Pull the three free levers before you charge anyone. Shorten how far ahead people can book, send one reminder about a day before, and reply fast when somebody misses. Each of those has a measured effect in the research, and none of them costs a client anything. A cancellation fee is the fourth lever, and it works - on the people who were coming anyway.
The reason to order them that way is not softness. It is that the first three change behaviour, and the fourth mostly changes who stays.
Key takeaways
- Booking horizon is the biggest free lever: attendance falls sharply as the gap between booking and appointment grows.
- One reminder about a day ahead moves non-attendance by roughly half in the pooled evidence.
- The channel barely matters, and the cost difference between channels is enormous.
- Speed after a miss roughly doubles the chance the client comes back.
- The tools let you charge from one hour's notice, and almost nobody tells clients the rule exists.
Lever one: stop booking so far ahead
This is the lever people never touch, and it has the largest documented effect.
Across 46,655 appointments at one eye clinic, no-shows ran at 9.1% (lead-time analysis, 2015) when the booking was made within two weeks, and reached 38.3% (same analysis) at a six-month horizon in the same clinic with the same staff.
A separate analysis of 332,281 scheduling transactions found nonarrival rising about 16% (scheduling horizon analysis, 2020) for every extra 30 days of horizon, and passing 50% (2020 analysis) once the gap exceeded 90 days.
Nothing about those clients changed. Only the distance did. If your booking page is open twelve months ahead because nobody chose otherwise, that default is quietly generating your no-show rate.
The practical move is to shorten the horizon for most services and keep a longer one only where clients genuinely plan far ahead. Then use a waitlist to absorb the demand that no longer fits, which is covered in what a waitlist has to store.
Lever two: one reminder, about a day before
The pooled evidence here is unusually clear. Across 29 studies, median non-attendance fell from 23.1% (systematic review, 2011) to 12.5% (same review) once reminders existed, at an average cost of €0.41 per reminder.
The channel is where money gets wasted. In a randomised trial of 6,450 patients who all received a reminder 24 hours ahead, missed appointments came in at 11.7% (non-inferiority trial, 2013) for text messages against 10.2% (same trial) for phone calls. Over the same period the texts cost €230 in total and the calls cost €8,910, of which €8,700 (2013 trial) was administrative labour.
So: automate the reminder, keep the phone for the appointments you genuinely cannot lose, and stop treating the channel as the decision. The timing is the decision, and it is covered in when to message a client about their next visit.
Lever three: move fast after a miss
A missed appointment is not the end of the relationship unless you treat it as one.
In a randomised trial, a prompt message after a no-show brought patients back within thirty days 22.2% (randomised clinical trial, 2024) of the time, against 11.6% (2024 trial) for the standard letter alone. Roughly double, from speed rather than cleverness.
This is also the moment where punishing and recovering diverge most sharply. The message that recovers a client and the message that charges them are not the same message, and sending the second one first ends the relationship for the price of one appointment.
| Lever. | Documented effect. | What it costs the client. |
|---|---|---|
| Shorter booking horizon. | 9.1% against 38.3% across two weeks versus six months. | Nothing. |
| One reminder a day ahead. | Median 23.1% down to 12.5%. | Nothing. |
| Fast reply after a miss. | 22.2% return against 11.6%. | Nothing. |
| Cancellation fee. | Not measured in these studies. | Money, and sometimes the relationship. |
The fourth lever: fees
Fees are legitimate. A late cancellation costs you a slot you cannot refill, and asking for a card is a normal way to make a promise real.
But two things are true at once. A fee is the only lever on this list that a good client can be caught by - the one whose child got ill, whose train was cancelled, who has been coming for three years. And the effect of fees is not what the three studies above measured, so anyone quoting a percentage reduction from a fee is quoting marketing rather than evidence.
The way to use the fourth lever without losing people is to make it visible and forgiving. State the window on the page. Apply it consistently. And keep the discretion to waive it, because the client who is charged once and forgiven once tells a very different story than the client who is simply charged.
[ORIGINAL DATA] The tools are precise, and almost nobody says so
Methodology. On 19 August 2026 we read the public help documentation of two booking platforms for what their cancellation and no-show features actually allow. We then cross-referenced that against the 62 real salon booking pages audited in what an online booking page needs. Sample size: two vendor documentation sets, 62 live pages. We quote only wording printed on the vendor's own page.
