The Regular Who Stopped Coming And Nobody Noticed

Someone asks: "Have you seen Mr Tanaka lately?"

Nobody has. Not for a while. Nobody can say how long.

You look it up. Last visit was fourteen weeks ago. Before that he came roughly every ten days for two years.

⚠️ He did not quit. He faded, and fading is invisible — because on any given night, one regular not being there looks exactly like a normal night.

The short version

• Regulars leave by stretching the gap, not by stopping
• A missed regular costs far more than a missed walk-in, and shows up far later
• The number to watch is each regular's own usual gap, not a fixed cut-off
• The window where you can still get them back is roughly twice their normal gap
• After about three months, a win-back message reads as a sales pitch

Why nobody notices

A regular's absence is only visible against his own pattern, and nobody holds twenty patterns in their head.

What the floor sees	What is actually happening
"Quiet tonight"	Four regulars all in week three of a stretch
"He's probably busy"	He went somewhere else on a Friday and liked it
"He'll be back"	He came every 10 days; it has been 40

⚠️ The staff who would notice are the ones who left. Institutional memory of regulars walks out with the person who knew them.

One — every regular has his own clock

The single most common mistake is using one rule for everybody.

"Contact anyone who hasn't been in for a month"

That rule is wrong in both directions:

Comes every 7 days  …a month means he is long gone. Too late
Comes every 6 weeks …a month means nothing. You just annoyed him

⇒ The usable rule is relative to his own gap:

His usual gap  = median days between his visits
Watch point    = usual gap × 1.5
Act point      = usual gap × 2

Every 10 days  ⇒ watch at 15, act at 20
Every 6 weeks  ⇒ watch at 9 weeks, act at 12

Use the median, not the average

⚠️ One long gap from a holiday will drag the average and hide the pattern. The median ignores it.

Two — the arithmetic of one lost regular

It is worth knowing the size before deciding how much effort this deserves.

Visits every 10 days      ⇒ about 36 visits a year
Average spend per visit   …25,000
⇒ 900,000 a year

Lose four such regulars in a year …3,600,000

⇒ That is usually larger than the entire advertising budget that was spent trying to replace them.

⚠️ And the replacement costs money while the retention costs a phone call.

Three — the window is shorter than people think

Getting a regular back is not equally possible at every point.

At 1.5× his gap  …he has not decided anything. A message lands as attentive
At 2× his gap    …he may have been somewhere else once. Still recoverable
At 3× his gap    …he has a new habit. You are now competing
Past 3 months    …a message reads as a sales pitch, not a relationship

⇒ The whole value of tracking this is to act inside the first two rows.

What this means in practice

Check weekly, not monthly

⚠️ A monthly check will put most regulars in the third row before you look.

Four — who makes contact matters more than the message

Weak  …a broadcast from the venue
Better…a message from the venue mentioning his usual girl
Best  …a message from the entertainer he asks for, in her own words

⇒ The relationship is usually with a person, not with the room. A message from the room is talking to him about a relationship he does not have.

⚠️ If his entertainer has left, say so before he finds out at the door. That is the single most common reason a long regular disappears without a word, and an honest message beats an awkward arrival.

What to say

Weak  …"We miss you! Come back for a special offer"
Good  …"It's been a while since you were in — I've got Thursday
        and Friday this week if you fancy it."

⇒ No discount. A regular who came 36 times a year at full price does not need a discount; offering one tells him his last two years were overpriced.

Five — the three reasons regulars actually fade

Worth knowing, because two of them you can fix and one you cannot.

One   his entertainer left or changed shifts     …fixable
Two   something went wrong and he did not complain …fixable if you ask
Three his own life changed                        …not fixable

⚠️ Two is the dangerous one, because it is invisible. He will not tell you. He will simply not come back.

The question that surfaces it

"Was everything alright the last time you were in?"

⇒ Asked once, early, by someone he knows. Not in a survey.

Six — the list that runs itself

You want a short list every week, not a database project.

For each regular:
  his median gap
  days since last visit
  ratio = days since ÷ median gap

Sort by ratio, descending. Look at the top ten.

⇒ That single ratio column replaces every argument about who counts as a regular.

Where to draw the line for "regular"

Three or more visits, and a gap you can actually measure

⚠️ Two visits is not a pattern. You cannot compute a gap from one interval with any confidence.

The numbers to keep

One   median gap between visits, per customer
Two   days since last visit
Three the ratio of the two
Four  which entertainer he asks for, and whether she is still on

⚠️ Four is the one that explains most of the list. When an entertainer leaves, her regulars fade over the following two months, and it looks like a general downturn.

A sheet for the office

────────────────────────────────────
  Who is drifting?

  For each regular:
    usual gap (median) …………
    days since last …………
    ratio = since ÷ usual …………

  1.5× → watch
  2×   → make contact this week
  3×   → probably gone

  Contact comes from HIS entertainer, by name.
  No discount.
────────────────────────────────────

Common objections

"Our regulars are loyal, they always come back"

⇒ Then the list will be short and the check costs nothing. If it is long, you needed it.

"We'd look desperate chasing people"

⚠️ At 2× his own gap, a message from someone he knows reads as attentive. At 3 months it reads as desperate. The timing is what separates them.

"We don't record who asks for whom"

⇒ That is the highest-value field on this page. Start recording it at booking this week.

"Some people just move on"

⇒ True, and that is the third reason. The point is to separate them from the two you could have fixed.

"A discount would bring them back faster"

⚠️ It also tells every returning regular that waiting produces a better price. You are training the behaviour you are trying to stop.

This week

One   pull every customer with three or more visits
Two   work out each one's median gap between visits
Three work out days since last visit, and the ratio
Four  take the top ten by ratio
Five  for each, check whether his entertainer is still on — then have her make contact

Doing it by hand, or on a screen

The first pass is doable by hand. The weekly rhythm is what fails.

By hand …export visits, sort per customer, compute gaps. Half a day, once
On screen …median gap, days since, and the ratio held per customer,
           with the requested entertainer and her current status alongside

⇒ tasteck keeps visit history per customer with the requested entertainer, so the drift ratio and the reason for it can sit on the same row rather than in two separate exports.

⚠️ It depends on visits being attached to the right customer record. If the same person exists three times under three phone numbers, every gap will be wrong and the list will be quiet when it should be loud.

Wrapping up

• Regulars leave by stretching the gap, not by stopping
• Use each customer's own median gap, never a fixed cut-off
• Watch at 1.5×, act at 2×, and accept that 3× is usually gone
• Contact should come from the entertainer he asks for, and carry no discount
• When an entertainer leaves, expect her regulars to fade over two months
• The whole thing rests on visits being attached to one clean customer record

Nobody notices a regular leaving, because leaving does not happen on any particular night. It only exists as a gap, and gaps have to be measured to be seen.