The Guest Who Stopped Coming and Nobody Noticed

A regular who stops coming does not tell you. There is no cancellation, no complaint, no moment where anything happens.

They came every second Friday for a year and then they did not, and the room was busy that night so nobody noticed, and by the time someone says "we have not seen him in a while," it has been four months and he has a new place.

The core difficulty is that every other problem in the venue generates a record. A guest not arriving generates nothing.

You cannot see an absence unless you were already counting the presence.

Why the window is short

The instinct is that a regular of two years has a lot of goodwill and will come back.

In practice the opposite is closer to true. A guest with a strong habit is easier to lose permanently, because the habit was doing the work. Once it breaks, they have to actively decide to return, and most people do not actively decide anything about where they drink.

The practical window in most venues is between six and ten weeks past their normal interval. Inside it, a reason to come back frequently works. Outside it, the new habit has formed somewhere else and the same message reads as a venue chasing business.

The number is not universal — it scales with the guest's normal frequency — but the shape holds everywhere: the value of noticing decays fast, and most venues notice at three to six months.

The mistake in "haven't seen them in 90 days"

Most venues that try this use a fixed window. Ninety days, sixty days, whatever the system defaults to.

A fixed window is wrong for almost every guest you have.

A guest who came weekly is in trouble at three weeks. A guest who came once a quarter is completely normal at ninety days and you would be contacting them for no reason — which is worse than not contacting them, because it signals you do not know who they are.

The window has to be relative to that guest's own pattern. Roughly: what was their typical gap between visits, and how many multiples of it has it been.

Two or three times their normal gap is where attention is warranted. For the weekly guest that is two to three weeks. For the quarterly guest it is six to nine months. Same rule, completely different clock.

This is the single most common reason churn detection fails in hospitality. The rule is right, the clock is wrong, and the venue concludes the whole approach does not work.

What to do when you notice

The instinct is a discount. It is usually the wrong first move.

A discount to a lapsing regular says the relationship was transactional, and it prices your room downward for someone who was previously paying full. It also trains the behaviour — the guest learns that going quiet produces an offer.

What works better is smaller and more specific:

• Something that changed since they were last in, if it is relevant to what they came for
• An invitation to something with a date on it, rather than an open-ended offer
• Contact from the staff member they actually knew, not from the venue generally

That last one carries most of the weight. A regular's relationship is almost never with the room. It is with a person in the room. A message from that person works when a message from the venue does not.

Which means the record you need is not just when they last came. It is who they came to see.

The record that makes it possible

Four fields, per guest. Nothing more is needed to start.

Last visit date. Obvious, and yet frequently absent because tabs are attached to tables rather than to people.

Visit count and first visit. Together these give the typical gap, which is what makes the window relative rather than fixed.

Attached staff. Who served them most often. Determines who reaches out.

What they came for. One line. A drink, a table, a night, a particular server. Determines whether the reason to return is relevant.

That is a small amount of data and it is the difference between a churn list that works and one that annoys people.

The version that goes wrong

The failure mode worth naming: a venue builds the list, sends a batch message to everyone on it, gets a low response, and concludes guests do not respond to outreach.

What actually happened is that the list mixed the weekly guest at three weeks with the quarterly guest at four months, the message came from the venue rather than from anyone, and it offered a discount to people who were not price-sensitive.

Three errors that each look like the same result. The approach was not tested; a bad implementation of it was.

If you try this, start with the smallest possible group — your ten most frequent guests who have gone quiet, contacted individually by the person who knows them. Ten conversations tell you more than a thousand messages, and they tell you why, which is the part the batch never surfaces.

What the "why" usually turns out to be

Venues that actually ask are frequently surprised.

It is rarely a complaint about the venue. The common answers are that the person they came to see left, their schedule changed, the group they came with dispersed, or something about the room changed in a way that mattered to them and to almost nobody else — the music got louder, their table got reassigned, the night they liked moved.

Most of those are either unfixable or trivially fixable, and knowing which is worth a great deal. A regular lost because their server left is a retention problem. A regular lost because their Thursday moved to Wednesday is a scheduling note.

The one number to watch

If you track a single thing, track the count of guests currently past two times their own typical gap, week over week.

Not who — just how many. It is a leading indicator that moves before revenue does, because the guests in it are still coming, just less. By the time it shows up in revenue, the ones at the front of the list have already gone.

A rising count with flat revenue is the clearest early warning a venue produces. Something changed recently and the effect has not reached the register yet.

Where the record has to live

The reason most venues cannot do any of this is not unwillingness. It is that the tab was attached to a table, the guest's name lives in a booking app, the staff relationship lives in a bartender's memory, and none of them join.

tasteck attaches visits to guest records with the serving staff member on them, so who has gone quiet relative to their own pattern, and who should reach out is a list rather than an exercise in remembering.

The guest already stopped coming. The only question is whether you find out in week six or month six.