How to Run Classes and Group Bookings — and Why a Class Is a Seat Count, Not a Timetable

BookrHub Research · Published 17 October 2026 · Capacity arithmetic and all product behaviour verified against the BookrHub application source in October 2026

If you run classes, workshops, group sessions or anything where several people share a start time, you have probably looked for a class schedule — a place to say “Pilates, Mondays and Wednesdays at 18:00, twelve places”. That screen does not exist, and looking for it is the single most common way to misunderstand how group booking works. What exists instead is a capacity number on a service. Switch on multiple bookings per slot and the slot stops being one appointment owned by one staff member and becomes a seat count shared by the whole business. Everything surprising about classes — why they appear at times you do not teach, why two instructors cannot run the same class in parallel, why a blocked-out holiday does not stop sign-ups — falls out of that one design decision. This page follows it all the way through.

A seat count, not a timetable

Most class-booking software is built around an event. You create a class, you give it a date and a time, you say how many people fit, and customers sign up to that specific instance. The class exists as an object whether or not anybody books it.

BookrHub does not work that way, and the difference matters more than it sounds. Here a class is an ordinary service with one extra property: how many people may hold the same slot at once. Turn that on and the slot stops being “an appointment with Ana” and becomes “a room with twelve seats”. Nothing is scheduled in advance. No class instance is created. The system simply offers every start time that still has a free seat, and keeps offering it until the seats run out.

This has one genuine advantage and several consequences you need to design around.

The advantage is that bookings are independent. Nobody waits for a group to assemble, nobody is told “this class needs four more people to run”, and there is no pending state to manage. Each customer books their own place and is confirmed on the spot, exactly like a one-to-one appointment.

The consequences are the rest of this page. If a class is not an event, then there is nothing to attach a timetable to, nothing to attach an instructor to, and nothing to cancel when you want to close a single session. Those three absences explain nearly every question people ask about group bookings.

The two settings that make a group service

There are only two, and they live on the service itself rather than anywhere class-specific.

Setting	What it does	Allowed values
Multiple bookings per slot	Switches the slot from one appointment to a shared seat count. This is the setting that changes the behaviour, not the number beside it.	On or off. Off by default. Requires the Business plan to take effect.
Maximum bookings per slot	How many people may hold the same slot. Required once the switch is on.	A whole number from 2 to 999

Two details are worth pulling out. The minimum is 2, not 1 — a capacity of one is expressed by leaving the switch off, which is a cleaner way to say the same thing. And the switch alone is not enough: if you turn it on without entering a maximum, the service will refuse to save until you do, which is the system declining to guess a class size on your behalf.

Everything else about a group service is just a service. It has one duration, one price, a name and a description, and it is created inactive like any other. There is no separate class area of the product, no class list, and no setting that says a service is a class — only the seat count, which is what makes the rest of the behaviour follow the service rather than a schedule.

Where class start times come from

This is the part that surprises people, and it is worth being precise about because there is no setting to change it.

Class start times are generated, not published. For a given day the system takes your opening hours for that weekday, walks forward from the beginning of each open stretch in steps of the class duration, and offers a start at every position that still has at least one free seat. It stops as soon as the next position would not leave room for the whole class before the stretch ends.

So a 60-minute class, with the business open 09:00 to 17:00, is offered at:

09:00, 10:00, 11:00, 12:00, 13:00, 14:00, 15:00, 16:00 — eight starts, the last one finishing exactly at closing time.

Not at 18:00, if you are closed by then. And — the important half — also at 09:00, 11:00 and 13:00, even if you only ever teach at 18:00. The system has no idea when you intend to teach. It knows when you are open and how many seats are free.

Because capacity multiplies the slot count, the stranded remainder costs you far more in a class business than in a one-to-one one. A 90-minute class on an eight-hour day gets five starts and leaves 30 minutes over; at twelve places, that remainder is not 30 wasted minutes but twelve wasted seats, every day you open.

