Regulars do not vanish suddenly. They come less often, then only to the big nights, then not at all — and it takes months. It is visible in your door data long before you notice it in the room. You only have to count the right thing.
The wrong number: headcount per night
The number everyone looks at tells you the least. A packed Saturday can be entirely first visits; a half-empty Tuesday can be your whole core crowd. Headcount measures the line-up, not loyalty.
The right number: visit frequency
It gets useful when you count, per guest, how often within what period. A workable split needs no data science:
| Group | Definition | What you do with it |
|---|---|---|
| First visit | once, in the last 3 months | invite for a second — that is the threshold |
| Returning | 2–4 visits in 6 months | invite deliberately to matching nights |
| Regular | 5+ visits in 6 months | tell them early, hold spots |
| Drifting | was a regular, absent 8+ weeks | reach out while it still lands |
The threshold that matters
The biggest lever sits at the second visit. Someone who came once decides again from scratch next time; someone who came twice has started a habit. If you only ever target one group deliberately, make it the first-timers of the last few weeks.
The early warning
Do not watch who is gone — watch whose interval is stretching. Someone who moved from every two weeks to every six is still reachable. Someone absent four months usually is not.
What it takes
None of this works without a stable guest identity across events, or you count the same person as three separate first-timers. The prerequisites are in managing guests across events, and the limits on keeping the data in GDPR for guest lists.
EventSync models these groups as dynamic segments — they update on every check-in without anyone maintaining a list.