count completed consults, you will understate load.

• Pull from appointment overlays plus walk-ins if both feed the same desk
• Separate Services if they never share Counters (billing vs clinical)
• Note seasonal peaks—renewal week at a council, flu season at a clinic, Saturday mornings at a salon

2. Service time (1/μ)

How long a Counter is occupied per Ticket, from call to done. Include the messy parts: printing forms, answering “one more question,” walking a visitor to another window. The stopwatch on a clean consult understates real occupancy.

Service time is often more variable than arrivals. A salon colour can run 90 minutes; a trim might take 20. A council licensing desk might finish most visits in 8 minutes and occasionally need 40. Plan with an average, then stress-test with a “long visit” scenario so one outlier does not surprise you.

3. Target wait (and how angry waits feel)

Capacity planning without a wait target is just arithmetic. Decide what “good” means for your room: under 10 minutes for most Tickets? Under 20? A hard cap before you open another Counter or pause walk-ins?

Perceived wait matters as much as clock wait. A live Ticket on a phone and a honest display board make the same minutes feel shorter than a silent room where everyone stares at reception. Software will not invent Counters—but it will stop people asking “was I skipped?” every three minutes, which is itself a drain on service time.

Utilisation: the number managers skip

If arrivals and service times are matched perfectly, Counters are busy 100% of the time. That looks efficient on a spreadsheet and catastrophic in a waiting room. Any small surge has nowhere to go. Queues form, waits explode, and staff burn out trying to “catch up.”

A practical rule for walk-in service desks: design for roughly 70–85% utilisation in the busy hour. Below ~70% you may be overstaffed for that window (or you have intentional slack for quality). Above ~90% you are one delayed visit away from a meltdown.

Utilisation for c identical Counters is roughly:

ρ = λ / (c × μ)

where λ is arrivals per hour, μ is Tickets one Counter can finish per hour (60 ÷ average service minutes), and c is the number of Counters. Keep ρ under your comfort threshold, then check whether the implied wait matches your target. The wait-time estimator is built for that back-of-envelope loop.

You do not need M/M/c tables on day one. You need to notice when your plan assumes robots never take lunch.

A simple method you can run on a whiteboard

1. Pick the busy hour you must survive (not the quiet Tuesday afternoon).
2. Estimate arrivals per hour for that window.
3. Estimate average service minutes; convert to μ = 60 / minutes.
4. Choose a starting Counter count c. Compute ρ = λ / (c × μ). If ρ > 0.85, add a Counter and recalculate.
5. Sanity-check wait: if the line still looks ugly at 80% utilisation, your service times are too long or too variable for that arrival rate. Fix process, split Services, or add capacity—software alone will not invent minutes.
6. Overlay breaks, training, and “only two of three desks are open on Wednesdays.” Capacity that exists only on paper does not serve Tickets.

Then price the desks. Esperaly bills per Counter with unlimited staff users—so the staffing answer and the subscription answer stay aligned. Details are on pricing.

Worked example: small clinic

A three-service walk-in clinic measures a busy morning: about 18 arrivals per hour across new patients, follow-ups, and a short billing window. Clinical visits average 12 minutes of Counter time; billing averages 6 minutes. For a first pass, blend to ~10 minutes overall (μ ≈ 6 Tickets/hour per Counter).

• With 2 Counters: capacity ≈ 12/hour, ρ ≈ 18/12 = 1.5 — impossible. The line grows forever during that hour.
• With 3 Counters: capacity ≈ 18/hour, ρ ≈ 1.0 — fully loaded. Any phone call or no-show recovery breaks the room.
• With 4 Counters: capacity ≈ 24/hour, ρ ≈ 0.75 — a workable busy-hour design if breaks are staggered.

Reality check: maybe only two Counters handle clinical work and one handles billing. Then you must split the arrival streams. If 14 of the 18 are clinical, two clinical Counters at 12 minutes (μ = 5) give capacity 10/hour against 14 arrivals—still short. The clinic either opens a third clinical Counter in peak hours, shortens clinical handling, or routes some demand to appointments. The whiteboard forces that conversation before you buy a fourth tablet.

For clinic-specific product fit (Services, waiting-room displays, what not to buy), see queue software for small clinics.

Worked example: council service centre

A municipal service centre runs licensing, rates enquiries, and general advice. Peak day: 40 arrivals per hour between 10:00 and 12:00 after a mail-out. Average handling is 8 minutes (μ = 7.5/hour), but 15% of visits are complex (20+ minutes). Using 8 minutes for the base plan:

• 5 Counters: capacity ≈ 37.5/hour, ρ ≈ 1.07 — still underwater in the peak.
• 6 Counters: capacity ≈ 45/hour, ρ ≈ 0.89 — tight but viable if supervisors can float.
• 7 Counters: capacity ≈ 52.5/hour, ρ ≈ 0.76 — healthier for a public-facing fairness mandate.