If you want something resembling a timetable, the only lever is your opening hours. Shape the open stretch for each weekday so that it starts when your first class starts and ends when your last class ends, and use split shifts when you teach twice in a day — an 18:00–19:00 stretch offers exactly one 60-minute class, which is as close to “Mondays at 18:00” as the system gets. This is a genuine workaround rather than a feature, and it has a cost: those same opening hours govern your one-to-one services too, so a studio that teaches evening classes and sells daytime appointments cannot express both with one set of hours.

Whose seats are they? Not the instructor’s

For an ordinary appointment, availability is personal. Each staff member has their own diary, and the system checks that the person you picked is free before it offers you a time.

For a group service it deliberately does not do this. Capacity is counted per service, per time slot, across the whole business, with the assigned staff member ignored entirely. The practical meaning:

• A 10-place class has 10 places at 18:00. Full stop. Not 10 per instructor. If you have three teachers, the class still holds ten people.
• Two parallel classes of the same service are impossible. You cannot run Ana’s ten-person class and Marco’s ten-person class at 18:00 as one service and get twenty places.
• The workaround is two services, because the seat count is scoped to the service. Create “Pilates — Studio A” and “Pilates — Studio B”, each with its own capacity, and you have two independent counts at the same hour. Name them for whatever actually distinguishes them — the room, the level, the instructor — because the customer chooses the service before they choose anything else, so the name is doing the work a timetable would otherwise do.

There is a sharp edge to this, and it is better heard now. Because a group service checks seats instead of diaries, it does not check staff double-booking at all. Nothing stops the same instructor being attached to two overlapping class services, or to a class and a one-to-one appointment at the same time. If you split classes across services to run them in parallel, the system will not protect you from assigning one human to both.

The protection runs one way only, which is worth knowing precisely: a class booking assigned to a staff member does occupy that person’s diary for their one-to-one services, so it will hide their individual appointment slots as you would expect. It is the reverse that is missing — their individual appointments do not reduce a class’s seats, and neither does another class.

Seats are counted by overlap

A small mechanical detail with a visible consequence. When the system decides how many seats are left at a given time, it does not look for bookings with the same start time. It counts every active booking of that service whose time range overlaps the slot.

For classes that all share one duration, this is invisible: the grid steps by the duration, so slots line up end to end and each booking falls inside exactly one of them. Two classes that merely touch — one ending at 10:00, the next starting at 10:00 — do not count against each other, which is correct.

It becomes visible when a booking sits off the grid, which happens more often than you would think: a booking created by hand from the admin calendar at a time you chose yourself, or an older booking left behind after you changed your opening hours and moved the anchor. Take a 60-minute class with ten places and one booking at 10:30:

• The 10:00 slot runs 10:00–11:00, which overlaps 10:30–11:30 → nine seats left.
• The 11:00 slot runs 11:00–12:00, which also overlaps 10:30–11:30 → nine seats left.

One person has taken a seat out of two different classes. With ten places that is a curiosity; with a capacity of two it is half your advertised availability gone from two slots at once. The lesson is narrow but practical: for group services, create bookings on the grid the customer sees, and after any change to opening hours, look at the affected days rather than assuming the existing bookings still line up.

What the customer sees, and what a full class looks like

On the service list, a group service shows its capacity, so a customer knows before they pick a time that this is a class for up to twelve rather than a private appointment.

On the time picker, each available start carries its remaining seats — “8 spots left”, and correctly “1 spot left” when it is down to the last one. That last-place label is the most useful thing on the page: it is honest scarcity, calculated rather than manufactured, and it is the one moment where a class booking behaves better than a phone call ever could.

Then the part to plan around. A full class is not shown as full. It is absent. A slot is offered only while its remaining seats are above zero, so when the last place goes, the time simply stops appearing. Nobody is told the class was popular. From the outside, a sold-out 18:00 and an 18:00 you never offered look identical, and a customer scanning a thin list of times is more likely to read it as “they are closed” than “this filled up”.

