The policy you have covers the cheaper failure
Almost every service business has a no-show policy. It is usually a line on the invoice or the booking confirmation, it usually mentions a fee, and it exists because a no-show is memorable - a technician stood on a porch, knocked, called, waited fifteen minutes, and drove away with nothing.
Far fewer have a cancellation policy, and almost none have a reschedule policy. As of September 2026 that is still the wrong way round, because the expensive failure in most schedules is not the no-show. It is the customer who calls at 7:50 in the morning to cancel an eight o'clock appointment, when the technician is already fifteen minutes into the drive and the day is already built. That call costs you more than a no-show, it happens more often, and in most systems it is recorded as nothing at all. The appointment simply disappears from the calendar and the day continues one job lighter.
This guide prices that hole and closes it. It separates cancellations from reschedules from no-shows and explains why each needs a different response, works the arithmetic on what a same-day cancellation actually costs, covers when a cancellation fee helps and when it just buys you a bad review, shows how to make rescheduling so frictionless that it becomes the default outcome, sets up the standby list that turns an empty slot back into revenue, covers deposits for the jobs that genuinely need them, and ends with a policy paragraph short enough that a customer will actually read it. The scheduling machinery underneath runs on Run with Jarvis, where booking, reminders, the phone line and the calendar are the same system.
Three events, three different responses
Lumping these together is the root mistake, because they differ in what they cost and in what you can do about them.
A cancellation ends the appointment with no replacement date. The revenue is gone unless you win it back later, and the slot is empty unless you fill it. It is the only one of the three where you lose both the time and the customer relationship in one move, and it is the one most likely to indicate something you could have prevented - a quote that felt too high after they thought about it, a competitor who could come sooner, a confirmation that arrived so late they assumed you had forgotten.
A reschedule moves the appointment. The revenue survives. The cost is the disruption to a day that was already routed, which is real but bounded, and a reschedule taken with notice frequently costs you nothing at all. Every reschedule is a cancellation that did not happen, which is why a reschedule rate going up while a cancellation rate goes down is a win rather than a problem.
A no-show is an appointment nobody cancelled. Your technician arrives and there is no access, no customer, or no answer. It costs you the drive both ways plus the wasted window, and it is the most annoying because it is entirely avoidable through confirmation. It is also, in most shops, the least frequent of the three - which is why building your whole policy around it is aiming at the wrong target.
The response differs accordingly. A no-show is a confirmation problem, and the fix is reminders that actually get read, covered in the appointment reminders guide. A reschedule is a friction problem, and the fix is making the new date easy to pick. A cancellation is partly a value problem and partly a capacity problem, and the fix has two halves: try to convert it into a reschedule while you still have the customer on the line, and if you cannot, fill the slot with somebody else.
What a same-day cancellation actually costs
Here is the arithmetic. Treat every figure as an illustration to replace with your own loaded rate, average ticket and margin, because the shape of the number matters more than the number.
Take a three-hour morning slot. Your fully loaded technician cost - wages, payroll burden, vehicle, insurance - is 68 dollars an hour, so the capacity you set aside is worth 204 dollars in cost terms. But cost is not the right measure of what you lost. What you lost is the gross profit on the job that would have filled it. Say the average job in that slot invoices at 385 dollars at a 52 percent gross margin, which is 200 dollars of gross profit that will not now exist.
Then add what you already spent. The technician was 22 minutes into the drive when the call came, which at 68 dollars an hour is about 25 dollars of loaded time that produced nothing. Add the dispatch rework - roughly fifteen minutes of somebody rerouting the day, calling the next customer to see if they can go earlier, and updating the board - at a loaded office cost of 28 dollars an hour, so about 7 dollars.
That is 232 dollars for one same-day cancellation, and the number is conservative because it assumes the rest of the day absorbs the gap cleanly. In a tightly routed day it does not; a hole in the morning often means the technician arrives at the afternoon job early, waits, and burns another half hour.
Now scale it. Six same-day cancellations a month is 1,392 dollars, or roughly 16,704 dollars a year. That is a number most owners have never seen, because nothing in their system adds it up. It sits in the gap between "we were a bit slow last week" and "the calendar looked full."
Cancellations at distance are cheaper because you can refill the slot, which is precisely why notice is the variable your policy should be built around. And if your routing already has long drive legs, every cancellation costs more than this arithmetic suggests, because the drive you cannot recover is longer - a point covered in the service radius and drive time guide.
