A schedule starts running late long before the first client complains. The problem often begins in the service setup: an appointment that actually takes 52 minutes is entered into the system as 45; five minutes of cleaning are left out of the calculation; a final conversation runs over; the next client arrives on time and finds the day already behind schedule. Setting the right duration is not about choosing a round number. It is about finding out how much time the service really consumes and where the schedule needs margin.
The most useful rule is this: block the time needed to leave the resource ready for the next appointment, not only the time during which the client receives the main service. Preparation, execution, wrap-up, and variability need to appear in the schedule design, even if not all of those parts are visible to the client.
First, separate four types of time that look similar but are not
Scheduling systems often treat duration, interval, and start-time frequency as different settings. Reservio, for example, separates how often start times are shown to the client from the duration actually blocked by the service. Zenamu, in turn, allows preparation time before the appointment and time afterward for cleaning or a break to be reserved. This distinction matters because changing one of these variables does not automatically correct the others.
Service duration
This is the time during which that service occupies the professional or another required resource. In some businesses, this duration almost completely matches the client's time on site. In others, there are waiting, processing, or drying stages during which the client remains there but the professional may become available.
Preparation beforehand
This is what needs to happen before the appointment begins: preparing the room, separating materials, sanitizing equipment, opening the client record, reviewing information, or setting up the workstation. When this has to be done between two clients, ignoring that time is equivalent to scheduling a delay.
Wrap-up afterward
This is the time used to clean, record information, receive payment, reorganize the environment, replace materials, or complete tasks that cannot be postponed. This interval may differ from preparation time and therefore does not need to be symmetrical.
Frequency of available start times
This is the grid of start times the client sees: every 10, 15, 20, or 30 minutes, for example. A 15-minute frequency does not mean the service lasts 15 minutes or that there should be a 15-minute interval between clients. It only defines which start times can be offered. A very fine grid can create gaps that are difficult to use when service durations do not fit it well.
Measure the service as it actually happens
The most reliable starting point is real history. For a few weeks, record when the professional was ready to begin, the actual start time, the end of the main work, and the moment when the space or resource was ready for the next client. If the schedule already records some of these events, use what exists and add only what is missing.
Do not try to measure a perfect day. The goal is to capture normal days, simpler appointments, cases that took longer, and transitions that seemed small but kept repeating. One unusually long appointment should not define the entire schedule, but a delay that appears every week cannot be treated as an exception.
Also, do not use only the average. Two services can both average 45 minutes and behave very differently. One may almost always finish between 43 and 47 minutes; the other may alternate between 30 and 60. The second needs a different strategy because the variation carries over to later time slots. Queueing simulation studies in service settings show precisely that greater variation in service time can significantly increase waiting even when the average duration stays the same.
Turn the measurements into an operational slot
A simple way to think about it is: operational slot = work time that needs to block the professional + mandatory preparation + mandatory wrap-up + margin for the variation you cannot predict or classify before the booking.
The margin should not be used to hide a poor estimate. First remove from the problem everything that is predictable. If a client's first appointment normally takes longer than a return visit, treat them as two service types. If a service with a certain characteristic requires more steps, create a variant. If an add-on adds time, make that add-on carry the extra duration as well. The more predictable causes are turned into clear categories, the smaller the generic cushion between every client needs to be.
A hypothetical example
Imagine a service that, in most ordinary appointments, takes close to 50 minutes of active work. Afterward, the professional needs about five minutes to finish records and reorganize the space. Entering only 50 minutes means the next client inherits that wrap-up time. A more coherent test would be to block 55 minutes or structure 50 minutes of service plus five minutes afterward, depending on the system being used.
Now suppose that some cases of this same service frequently reach 70 minutes because of a characteristic that can be identified when the booking is made. In that situation, increasing every slot to 70 minutes creates idle time in ordinary cases. It is better to separate the longer variant, provided the client or the team can classify it before the appointment.
The interval should absorb work and uncertainty
There is no universal interval. Five, ten, or fifteen minutes may be appropriate in one operation and insufficient or excessive in another. The right interval comes from the tasks that need to happen between clients and the risk that an appointment will run beyond the main service time.
Start with what is mandatory. If there are eight minutes of cleaning, those eight minutes need to appear somewhere. If preparation can be done before opening or by another person without blocking the professional, it does not necessarily need to occupy the same resource. If administrative closing tasks can be completed in batches without affecting the next appointment, part of that work may not need to become an individual buffer.
