The Maintenance You Only Do When It Breaks

The ice machine works until the Saturday it does not.

Almost every venue operates on run-to-failure, which is a legitimate strategy for some equipment and an expensive one for the rest. The problem is that nobody has decided which is which.

---

The short version

Some equipment should run to failure. Most of the expensive failures should not
The cost is never the part — it is the hour it chose to fail in
Six items on a schedule beats twenty on a list nobody follows
The cheapest maintenance is cleaning, and it is the first thing dropped
Record every failure with its date. Without that, nothing below is answerable

⚠️ The last line is the whole thing. Most venues cannot say when the ice machine last failed, so they cannot tell whether it is failing more often.

---

One — what failure actually costs

The part is the small number

The part …a known, usually modest figure
The emergency call-out …two to four times a scheduled visit
The night it ran without …the expensive one, and it never appears anywhere

⚠️ The third line is the one that decides the argument, and it is the only one nobody counts.

What a bad night costs

Work it out once, roughly, for your two or three critical items:

Item	If it fails on a Saturday	Rough cost
Ice machine	Buy bagged ice, hourly runs, service slows	Staff hours + panic purchasing
Card terminal	Cash only, queue, some walk-outs	Direct lost sales
Cellar cooling	Product at risk	Stock write-off
Air conditioning	Room empties by 23:00	The whole night

The last row is the one operators underestimate most. A hot room does not produce complaints — it produces an early, quiet exit.

---

Two — the split

Run to failure is fine for

Cheap to replace
Spare available or easily bought
Failure is an inconvenience, not a stop

Examples: most small equipment, lighting, non-critical fixtures.

Schedule maintenance for

Expensive to replace or repair
Long lead time on parts
Failure stops service or risks stock

⚠️ The lead-time question is the one people forget. A part that takes five days to arrive turns a small failure into a week.

The test

"If this failed at 22:00 on a Saturday, what happens?"

• Nothing much → run to failure
• We work around it → run to failure, keep a spare
• We stop, or we lose stock → schedule it

Most venues end up with four to six items on the scheduled list. More than that and the schedule stops being followed.

---

Three — the cheapest maintenance is cleaning

Filters, coils, drains, seals.

Costs almost nothing
Extends life substantially
Is the first thing dropped when the venue is busy

⚠️ A dirty condenser is the single most common cause of avoidable cooling failure, and it is a fifteen-minute job that nobody owns.

Put it on the closing or weekly list with a named owner — not "when someone has time", which is never.

---

Four — the record

One line per failure

Date / item / what happened / who fixed it / cost / downtime

Six fields, thirty seconds. Kept for a year, it answers questions you cannot otherwise answer:

Is this item failing more often? ⚠️ The strongest replace-versus-repair signal there is, and it needs history rather than opinion. What a second failure tells you — as opposed to a first — is covered separately in The Second Time Something Breaks; this page is about which equipment earns a schedule in the first place.

Are we paying emergency rates repeatedly for the same thing? Usually yes, for one or two items.

Is a supplier's response time actually acceptable? Feels acceptable until it is written down.

The replace decision

Repairs in the last 12 months × average repair cost
vs
Replacement cost ÷ expected remaining years

⚠️ Most venues replace far too late — they repair a failing item five times because each repair is individually small, and never add them up.

---

Five — the supplier side

Response time matters more than hourly rate. A cheaper engineer who arrives in three days is more expensive than a dearer one who arrives the same day, for anything on the scheduled list.

Ask about parts before you need them. ⚠️ Which parts are held locally is the question that determines your actual downtime — and it is answerable in one phone call, months before it matters.

One named contact per critical item. Not a general number found during a crisis at 22:30.

---

Six — the seasonal items

Some equipment fails predictably by season.

Cooling …first hot week
Heating …first cold week
Drainage …heavy rain

⚠️ Servicing these the week before the season, rather than during it, is the single highest-return maintenance decision — and it fails to happen because the season has not started, so it does not feel urgent.

---

Three to hold

Failures per item, rolling twelve months. ⚠️ The replace-versus-repair signal.

Emergency call-outs versus scheduled visits. A ratio that should be falling.

Downtime hours during trading. The number that connects maintenance to revenue, and the one nobody keeps.

---

Where the record has to sit

A maintenance record kept in a folder in the office answers questions only when somebody goes and looks, which happens after a bad failure rather than before one. The useful version is a recurring task that appears on schedule and leaves a trace when it is done — otherwise "we service that quarterly" is a belief rather than a fact.

tasteck keeps recurring tasks on the staff-facing screens with a record of completion, so a quarterly service is something that appears and gets logged rather than something somebody is supposed to remember.

---

Related reading

• Nightclub KPI Guide: 12 Metrics Every Operator Must Track
• When The Supplier Changes Something
• The Second Time Something Breaks
• The Nightly Task Nobody Checks
• For nightclubs

---

Run-to-failure is a decision. In most venues it was never made — it is simply what happens when nobody decides.