The four scenarios, priced
| Scenario | What it costs you | Recommended response | What they should have been told at booking |
|---|---|---|---|
| Cancelled more than 24 hours out | Usually nothing - the slot is refillable at normal notice | Offer a new date immediately, release the slot to booking, no fee | That you appreciate notice and can hold a new time right now |
| Cancelled same day | Full gross profit on the slot plus drive already started plus dispatch rework | Convert to a reschedule on the call, then work the standby list; fee only if committed or repeat | That same-day changes may carry a fee, and how much notice avoids it |
| Reschedule with any notice | Routing disruption only, revenue survives | Make it one tap, confirm the new date instantly, never charge for it | That they can move the appointment any time from the reminder link |
| No-show, never cancelled | Drive both ways plus the wasted window plus a follow-up cycle | Contact the same day, rebook once, then apply the trip charge on repeats | The arrival window, who must be present, and the trip charge if nobody is |
Two things in that table are doing most of the work. The last column is the one businesses skip, and it is the reason fees turn into arguments - a fee the customer learns about at the moment you charge it reads as a penalty, while a fee they were told about at booking reads as a term. And the reschedule row deliberately has no fee in it, because charging for a reschedule teaches customers to cancel instead, which is the opposite of what you want.
When a fee helps and when it costs you the customer
A cancellation fee is a tool with a narrow correct use. It works when the slot was genuinely committed and hard to refill, and it works as a deterrent on people who have done it before. It fails badly everywhere else.
Charge it when you ordered parts specifically for the job, when you scheduled multiple technicians or a full day, when the customer has cancelled late more than once, or when the appointment was an after-hours or premium window somebody else wanted. In those cases you have committed real money or real scarcity, and a fee is the reasonable expression of that.
Do not charge it on a first-time customer who cancelled with notice and apologized. The arithmetic there is straightforward: a 45 dollar fee against a customer whose lifetime value is several hundred dollars in future work plus whatever their referrals are worth. Collecting the 45 dollars is one of the worse trades available to you, and it converts a mildly awkward moment into a permanent one.
Three rules keep a fee from doing damage. Disclose it at booking, in the confirmation, and in the reminder, so it is never a surprise. Give the technician or the office the discretion to waive it without asking permission, because a fee you always enforce mechanically will eventually be enforced on somebody whose father just died. And always offer the reschedule before the fee - the sentence is some version of "I can move this to Thursday morning at no charge, or if you would rather cancel outright there is a fee, which would you prefer?" A large share of people take Thursday.
The deeper point is that a fee is compensation, not a solution. It converts a lost slot into a small payment. Filling the slot converts it into a whole job, which is worth roughly five times more, and that is where the effort belongs.
Make rescheduling easier than cancelling
Most cancellations are not decisions to stop being your customer. They are decisions made at the moment somebody realizes the time no longer works, and the option they take is whichever one is in front of them. If the only obvious action is "cancel," you get a cancellation. If a reschedule link is right there showing real times, a large share of those become date changes.
That means the reschedule path has to be genuinely frictionless. Every reminder message should carry a link that opens actual availability rather than a form that submits a request to a human. The customer picks a time, the calendar updates, the old slot returns to availability, and nobody in your office touches it. If the customer has to call during working hours to move an appointment, you have built the friction that produces cancellations, because calling a business at 2pm from a job of their own is exactly what people avoid.
The other half is the phone. Cancellations arrive by phone constantly, and often outside office hours - the evening before, or at 7:50 in the morning. An unanswered cancellation call is the worst outcome available: you keep the appointment on the board, you send a technician, and you get a no-show on top of a cancellation. On Run with Jarvis the AI receptionist takes those calls around the clock in English and Spanish, and the important behaviour is that it does not simply accept the cancellation - it offers a new time first, because the caller is right there and the alternative is losing them entirely. The mechanics of how that booking conversation works are covered in how AI appointment booking works, and the broader comparison of self-service against the phone is in online booking versus phone calls.
For the calls that still go unanswered, an immediate text back closes the loop with the customer before they call your competitor, which is the entire argument in missed call text-back.
The standby list turns the hole back into revenue
Every cancellation creates a slot that somebody else wanted. Most businesses never offer it to them.
A standby list is the simplest fix in this whole guide. When a customer books further out than they would like, or asks whether anything sooner might open up, you record that they would take an earlier slot and roughly what times work. When a cancellation happens, you have a list of people who already want that exact window rather than starting from nothing.