And there is no waiting list. Nothing records who looked for a place they could not get, nobody is contacted when a seat frees up, and no report tells you which classes filled first. This is the single biggest gap in the feature, and it is worth being clear-eyed about: in a class business, the demand you cannot see is the demand you would most want to act on. If a class fills reliably, your options are to raise the capacity, open another start time by widening the opening hours for that day, or capture interest somewhere outside the booking system and work the list yourself.

Cancellations and reschedules

This part is handled well, and in a class business it matters more than it does for appointments, because a freed seat is only worth anything if it comes back fast enough for somebody else to take it.

A cancellation frees its seat immediately. Seats are counted by asking how many active bookings overlap the slot, and cancelled bookings are excluded from that count. There is no job to wait for and no cache to clear: the moment the booking’s status changes, the place is available again, and the slot reappears in the picker on the next load with its count up by one. If the class was full, the time comes back from being absent.

A reschedule is capacity-checked in both directions. Moving a booking into a slot is refused if that slot is already at capacity, so a reschedule cannot overfill a class, and moving a booking out frees its seat in the slot it left. Worth noting that the same staff-blindness applies: a class reschedule checks seats, not whether any particular person is free.

What you cannot do is cancel a single session. Because no class instance exists, there is nothing to cancel except the individual bookings in it. Closing one evening means cancelling each booking by hand and telling those customers yourself — and then closing the slot, or the system will cheerfully sell the places again. Deactivating the service closes it for every day at once, not just the one you meant.

Paying for a seat

If the service requires payment, three things are true, and only the third is unusual.

Each participant pays separately. Every booking is its own transaction with its own confirmation. There is no group invoice and no way for one person to pay for several places — a parent booking three children books three times and pays three times.

The amount is the full price. BookrHub has no deposit or part-payment option of any kind: a class is either free to book or paid in full at the moment of booking. This is the most commonly assumed feature in this category, so it is worth stating flatly rather than leaving to be discovered. If you want a smaller commitment up front, the only honest route is to price a cheap bookable service as a booking fee and settle the balance in person — a workaround, not a feature, with the obvious downside that your service list now contains a price that is not the real price.

An unpaid seat is released automatically, but not instantly. This is the mechanism that makes paid classes workable, because without it an abandoned checkout would hold a place indefinitely. When a booking needs payment it is given a deadline ten minutes out, and the payment page shows the customer a live countdown. A background sweep then cancels anything past its deadline, releases the seat, and emails the customer to say the booking expired.

The precise timing is worth knowing if you ever watch this happen: that sweep runs every five minutes in production rather than continuously, so a seat comes back somewhere between ten and roughly fifteen minutes after the checkout was abandoned, depending on where the abandonment fell between runs. The cancellation is also written conditionally, so a payment that completes while the sweep is mid-flight cannot have its booking cancelled out from under it — a detail that matters precisely because a class slot is contended.

One consequence for a class that is nearly full: during that ten-to-fifteen minute window, the abandoned seat is genuinely unavailable, so a class can read as sold out and then quietly have a place again a few minutes later. That is correct behaviour rather than a glitch, but it is the kind of thing worth knowing before a customer tells you about it.

What does not close a class slot

This is the sharpest edge on the whole feature and the one most likely to catch you out, so it gets its own section.

Class availability is built from exactly two inputs: your opening hours for that weekday, and how many seats of that service are already taken. That is the complete list. The capacity path never consults anything else, which means two things that look like they should close a class do not:

• Blackout dates do not close class slots. A day you have blocked out — a public holiday, a week away, a studio closure — still offers its class times and will still accept bookings. Blackouts are applied by the availability builder that one-to-one services use, and the capacity path does not go through it.
• Fully booked staff do not close class slots. A class slot is offered even when every member of staff is booked solid with individual appointments at that hour, because the seat count does not look at anybody’s diary.

So how do you actually close a day for classes? Two reliable ways, both blunt:

• Close the day in your opening hours for that weekday. This is the mechanism the class grid actually reads, so it works — but it is weekday-shaped rather than date-shaped, which means it closes that weekday generally rather than one specific date.
• Deactivate the service. Immediate and total: it closes the class for every date, not just the one you wanted, and you must remember to switch it back on.

