How to reduce appointment no-shows

Search for no-show advice and you'll get the same article forty times: a big scary number, seven tips, a call to action for whichever booking tool published it. The numbers get copied from one post to the next until nobody can tell you where they came from, and the tips are ordered by what's easiest to sell rather than what actually works.

This is the version I wanted to read. It separates the levers with real randomized-trial evidence behind them from the ones where the only evidence is a vendor's own blog, and it puts them in the order I'd actually apply them. I build booking software, so I have an obvious interest here — which is exactly why I've flagged where the evidence for my own product category is thin.

What the data actually says

Three things are worth knowing before you plan anything.

The famous cost figure is softer than it looks. You'll see "no-shows cost the US healthcare system $150 billion a year" everywhere. It traces back to a 2020 Health Management Academy report and gets recycled endlessly, with the underlying method — roughly, an average appointment value multiplied by an assumed national no-show rate — rarely stated. It's a reasonable order-of-magnitude estimate, not a measurement. Treat it as a headline, not a data point, and don't build a business case on it.

Aggregated "average no-show rate" numbers are mostly vendor surveys. The commonly quoted figures — around 23% across industries, 27–30% for healthcare and dental, with 5–10% cited as a healthy target — come from statistics round-ups rather than primary research, and they're bundling wildly different businesses. Your own rate over the last 90 days is more useful than any of them.

The reminder research is real, and modest. This is the part that actually holds up. A pragmatic randomized study at Kaiser Permanente Washington, published in The Permanente Journal, found that an additional text reminder cut the chance of a no-show by roughly 7% for primary care visits and 11% for mental health visits — relative reductions, not percentage points. A more recent randomized trial reported in NEJM Catalyst found that adding an automated voice call on top of text reminders moved the no-show rate from 11.3% to 9.6%. And a randomized controlled trial in a paediatric clinic found a much larger effect — 23.5% no-shows in the reminder group versus 38.1% in the control group.

Notice the spread. The same intervention produced a few percent in one population and a third of the no-shows in another. That's the real lesson: effect size depends far more on who your customers are than on which tool sends the message. A paediatric clinic serving families with chaotic schedules has enormous headroom. A barbershop with a loyal regular book does not.

Why people actually miss appointments

Three different causes, three different fixes. Most no-show advice fails because it applies one fix to all three.

Forgetting. The person intended to come and lost track. This is the cause reminders solve, and it's why reminder effects are biggest where booking-to-appointment gaps are long and schedules are chaotic. If someone books three weeks out, forgetting is your dominant cause.

Low commitment. The person booked casually, with no real intention or a weak one — the "I'll book it and decide later" appointment. Reminders barely touch this. Deposits and friction at booking do, because they filter at the moment of intent rather than reminding someone who was never coming.

Friction in cancelling. The person knew they couldn't make it and didn't tell you, because telling you meant calling during business hours and having an awkward conversation. This one is entirely your fault and entirely fixable, and it's the most under-addressed of the three. A cancellation is not a no-show: you get the slot back and can refill it.

Before choosing levers, work out which of the three you actually have. If most of your no-shows are people who booked a week ago and vanished without a word, you have a cancellation-friction problem and no amount of deposit policy will fix it.

The playbook, in order

Reminders: the details that matter

The evidence supports reminders. It doesn't say much about the specifics, so this is where judgement comes in.

Timing. One reminder roughly 24 hours before is the workhorse — far enough ahead that the person can still cancel and you can still refill, close enough that they won't forget again. For appointments booked more than a week out, add a second at booking time; the confirmation is also a reminder that the thing exists.

Channel. SMS outperforms email for adherence in most published work, and the NEJM Catalyst trial suggests a voice call adds something on top of texts for higher-risk groups. But channel cost matters: SMS is per message, and at a few thousand appointments a month that's a real line item. Email plus SMS only for high-value or first-time bookings is a reasonable compromise.

Content. Date, time, location, what to bring, and — critically — a link that cancels or reschedules in one tap. A reminder without a cancel link converts "I can't make it" into a no-show, because the alternative is phoning you.

Frequency. More reminders are not linearly better, and past a point they read as nagging and get muted. Two is plenty for most service businesses.

Making cancellation easy is not a concession

This is the recommendation people push back on hardest, so it's worth being direct: your cancellation rate going up is not a bad outcome.

A no-show is a slot that earns nothing and that you learn about when the customer doesn't walk in. A cancellation 24 hours out is a slot you can offer to someone else. The same person, the same missed appointment, and one of them is recoverable. Every barrier between "I can't make it" and you knowing about it converts the recoverable one into the unrecoverable one.

