Breakfast shop equipment technician - the mid-rush equipment failure keep-trading SOP: three-minute self-check, reduced-capacity running, menu switch, logged follow-up - archivable

Please organize our breakfast shop's "equipment failure mid-rush"
response into an archivable SOP. Scope: any equipment failing
suddenly between opening and closing (especially the 6:30-9:00
rush). The single governing principle: keep the doors open first,
protect the menu second, close only as the last resort - organize
the flow below under that principle, completing each step's
judgment criteria and red lines.

Step 1: three-minute self-check (user-side only, no disassembly)
- Steamer: power / water level / inlet valve / door seal seated
- Dough mixer: e-stop reset / belt fallen off
- Fryer: thermostat dial position / breaker tripped
- Soy grinder: beans jammed / reset button
- Criteria: the check list is taped to each machine; three boxes
  ticked with no result means proceed to step 2 - no continued
  troubleshooting

Step 2: triage (match against criteria, assign on the spot)
- Class A "run degraded": function present but slow (steamer slow
  to steam, fryer slow to heat) - action: stretch production
  intervals, protect hero items, drop the most machine-time-hungry
  items
- Class B "function lost but substitutable": machine down, but a
  workaround exists (mixer down - hand-mixing or bought wrappers
  for today only; by tomorrow it must be repaired or replaced)
- Class C "irreplaceable line down": no substitute (grinder down
  with no backup stock - the soy line comes off today)
- Criteria: unsure whether A or B? Treat as B - round up, never
  down

Step 3: menu switch and customer notice
- Cut order: by the item's share of that machine's time, highest
  first; hero items hold out longest
- Notice wording: chalkboard / customer group gets one sentence -
  "X is unavailable today; restoration per in-store notice" - no
  specific restoration time (unless the technician gave a firm
  one), no technical explanations
- POS sync: the item key gets disabled on the spot - never take
  orders you cannot fill

Step 4: repair call and logging
- Call essentials: symptom + when it started + what was ruled out
- Fault card gets three lines on the spot: time of failure /
  symptom / self-check and response - completed after closing

Step 5: review within 24 hours
- Second occurrence of the same fault - goes to a
  replace-versus-repair evaluation, with the prior repair record
  attached
- The review answers exactly two questions: what did we fail to
  sell this time, and could it have been caught earlier

Safety red line (outranks the governing principle):
- Smell of gas, sparks, or smoke: shut gas and power, ventilate,
  evacuate - the keep-trading principle does not apply; treat as
  an immediate closure

Organization requirements and constraints:
- Every step states: who does it, within how many minutes, what
  the output is
- No "contact a qualified professional" filler - state whom to
  contact and what to say on the phone
- The equipment list reflects our shop's actual machines; do not
  extend to equipment we do not own

Why it is written this way

The response logic for a breakfast rush failure runs opposite to the generic playbook. Standard equipment-emergency doctrine says "safety first, shut down and inspect immediately" - a breakfast shop cannot afford that: every minute of rush downtime loses customers for the day and, worse, for tomorrow, because they simply switch to the shop across the street. Hence the ordering keep trading > protect menu > close, with exactly one exception placed above the governing principle: the gas red line. That carve-out must exist and must sit above everything else, because a "keep trading" SOP applied to a gas smell is not scrappy, it is dangerous.

"Round up, never down" in triage follows from asymmetric costs: mistaking B for A costs a few minutes of unnecessary caution; mistaking A for B costs a mid-rush capacity collapse. Likewise, the cut order runs by machine-time share rather than item importance: customers can live without the mixed-grain cake but not without the buns - and the way to save the buns is to give that machine's time to its hero output. The notice ban on specific restoration times is promise management: promise two o'clock, miss two o'clock, and the damage lands on the shop's name; "per in-store notice" pulls the promise back inside what you actually control.

The three-minute self-check ending in "three boxes ticked, move on" is a guardrail for the panicked moment: mid-rush there is a real pull to keep tinkering - which is also a way of postponing the decision - so the check list turns "what have I looked at" into three ticks, and ticking them forces the transition into triage. The fault card takes only three lines on the spot (time, symptom, response), completed after closing; the real mechanism lives in step 5: the second occurrence of the same fault goes straight to replacement evaluation. The true law of small-shop equipment is that the third repair costs half a new machine, but nobody keeps score, so each time the choice is "fix it again" - the second-occurrence rule turns that feeling into a trigger condition, with the prior repair record attached so the evaluation runs on paper, not memory.

Variables