The capability is far more precise than most businesses realise.
Square documents three fee shapes: a flat fee per appointment, a flat fee per service, or a percentage of the service price. On the notice window it is explicit: "You may select between one hour to two weeks or none". On recovery it states, "You can charge a cancellation fee up to 14 days after the appointment". The feature is limited to Appointments Plus and Premium subscribers and to Square Plus and Premium sellers with advanced bookings added. No maximum fee or percentage is stated anywhere on the page.
Fresha documents the charging step rather than the limits, instructing the business to "Enter the amount to charge the client, up to the limit agreed in your payment policy", with fees able to be "set up as a percentage beforehand".
Now set that against the pages. Of the 62 live salon booking pages we audited, the word "cancel" appeared on 3, a deposit was mentioned on 2, and "no-show" on 1.
| Signal. | Finding. |
|---|---|
| Notice window configurable from. | One hour, per Square's documentation. |
| Card chargeable after the appointment for. | Up to 14 days. |
| Maximum fee documented. | None stated. |
| Live pages mentioning cancellation. | 3 of 62. |
[UNIQUE INSIGHT] The mismatch is the finding, and it runs the opposite way to the usual complaint. The industry conversation assumes businesses are too aggressive with fees and clients are being nickel-and-dimed. What the documentation and the pages together show is a capability that can be set to an hour's notice and charged for a fortnight afterwards, sitting almost entirely unannounced - three pages in sixty-two. That combination is the worst of both worlds. The client cannot see the rule, so it cannot deter anything, which is the whole point of a policy. And when it is eventually applied it arrives as a surprise, which is precisely the condition under which a reasonable charge feels like a punishment. A fee nobody was told about does not protect the calendar. It only converts one lost slot into one lost client. If you are going to switch the fourth lever on, the sentence on the page is not the optional part. It is the part that does the work.
What "not punishing good clients" means in practice
- Shorten the booking horizon before you touch anyone's card.
- Send one reminder about a day ahead, automatically, on a channel they read.
- Reply within hours of a miss, with an offer rather than an invoice.
- State the cancellation window in one sentence on the page and in the confirmation.
- Keep the right to waive it, and use that right for people with history.
The first three cost nothing and carry the measured effects. The last two make the fourth lever usable without turning it into a trap.
Definitions used in this article
A no-show is an appointment the client neither attended nor cancelled, such as a booked slot that simply passes empty.
Booking horizon is how far into the future clients may book, and it changes attendance rather than just convenience.
A cancellation window is the period before an appointment in which a client may cancel without a charge, for example twenty-four hours.
No-show protection is a card held against a booking so a fee can be charged if the client does not attend.
Waiving is choosing not to apply a stated fee in a specific case, and it is what keeps a policy from costing you regulars.
FAQ
What reduces no-shows the most?
Shortening how far ahead people can book. Attendance fell from 9.1% to 38.3% as the horizon stretched from two weeks to six months (2015 lead-time analysis).
When should the reminder go out?
About a day before. A trial of 6,450 patients used 24 hours and reported missed appointments near 10 to 12% (2013 non-inferiority trial) on both channels tested.
Should I charge a cancellation fee?
It is legitimate, and it should be the fourth thing you do rather than the first. State it on the page, apply it consistently, and keep the discretion to waive it.
How short can the notice window be?
Square documents a range "between one hour to two weeks or none", so the tooling is far more precise than most policies written with it.
How long can a card be charged afterwards?
Square states up to 14 days after the appointment. Whether you should use the full window is a separate question from whether you can.
What should I send right after a no-show?
An offer of a new time, quickly. A prompt message roughly doubled thirty-day return against a standard letter (2024 trial).
Where this fits
The page-side view of all this is in what almost every booking page gets wrong, where the cancellation sentence turned out to be the single biggest gap. Reminder timing is in when to message a client about their next visit, and what to do with the slot a cancellation frees is in how to fill a last-minute cancellation.
The reminder lever described here sits in the Belom messaging automations, where timed reminders send within consent and contact limits.
The fourth lever, the one that costs the client money, is booking deposits and prepaid balances in Belom.
Editorial note: written by the Belom product team from how the product actually works, not from vendor marketing. More about Belom, or contact us with a correction. Reviewed 19 August 2026.