Appointments nobody turns up to: reducing no-shows without hounding the customer
There is a loss that appears on no profit-and-loss statement: the hour you had booked for someone who did not come. There is no invoice to issue and no expense to justify, but the mechanic was there, the bay was there, and the morning passed all the same. And unlike material left in stock, an empty hour is never recovered.
In many businesses that work by appointment — workshops, physiotherapists, hairdressers, accountants, vets — no-shows sit between 10% and 20% of visits. Translated: one day in seven, or more, you work for nobody.
It is almost never bad faith
It is worth pausing on this, because the solution depends on the cause. When you ask customers who did not turn up, the answers repeat:
- They forgot. They booked three weeks ago by phone and wrote it down nowhere.
- They did not know how to let you know. Calling to say "I can't make it" is a chore and often embarrassing; if the line is busy, they do not try again.
- They said yes out of politeness and never seriously intended to come.
- Their day changed and it did not occur to them that the appointment could be moved.
None of these four is solved by scolding the customer or charging a penalty up front. Three of the four are solved with information at the right moment and an easy way out.
The three moments of a reminder that works
A reminder system is not "send a message the day before". It is three moments, and each does a different job:
- When the appointment is booked. An immediate confirmation with day, time, address and what to bring. This is where the customer writes it down: if they do not get it in writing, it does not exist.
- 48 to 72 hours before. This is the important reminder, and almost everyone skips it. It is the only moment when you are still in time to refill the slot if the customer says they cannot come. A reminder the day before is too late for you.
- The same morning. Short, practical: time and place. This prevents the last-minute lapse, not the cancellation.
If you can only set up one, set up the 48–72 hour one. It is the one that changes your number.
Make it easy to cancel (yes, really)
It sounds counter-intuitive, but it is the part that prevents the most no-shows: every reminder has to carry a one-step way of saying "I can't" or "I'd like to change it". A link, a button, a one-word reply.
A customer who cancels three days ahead is not a lost customer: it is an hour you can sell again. What hurts you is not the cancellation, it is the silence. Making it hard to cancel brings nobody in; it only turns cancellations into no-shows.
And when a slot frees up, the natural next step is to notify the waiting list. Here a simple automation does in two minutes what nobody ever does by hand.
A concrete example. A physiotherapy centre with three rooms had a 16% no-show rate. They did not change their booking software: they added three automated notices (confirmation, a 72-hour reminder with a reschedule button, and a same-morning notice) and a waiting list by time slot. Three months later, no-shows were at 7%, and half the slots that opened up were refilled the same day. Nobody had to make more calls; they just got fewer surprises.
What the automation must never do
This is where these systems become unpleasant. Three rules:
- Never send more than three notices per appointment. The fourth is no longer a reminder, it is harassment, and the customer mutes the number.
- No messages at unsociable hours. A reminder at 10:40 p.m. helps nobody and looks terrible. Working hours, always.
- Let people opt out of notices and respect it the first time. Whoever prefers a phone call gets a phone call.
And a fourth, more important than the other three: if the customer replies anything other than "confirm" or "cancel", a person has to read it. "I'm in a lot of pain, can it be brought forward?" is not something an automation can handle.
And your customers' data?
We are talking about names, phone numbers and, depending on the sector, reasons for the visit — which may be health data. That is not just any list. It is worth setting conditions from day one:
- Data stored inside the European Union.
- Not used to train third-party models.
- The reminder message carries no more than it needs: time and place, not the diagnosis.
- A retention period that has actually been decided, after which the data is deleted.
This article is general information, not legal advice for a specific case.
About this article
iaetica.cat is a space for writing in the open: what I am learning about artificial intelligence, its ethics and its law.
— Laura Graells