What actually reduces no-shows.

Almost every claim made about appointment no-shows in marketing material is either unsourced, mis-cited, or contradicted by the trial it supposedly rests on. We went looking for the primary sources. Here is what survives.

Start with the rate, because it is usually wrong

Across 105 studies and multiple specialties, the average no-show rate is 23% (Dantas et al., Health Policy, 2018), with Europe at 19.3%. But that review spans decades, and rates have fallen sharply: NHS England hospital outpatient non-attendance was around 5.6% in 2024–25.

For appointment-based consumer businesses — salons, barbers, clinics — there is no peer-reviewed literature at all. Searches of PubMed and Europe PMC return nothing relevant. Every number in circulation is vendor-sourced.

Of those, only two disclose a sample. Fresha, analysing over three million UK appointments between May and August 2024, reported 3.8% overall — 3.16% for hair salons but 21.16% for barbershops. Zenoti reported 1–5% across US and Canadian segments.

Why this matters commercially: a business told its no-show rate should be 30% will happily pay for a tool that gets it to 15%. A business that measures its actual rate at 4% is making a very different decision.

SMS reminders work. The effect is smaller than advertised.

This is the best-evidenced intervention, from a Cochrane systematic review (Gurol-Urganci et al., 2013) covering eight randomised trials and 6,615 participants.

SMS versus no reminder gives a risk ratio of 1.14 (95% CI 1.03–1.26), moving attendance from 67.8% to 78.6%. But heterogeneity was severe (I²=90%), and excluding a single outlier gives RR 1.08 (1.05–1.12) with I²=5%. The lower figure is the honest one.

SMS also performs about as well as a phone call — RR 0.99 — at 55–65% lower cost. That is the real argument for automating reminders: not that texts are magic, but that they are nearly as good as a human calling, for a fraction of the effort.

The review's own authors concluded that the evidence “remains insufficient to conclusively inform policy decisions”, and it is now over a decade old.

Reminder timing does not matter. Reminder count does.

Four independent meta-analyses agree that sending a reminder anywhere between 24 hours and 7 days ahead produces no measurable difference in attendance.

What does help is sending two. In a randomised trial of 54,066 appointments (Steiner et al., 2018), a three-day plus one-day pair achieved 4.4% no-shows, against 5.3% for one-day only and 5.8% for three-day only. Note the ordering: three-day-only performed worse than one-day-only.

So the case for an early reminder is not that it aids memory. It is that it creates time to refill the slot if the answer is no.

Three things that do not do what you were told

Deposits: one trial, null result

In a randomised trial of 6,746 Danish outpatient appointments (Blæhr et al., BMJ Open, 2018), a €34 fine produced 5% no-shows in both arms — 130/3,333 versus 131/3,413. 79% of the fines went unpaid.

Studies reporting large reductions from deposits are observational and typically target people who have already no-showed, where regression to the mean does much of the work. Interestingly, the reverse incentive showed signal: a $15 gift card for attending gave an odds ratio of 1.94.

Two-way confirmation: no attendance benefit

Requiring a reply does not beat a one-way message: RR 1.03 (0.95–1.12) across five trials. A head-to-head trial of 3,695 appointments found 27.9% versus 27.2%, p=0.622, with only 19.5% of patients replying.

Its genuine value is different and still worth having: it converts silent no-shows into cancellations you can refill.

The “9 out of 10 people attend” message failed

The social-norms nudge everyone copies was tested in a randomised trial of 10,111 appointments (Hallsworth et al., PLoS ONE, 2015) and was not significant — 10.0% versus 11.1% control. What worked was stating the specific cost: “costs the NHS £160 approx” gave 8.4%, OR 0.74, p<0.01. A vaguer “wastes NHS money” was significantly worse.

Social norms did significantly increase cancellations, which frees slots by another route. Most write-ups of this trial get it backwards.

The largest lever is your booking window

Lead time is the single most consistent predictor of non-attendance in the literature — significant in 41 of 49 studies in the Dantas review. One ophthalmology study of 51,529 appointments found 9.1% no-shows at 0–2 weeks against 38.3% at six months.

The clearest demonstration is a London general practice that cut its maximum booking lead time from 28 days to one working day and watched non-attendance fall from 7.8% to 3.9% (Margham et al., BJGP, 2021). The authors' summary is worth quoting exactly:

And the risk is wildly concentrated

Blanket interventions are inefficient because no-show risk is not evenly spread. In one large analysis the lowest three risk quartiles missed 0.4% of appointments while the highest missed 24–29%. Finnish register data found 5% of patients accounted for 21.8% of all no-shows.

The operational implication is that reminder intensity should be targeted. Deposits, confirmation requirements and phone calls are worth reserving for the small group that actually drives the number — where the friction is justified — rather than imposing them on the majority who were always going to turn up.

What a no-show costs

Two warnings about using these. The widely-circulated US figure of “$150 billion a year” has no research basis — it traces to an unsourced opinion piece by a scheduling-software executive. And multiplying missed appointments by average appointment cost overstates the loss, because slots get backfilled and staff are paid either way.

What we would actually do

1. Measure your real rate first. It is probably far lower than the number you have been quoted, and it determines whether any of this is worth doing.
2. Shorten the booking window where the business model allows. This has the largest measured effect of anything here.
3. Send two reminders, not one. Timing within the 24-hour-to-7-day range is not worth optimising; the second message is.
4. Make cancelling trivially easy. A cancelled slot can be refilled; a silent no-show cannot.
5. Target friction at the high-risk minority rather than everyone.
6. If you use a cost message, use a specific number. Vague appeals to waste performed worse than saying nothing.

Planaro handles the mechanics — scheduling, Greek SMS reminders, calendar sync, deposits and wait-list — but the evidence above is clear that configuration matters more than the feature list. Our own product page cites a clinic reporting a 62% reduction in no-shows; that is one customer's self-reported result, not a controlled trial, and we would rather you set expectations from the research than from our best case.