What actually reduces no-shows, measured

We quote a 31% median reduction in no-shows. People reasonably ask which part does the work, so here is the breakdown, with the caveats first.

The caveats

This is observational, not a trial. Accounts choose which features to switch on, and the kind of business that turns on deposits is probably already the kind that was going to tighten up. Treat everything below as directional.

The two that work

The three-hour SMS reminder. The single biggest effect we can see. Not the 48-hour one — that one gets read and forgotten. Three hours out is when someone is deciding whether to leave the house, and it is close enough that a reply of “sorry, can’t make it” still gives you a chance at the slot.

Deposits. Clear effect, but smaller on its own than people expect. A deposit changes who books rather than who shows up: it filters out the speculative booking made at midnight that was never going to happen.

Together they are considerably better than either alone, which is why we quote the combined figure and say so.

The two that mostly do not

The 48-hour email reminder. Almost no measurable effect on attendance. It is not useless — it is where rescheduling happens, and rescheduling is a good outcome — but it does not reduce no-shows on its own.

Confirmation requests (“reply YES to confirm”). We expected this to help. It does not, much, and it generates a meaningful number of confused replies. People who confirm were mostly going to come; people who ignore it are a mix of no-shows and people who do turn up.

The thing nobody counts

Recovering the slot. A no-show you filled is not really a no-show, and the waitlist is doing more work for our accounts than any reminder. When a slot frees up at any distance from the appointment, somebody who wanted it gets ten minutes of exclusivity before it goes public.

If you only do one thing after reading this, it is not the deposit. It is turning on a waitlist and letting it fire automatically.