Then look at variability. When most appointments fit within the main service time but some exceed it by a few minutes for no identifiable reason, the buffer can act as a shock absorber. If the overrun is frequent and large, simply increasing the interval is a sign that the service duration itself is underestimated or that different service types are being mixed under the same name.
Not all of the client's time needs to block the professional
Some services have a passive stage. A classic example is a procedure in which the professional applies a product, waits for it to process, and then returns to finish. Square describes this structure as initial time, processing time, and final time, allowing the processing phase to be treated differently from the professional's active occupancy.
This separation avoids two opposite mistakes. The first is blocking the professional for the client's entire stay and losing capacity when there is a genuinely free period. The second is releasing the time too early and discovering that the final stage of two clients overlaps. The right design needs to consider who or what is occupied in each phase: the professional, room, chair, equipment, or another resource.
When to create different durations for the same service
A different duration makes sense when the time difference is predictable before scheduling. First visit versus return visit, simple versus complete version, known size or complexity, the presence of an add-on, or the need for specific equipment are examples of factors that may justify variants.
If the difference is discovered only in the middle of the appointment, creating many options for the client does not solve the problem. In that case, the schedule needs margin or an earlier screening step that turns that uncertainty into information before the booking. The goal is not to have dozens of nearly identical services, but to avoid placing appointments with very different behavior inside the same time block.
What to do when there is no history yet
A new service requires an estimate, but that estimate should come with a review date. Mentally separate the technical service time, preparation, wrap-up, and uncertainty. Configure a block that is conservative enough to keep the first appointments from pushing the entire schedule back, and start measuring from day one.
After you accumulate enough cases to see a pattern, adjust it. If time is left over almost every time, find out which part was overestimated before reducing everything at once. If delays repeatedly appear, identify which stage causes them. A common mistake is to react to delays by adding minutes at the end without knowing whether the problem is in the main service, the transition between clients, late arrival, or a service variant that should have its own duration.
Do not confuse a full schedule with an efficient schedule
A schedule with no open space may look productive, but it is fragile. In systems with variability, operating continuously at the limit causes small deviations to accumulate. Research on queueing and scheduling shows that variation in arrival and service duration can produce growing waits over the course of a shift, even when the average duration appears compatible with the planned interval.
That does not mean adding large breaks after every client. It means recognizing that useful capacity is not the same as 100% time occupancy. In operations where delays tend to accumulate in the second half of the period, it may make more sense to test a recovery window at a strategic point in the shift than to spread a large margin across every service.
Adjust the schedule by watching three signals at the same time
The first signal is punctuality: how many appointments start on time or within the tolerance the business considers acceptable? If the schedule starts well and always ends late, variation is accumulating. If it starts late from the beginning, the cause may be initial preparation, client arrivals, or the first slot.
The second signal is usable idle time. It is not enough to count empty minutes. A 12-minute gap between two appointments may exist on the calendar and still be useless for selling another service. That is why the frequency of offered start times needs to work with the registered service durations. Reservio's documentation specifically calls attention to the risk of creating idle gaps when the start-time grid and duration do not fit together well.
The third signal is shift overrun. If the professional regularly finishes after closing time even with few client delays, the slot design is probably too compressed. If the day often ends much earlier, there may be room to reduce margins, reorganize variants, or offer start times on a more efficient grid.
Make small adjustments and compare before and after
Avoid changing every service at the same time. Choose the services that cause the most delays or account for the greatest volume, adjust the duration or buffer, and observe the behavior over a comparable period. What matters is whether delays decrease without turning the schedule into a sequence of empty spaces.
When the system allows it, prefer recording preparation and post-service time separately instead of adding everything to the duration shown to the client. This makes the logic more transparent and makes future changes easier. When the system does not provide those fields, the same reasoning can be applied by incorporating operational time into the total block.
The right duration needs to work until the end of the day
A well-sized slot is neither the shortest time in which the service has ever been completed nor the worst case ever recorded. It is a duration that represents the normal appointment, separates predictable variations into their own service types, and preserves enough margin so that small differences do not become a chain of delays.
Before publishing the schedule, answer three questions: how long does the service really occupy resources, what needs to happen before and after it, and which duration differences can be identified before booking. When those answers become part of the configuration, the interval stops being a guess and becomes an operational tool. The schedule may even look a little less tightly packed on paper, but it tends to be much more usable in day-to-day work.