Neither is a satisfying answer for “I am closed on 25 December”, and we would rather say so than pretend otherwise. If you run classes and take holidays, the practical habit is to check your class services whenever you block out a date, because the blackout you just created has not touched them. It is also worth saying plainly that this asymmetry looks more like an oversight in the capacity path than a deliberate design choice, and it may well change; what is written here is how it behaved when this page was checked in October 2026.

Planner: seats, stranded time and seats per open hour

Put your own numbers in. The slot and seat arithmetic matches the engine exactly; the money row is a ceiling at full occupancy, which is useful for comparing two class formats against each other and useless as a forecast.

Choosing a capacity

Set it to the number of people you can genuinely teach, equip and fit at once, and then stop thinking about it. The number does nothing except decide when bookings stop being accepted — it does not affect pricing, scheduling, staffing or anything a customer sees beyond the seats-left label.

We are not going to give you an ideal class size, because there is no credible benchmark to give. The figures that circulate in booking-software articles about optimal class capacity trace back to vendor blog posts citing unnamed industry research, several of them contradicting each other, and a number invented for a yoga studio in one country has no bearing on your room. What the arithmetic above can tell you is what each choice costs in places per open hour, which is the comparison worth making.

Three things are worth knowing before you pick:

• Setting it above your real limit is unprotected. The system counts seats; it has no idea how many mats fit. If you type 20 for a room that holds 12, it will sell 20 places and the first you hear of it is in the room.
• Setting it far below demand is invisible. With no waiting list, the classes that fill in an hour look exactly like the ones nobody wanted. You will not get a signal telling you to raise it.
• Raising it later is safe and immediate. It is one edit, it applies to every future slot at once, and it does not disturb bookings already made. So starting at your true physical limit and adjusting upward as you learn is the low-risk path — there is no reason to guess high.

One more, less obvious: capacity interacts with the stranded remainder from the duration grid, and in a class business it multiplies. A 90-minute class on an eight-hour day strands 30 minutes, and at twelve places that is twelve places a day you never offered — roughly 3,700 places a year for a studio open six days a week. Before tuning the capacity, check whether the duration divides your open stretch cleanly, because that choice moves more seats than the capacity does.

The Business plan gate, and what a downgrade does

Group booking requires the Business plan. On the Independent and Team plans the switch is visible but inert — you will be prompted to upgrade when you reach it — and a service holds exactly one person per slot however the capacity field is set.

The behaviour on a downgrade deserves spelling out, because it is quiet. The capacity setting is not erased: the switch stays on and your number stays in the field. But the check that honours them requires the Business plan, so with the plan gone, every affected service silently falls back to one person per slot. Restore the plan and your numbers take effect again exactly as they were.

What this looks like from the outside is a studio where every class suddenly has one place and then disappears after a single booking. The services screen will show capacity 12 while the booking page behaves as though it were 1, which is a confusing pair of facts to debug from the service settings. If classes start behaving like private appointments, check the plan before you check the service.

Bookings already in the diary are not touched by a downgrade. Capacity is enforced when a booking is created or moved, not re-checked afterwards, so a class of twelve that is already booked stays booked — it simply will not accept a thirteenth, or a second, until the plan is back.

One related gate worth knowing if you are comparing plans: the “any specialist” booking flow, where a customer picks a time without choosing a staff member, is also Business-plan behaviour. For classes that pairing is natural, since nobody picks an instructor to attend a class.

The 12-point group booking audit

Run this against a class service you already have, or use it as a setup order for a new one.