Councils also care about visible fairness. Even with enough Counters, a single shared paper list invites accusations of queue jumping. Multi-service desks need clear Service selection at join time so licensing does not starve behind long advice visits—or the reverse, depending on policy. A QR-led Ticket with Service choice is often the operational fix; see government & council queues and the companion piece on QR code queue systems for councils.

Peak-day planning tip: staff the busy hour, not the daily average. A centre that is quiet after 2:00 p.m. should not use the all-day mean to decide morning Counter count. Flex staffing and clear “Counters open” status beat permanently overstaffing empty windows.

Worked example: salon or beauty studio

Salons break the “identical servers” assumption. Chair time varies wildly, and a colourist is not interchangeable with a junior stylist for every Service. Still, the same skeleton helps.

Suppose Saturday morning brings 10 arrivals per hour for mixed services averaging 45 minutes of chair time (μ ≈ 1.33/hour per Counter). Then:

• 6 chairs (Counters): capacity ≈ 8/hour, ρ ≈ 1.25 — backlog builds through the morning.
• 8 chairs: capacity ≈ 10.7/hour, ρ ≈ 0.94 — still aggressive unless no-shows pad the schedule.
• 9–10 chairs or shorter average services: utilisation drops into a survivable band.

Practical salon moves that beat blind hiring: split Services (express vs colour), limit walk-in categories during peak, and use a Ticket so clients waiting for a specific stylist do not block a free chair that could take a different Service. Capacity is not only headcount—it is matching the right Counter to the right Ticket.

When adding a Counter is the wrong answer

Sometimes the maths say “hire,” but the process says “fix the visit.” Watch for these patterns before you open another desk:

• Service mix is wrong — long and short visits share one line, so short Tickets starve. Split Services or dedicated express Counters.
• Join friction burns staff time — reception hand-writes every Ticket. Self-serve QR join returns those minutes to serving.
• Callbacks and transfers are chaotic — people rejoin incorrectly and inflate arrival counts. Transfer Tickets instead of creating ghosts.
• Displays are lying or blank — visitors cluster at the desk asking status, which lengthens every service time. A honest board is capacity.
• You are measuring the wrong hour — overstaffing the quiet afternoon does not fix the morning spike.

Software will not replace a missing Counter when ρ is already above 1. It will make the true shortage visible—and stop you from adding mystery work around a paper list.

Part-time desks, lunch, and “sometimes open” Counters

Real operations rarely run all Counters all day. Model the schedule the way customers experience it:

• Which Counters are open in the busy hour?
• Which Services can each Counter take?
• What happens when one specialist is off—does that Service pause or spill?

A clinic with four Counters on paper and two open at 11:00 a.m. has two Counters of capacity. Pricing models that charge per location regardless of open desks obscure that. Per-Counter pricing tracks the resource you actually schedule—see Esperaly pricing for how that maps to monthly cost once you know your number.

From estimate to experiment

Treat the first Counter count as a hypothesis. Run one or two busy periods with Tickets, then look at:

• Average wait by Service
• Share of Tickets waiting longer than your target
• Counter idle gaps vs stacked waits
• Missed calls / no-shows after call

If waits are long while Counters sit idle, you have a routing or Service assignment problem. If waits are long and Counters are saturated, you need capacity or shorter visits. That diagnosis is worth more than a vendor feature matrix.

For a fuller cost picture—not only Counter count but hardware and messaging traps—read how much a queue management system costs. For a same-day launch checklist once staffing is clear, see set up a digital queue in under an hour.

A practical checklist before you decide

1. Measure busy-hour arrivals for each major Service.
2. Time 20–30 real visits; do not trust the brochure average.
3. Set an explicit wait target staff can say out loud.
4. Compute utilisation for 2–3 Counter options; discard anything with ρ ≥ 1 in the peak.
5. Stress-test with +20% arrivals or +25% service time (mail-out week, trainee week).
6. Map lunch and absences; count only Counters that will actually be open.
7. Run the numbers through the wait-time estimator, then sanity-check cost on pricing.
8. Pilot with Tickets for a week before permanently changing headcount.

Bottom line

The right Counter count is the smallest number that keeps busy-hour utilisation out of the danger zone and waits inside the promise you make to the public. Arrival rate, service time, and utilisation are enough to start; industry colour (clinic, council, salon) tells you how to split Services and which desks are interchangeable.

Start with measurement, not folklore. Use a simple estimator to explore scenarios, then buy capacity—and software—that matches desks you can staff, not seats on a pricing page you will never fill.