Why desk booking fills up with no-shows, and how to stop it

Booking policy6 min readThe product team

No-shows are what kill a desk booking system. Not bugs, not adoption: the slow drift where the plan says a desk is taken, the desk is visibly empty, and people learn that the plan is fiction. Once that happens they stop checking it and start walking the floor again, and the tool becomes a thing IT pays for that nobody opens.

The causes are almost always configuration rather than character. Four settings account for most of it.

1. There is no check-in, so nothing is ever wrong

If a booking is never confirmed against reality, the system has no way to know the difference between a desk in use and a desk that was claimed three weeks ago by someone who is now on holiday. Adding a check-in step gives it that signal, and gives the person a small nudge that the booking is real.

Set a grace window after the start time and release the space automatically when it passes. Fifteen minutes suits desks. Meeting rooms usually want less, because a room sitting empty for half an hour is the most visible failure a booking system can have.

2. The booking horizon is too long

When people can book three months out, they book against uncertainty. Nobody knows where they will be on a Thursday in October, so they reserve the good desk and decide later. Most of those bookings become no-shows, and they lock out the people who did know.

Shortening the horizon to two to four weeks removes most speculative booking without anyone having to be told off. It also makes the plan a better picture of next week, which is the week people are actually planning.

3. There is no cost to holding a booking

If one person can hold five desks across five days and lose nothing by not turning up, some people will. A weekly cap per person is a blunt instrument and it works. It does not need to be tight, it needs to exist, so that holding a booking is a choice with a limit attached rather than a free option.

The same applies to guests. A cap on guests per host per day is the difference between a visitor policy and a visitor situation.

4. Cancelling is harder than not turning up

This one is often overlooked and it is the most fixable. If cancelling means logging in, finding the booking and clicking through two screens, and not turning up means doing nothing, then the system has priced not turning up as the cheaper action.

Cancelling should be possible from wherever the person already is: a link in the reminder email, a tap on the calendar event, one action from the bookings list. When cancelling is easier than ghosting, the no-show rate drops without any policy at all.

Worth watching in the numbers: a no-show rate that is concentrated in a handful of people is a conversation, and a no-show rate spread evenly across everyone is a configuration problem. They look identical in a summary figure, which is why the per-person breakdown is the one to open first.

What good looks like

There is no universal target, because it depends on how much of your booking is speculative by design. What matters more is the direction and the shape: a rate that is falling after you tighten the horizon, and one that is not concentrated in the desks near the window.

The point of measuring it at all is not enforcement. It is that a booking plan people trust is worth more than an accurate one, and the two are the same thing right up until the moment they are not.

More on booking policy