Three things make it work. Keep the list short and current, because a name added six weeks ago has probably already solved their problem elsewhere and calling them wastes the twenty minutes you have. Reach them fast, because a slot that opens at 7:50 for an 8:00 appointment is worth almost nothing by 9:15 - speed is the whole product here. And contact several at once by text rather than working down a list by phone, first reply takes the slot; a sequential phone list is too slow for a same-day gap.
This is one of the places where outbound automation earns its place. The platform's AI outbound follow-up calls can work a standby list by phone when a slot opens, which is follow-up on your own existing customers rather than cold prospecting, and it is the difference between an office intending to call eight people and actually reaching them inside twenty minutes. See the pricing page for how outbound capability sits across the plans.
The payoff is worth stating plainly. Using the earlier illustration of 232 dollars per same-day cancellation and six a month, filling half of them recovers roughly 700 dollars a month, about 8,400 dollars a year, from customers who were already asking to come sooner. No marketing spend, no new leads, just a list and a fast message.
Deposits, for the jobs that actually need them
A deposit is a different instrument from a fee. A fee punishes after the fact; a deposit prevents by making the customer commit before you spend anything.
Take one when you are committing money or scarce capacity ahead of arrival. Special-order parts are the clearest case, because a cancelled job leaves you holding a part you may not be able to return. Equipment installations and multi-technician days are the next clearest, because you are allocating a large block of capacity that cannot be refilled at short notice. Bookings scheduled several weeks out are a softer case but a real one, since intent decays with time.
Do not take one for a routine diagnostic visit. The deposit adds friction at exactly the moment you are trying to convert a caller into a booking, and it protects almost nothing, because a one-hour diagnostic slot is the easiest thing in your schedule to refill.
Sizing is simple. Cover the part you cannot return plus a meaningful fraction of the labor commitment - enough that cancelling costs the customer something real, not so much that booking feels like a purchase. Say what it applies to, that it comes off the final invoice, and under what conditions it is refundable, and say all of it before you take the money. Taking the deposit at the moment of booking, by link, while the customer is still engaged, is the difference between a policy and an intention; general consumer-disclosure guidance for businesses is available at ftc.gov.
The policy, in one paragraph
Long policies do not get read. Here is the shape of one that does, as an illustration to adapt rather than copy - and have your own attorney review anything you plan to make a term of business, since consumer rules vary by jurisdiction.
We hold your appointment window for you, so please give us as much notice as you can if plans change. You can move your appointment any time, free, using the link in your confirmation or reminder. If you cancel outright on the day of service we may apply a [fee] to cover the reserved time. If nobody is available at the property during your arrival window, a [trip charge] applies. Jobs that require special-order parts take a deposit at booking, which is applied to your final invoice.
Four sentences, and every one of them is doing a job. Notice is requested rather than demanded. Rescheduling is named first and explicitly free, which is the behaviour you want. The same-day fee is disclosed with the word "may," which preserves your discretion to waive it. And the deposit is explained as applied rather than extra, which is what customers actually want to know.
Put it in the booking confirmation, the reminder, and on the quote. Not the footer of your website, where it exists only to be cited in an argument.
Measure the three separately
If cancellations, reschedules and no-shows all land in the same bucket in your system, you cannot manage any of them. Split them at the point of recording and track three rates against total booked appointments.
Cancellation rate, split by notice - more than 24 hours versus same day - is the number that carries the money. A rising same-day rate is worth investigating quickly, because it usually has a cause: quotes that look different in writing than they sounded, arrival windows that are too wide, confirmations arriving so late the customer assumed you had forgotten.
Reschedule rate is not a failure metric, and treating it as one is a mistake. If it climbs while cancellations fall, your reschedule path is working exactly as designed.
Standby fill rate - the share of cancelled slots you refill - is the one that tells you whether the recovery half of this is real. It will start near zero and it is the fastest to move.
Record a reason on cancellations, from a short fixed list rather than free text, and read them monthly. Reasons cluster, and clusters are fixable in a way individual cancellations are not.
Put it together
The no-show policy you already have covers the failure that costs you least. The one that costs most - the same-day cancellation - is usually unpriced, unmeasured, and met with nothing but a sigh and a rerouted van.
Fix it in four moves. Price it, so the conversation stops being about inconvenience and starts being about the roughly 16,700 illustrative dollars a year it represents. Separate the three events in your records so you can see which one is actually moving. Make rescheduling one tap and always offer it before a fee, because a moved job is worth five times a collected fee. And build the standby list so an empty slot goes back out to somebody who already asked for it, fast enough to matter.
Compare current plans on the pricing page, and if you want help wiring reschedule links, standby outreach and cancellation reasons into your own booking flow, get in touch.
