The Shift You Filled At Six

18:00. Someone cannot make tonight.

The next hour is a manager sending messages, one at a time, in order of who is most likely to say yes.

It usually works. That is why nobody has ever counted what it costs.

What it actually costs

Three things, and only the first is obvious.

The manager's hour. At the worst time of day — the hour before service, which is the cheapest hour to prepare in and the most expensive to lose.

The same three people, repeatedly. ⚠️ Whoever answers gets asked first next time, so the load concentrates. The person who always says yes is being penalised for it, and they will notice long before anyone thanks them.

The quality of the fill. Asked in order of likelihood to accept, not suitability. The person who ends up covering a busy section may be the person least suited to it.

Why it takes an hour

Because it is a serial search: ask, wait, ask, wait.

The wait is the whole cost. Ten messages sent in sequence with five minutes of hoping between them is an hour. The same ten sent at once is five minutes.

⚠️ The reason people ask one at a time is fear of two people saying yes. That is a real problem and it is a much smaller one than the hour.

The three changes

One — ask everyone at once, first come first served. State it plainly: "first to reply gets it." Removes the serial wait and the double-yes problem in the same sentence.

Two — keep a standing list of who wants extra shifts. Not who is available — who wants them. ⚠️ These are different, and most venues only track the first. People who want extra hours will answer faster and resent it less.

Three — record who covered. The single highest-return field on this list. It turns "who owes who" from an argument into a record, and it makes the concentration visible before the person carrying it leaves.

The pattern worth checking

Which shifts get dropped.

If it is scattered, it is life. If it clusters — a specific night, a specific person, a specific week of the month — it is not really a cover problem.

⚠️ A shift that is dropped repeatedly is a shift somebody does not want, and that is a scheduling conversation rather than a series of emergencies.

What not to do

Do not make dropping harder. Penalties, notice requirements, "you have to find your own cover." These produce people who come in ill, or who stop telling you until later, which is worse.

The goal is to hear about it at 14:00 instead of 18:00, and everything punitive pushes it the other way.

The number that changes the conversation

Covers provided, by person, over three months.

Almost every venue has two or three people carrying most of it, and almost none of them know who. ⚠️ Once the list exists, the fix is usually simple gratitude and a slightly fairer ask order — and that costs nothing.

Filled is not the same as fine

Looks like  a gap that was closed
Actually was an hour of somebody's evening, spent on the phone

⚠️ The cost is not the gap. It is the hour, and whose hour it was — almost always the manager's, taken from whatever they were going to do that night instead.

Which is why it never gets counted

⇒ ⚠️ A filled shift produces no incident and no record. The only trace is that something else did not get done, and nobody writes that down.

---

Where the hour goes

Finding contact details  scattered across three places
Not knowing who is free  so you guess, and guess wrong
Asking one at a time   each no costs a full round trip
Not recording the noes  so next time you ask in the same order

⚠️ All four are structural, not personal. Nobody is being slow; the process has no memory and no parallelism.

---

Three changes, in order of effect

Know who is free before you need to  the order of asking is decided in advance
Ask several at once         the waiting disappears
Record every no           next time starts somewhere new

⚠️ The third is skipped most and returns the most. A refusal is information; discarding it means starting from the same place every time.

What a recorded no actually contains

Who, when asked, for which shift, and whether a reason was given

⇒ ⚠️ The reason matters more than the refusal. "Not Tuesdays" is a permanent fact about the rota; "not this week" is not.

---

Read when the gaps fall, not how many

Same weekday recurring  the rota is wrong
Same person recurring  their circumstances changed and nobody asked
Same hour recurring   the start time is wrong

⚠️ All three show up in when the gaps fall, not in how many there are. A count alone cannot distinguish any of them.

---

The person you must not call first

Call the person who always says yes, first

⚠️ They will stop saying yes, and it will all fail at once. 　 ⇒ The most accommodating person is the one who burns out first, and 　 　 their departure removes your entire contingency plan in one week.

---

Two numbers that end the debate

"We occasionally have gaps"  gets debated
"Seven times last month, and all seven were my hour"  does not

⚠️ Count two things: how often it happened after six, and whose hour paid for it. The second number is the one that gets the problem solved.

---

---

Three to hold

Time from drop-out to filled. Falls immediately when the ask goes parallel.

Covers by person, rolling three months. The concentration.

Drop-outs by shift and by night. Separates emergencies from a shift nobody wants.

Where the record has to sit

If the drop-out is a message and the cover is a reply and the rota is a spreadsheet, none of it is recorded — so the same three people carry it invisibly and the pattern in the dropped shifts is never seen.

tasteck keeps shift submissions, swaps, and cover on the same schedule, so who actually covered is a record rather than something people argue about in March.

The hour is not the manager being slow. It is a serial search that nobody has ever made parallel.

---

Read next

• Night Venue Scheduling: Why the Roster Never Matches the Room
• Why Your Best Staff Leave in Month Four
• The Shift Swap That Never Got Recorded
• Two Venues, One Person
• The Shift You Can Never Fill