So: a self-service customer portal or a cancel link in every message, no login, no phone call, no business hours. Then measure no-shows and cancellations separately, because if you only track "appointments that didn't happen" you'll read a genuine improvement as a decline.

Deposits: probably effective, poorly evidenced

Deposits are the lever most booking vendors push hardest, and I'd be doing exactly the same thing if I skipped this section. So, plainly: I could not find peer-reviewed randomized evidence that deposits reduce no-shows in service businesses. What exists is vendor-reported — restaurant platforms citing 40–55% reductions after introducing deposits, one reservation platform reporting a 1.7% no-show rate among customers taking deposits, salon software reporting rates below 3%. These are almost certainly biased upward by selection: the businesses that adopt deposits are the ones with the operational discipline to do a lot of other things right too.

That said, the mechanism is sound. Deposits work on commitment rather than memory, which is the one cause reminders can't touch, and a customer who has paid something has a reason to show up or at least to tell you. My honest read is that deposits do work, probably better than reminders, and that the published effect sizes are inflated.

Practical guidance that follows from that uncertainty:

• Apply them selectively. First-time customers and long or expensive services. Charging a regular a deposit to rebook the appointment they've kept monthly for two years is a good way to annoy someone who was never the problem.
• Keep them small. Enough to create commitment, not enough to become a barrier. Roughly 20–30% of the service value is the commonly cited range, and it's a reasonable starting point.
• Make the refund rule explicit and generous inside the window. Full refund if cancelled with 48 hours' notice, forfeited after. State it at the point of booking.
• Measure conversion, not just no-shows. A deposit that cuts no-shows by a third while cutting bookings by a fifth has made you poorer. Track both, and compare like-for-like periods.

If you're running deposits on Shopify specifically, the mechanics are covered in taking deposits on Shopify bookings. In Opencals, deposits and cancellation windows are configured per service on every plan, so running this as a controlled experiment on one service is straightforward.

Which levers are worth what

The last row is worth dwelling on. Retrospective no-show fees are popular in policy documents and rare in practice, because collecting money from someone who has already demonstrated they don't want to deal with you is difficult, and the attempt costs goodwill. A deposit taken up front solves the same problem without a collections conversation.

Measure it properly

Three mistakes make no-show data useless, and almost everyone makes at least one.

Mixing cancellations into the no-show number. They're different events with different fixes. If your reporting collapses them, an improvement in one hides a regression in the other.

Using the whole business as the unit. No-shows cluster. One service, one staff member, one time of day, one booking channel — that's usually where the problem lives. An analytics view that lets you segment turns "we have a 14% no-show rate" into "first-time clients booking the long treatment on Saturday morning have a 31% rate", which is a fixable statement.

No baseline. Changing three things at once and looking at last month's number tells you nothing. Take a 90-day baseline, change one lever, wait long enough to have a comparable sample, then read the result.

How this is set up in Opencals

Nothing here is unique to one platform — every lever above exists in most booking software, and the tool matters less than whether you actually configure it. For completeness, in Opencals: automated confirmations and reminders are on every plan with per-service timing, deposits and cancellation windows are configured per service under deposits and cancellations, customers get a self-service portal to reschedule or cancel without contacting you, and checkout questions let you capture a phone number so SMS reminders can actually reach someone.

The pricing model is relevant to one specific decision here: Opencals charges $0.99 per completed booking rather than a monthly subscription, so a no-show doesn't cost you a platform fee on top of the empty slot. That's a small thing, but it's the kind of small thing that adds up in a business with seasonal swings.

The short version

Measure your real rate for 90 days. Turn on reminders — they're the only lever with proper evidence, and expect a modest improvement rather than a transformation. Make cancelling one tap and stop treating a rising cancellation rate as a failure. Add deposits selectively for first-time and high-value bookings, and watch your booking conversion while you do. Build something that refills the gaps you can't prevent.

And be sceptical of anyone quoting you a big round number about what no-shows cost — including this article, where I've tried to show my working. If you want a second opinion on which lever fits your business, get in touch.

---

Sources: Kaiser Permanente Washington pragmatic randomized study, The Permanente Journal (2022); randomized trial of automated calls added to SMS reminders, NEJM Catalyst (2025); text message reminder RCT in a paediatric clinic, Clinical Pediatrics; Health Management Academy, The $150 Billion Cost of Missed Appointments (2020). Deposit effect sizes are vendor-reported and are flagged as such above.