1. Is the plan right? Group capacity needs Business. If classes accept one person, this is the first thing to check, not the last.
2. Is the switch on and a maximum set? The switch is what changes the behaviour; the number is required once it is on.
3. Is the capacity your real physical limit? The system will sell whatever you type. Nothing checks the room.
4. Does the class duration divide your open stretch exactly? Run the planner. A remainder costs you a whole class’s worth of places every day, multiplied by the capacity.
5. Do your opening hours describe when you actually teach? They are the only lever on class start times. Anything you leave open will be offered.
6. If you teach twice a day, are those two split shifts? One long stretch offers every position in between; two stretches offer only what is inside each.
7. If you need parallel classes, are they separate services? One service is one seat count, regardless of how many instructors you have.
8. If you split classes across services, who is actually teaching each? Nothing stops one person being attached to two overlapping class services.
9. Are your service names doing a timetable’s job? The customer picks a service before a time, so the name carries the room, the level or the instructor.
10. Is the price the full amount you want charged at booking? There are no deposits; it is all or free.
11. Have you checked your class services after blocking out a date? The blackout did not touch them. Close the day in the opening hours, or deactivate the service.
12. Do you have a plan for demand you cannot see? There is no waiting list. Decide now where interest goes when a class fills, or it goes nowhere.

What BookrHub does not do

Stated plainly, because knowing the edges before you build a term’s worth of classes is worth more than finding them afterwards.

• No class schedule or recurring sessions. There is no timetable to publish and no repeating class to create. Start times are generated from your opening hours and the class duration.
• No waiting list. A full class disappears, and the people who wanted a place are not recorded anywhere.
• No per-instructor capacity. Seats belong to the service across the whole business. Parallel classes need separate services.
• No staff double-booking protection on class services. The seat check replaces the diary check rather than adding to it.
• Blackout dates do not apply to class slots. Neither do other staff bookings. Only opening hours and the seat count close a class.
• No cancelling a single session. With no class instance, there is nothing to cancel but the individual bookings — and deactivating the service closes every date.
• No group or family booking. One booking is one place. A parent booking three children books three times and pays three times.
• No deposits or part-payments. Full price at booking, or free.
• No attendance tracking, packs, passes or memberships. There is no class pass, no multi-session package, no credit balance and no check-in or attendance mark.
• No per-class minimum. You cannot require four sign-ups for a class to run; every booking is confirmed on its own.

Several of these point the same way. BookrHub’s group booking is a capacity feature on an appointment system, not a class-management product. For a studio whose day is a handful of classes with places to sell, that is genuinely enough, and the independence of each booking is an advantage over systems that make you manage class instances. For a business whose operation is memberships, passes, attendance registers and a published term timetable, a dedicated studio-management system will fit better — and it is more useful to read that here than to discover it halfway through setting up a term.

Sources and method

• All product behaviour on this page was verified against the BookrHub application source in October 2026: the service model and its capacity validations (2 to 999, required once the switch is on), the capacity slot builder that generates class start times, the overlap-based seat count and its exclusion of cancelled bookings, the booking validation that replaces the staff-overlap check for group services, the reschedule path’s capacity check, the ten-minute payment deadline and the sweep that enforces it, and the plan gating on both multiple bookings per slot and the “any specialist” flow.
• The capacity arithmetic was derived from the slot builder and then checked by re-implementing it independently: class starts, last bookable starts, stranded minutes and seats per day were computed for durations of 30, 45, 50, 60, 75, 90 and 120 minutes against a 480-minute stretch, and the closed forms (⌊L ÷ D⌋ starts, ⌊L ÷ D⌋ × C places, L mod D stranded minutes) match the builder’s output for every one of them. The planner on this page runs the same loop the builder does.
• The blackout-date and staff-availability findings in the section above were established by tracing which inputs the capacity path reads — opening hours for the weekday and that service’s own active bookings — and confirming that the builder applying blackout dates is not on that path, and that no blackout check exists at booking creation either.
• No figure for “ideal class size” or “optimal class capacity” is cited here. A search of the available material found such numbers only on booking-software vendor pages attributing them to unnamed industry research, several of which conflict with each other. They are omitted rather than repeated.

Related reading: setting up your service menu, where the duration grid comes from, booking for yoga and Pilates studios, group fitness studios, martial arts and dance studios, swim schools, taking payment at booking, tracking what you actually earned, and staff management.