Coffee shop espresso-machine-down SOP - the first fifteen minutes, the menu downgrade ladder, customer scripts, the service call, and the restart

You are an operations coach for a coffee shop. Turn the
espresso-machine-down emergency flow below into an
archivable SOP document, numbered by section, printable
and postable inside the bar. Store-specific details stay
as 【】 placeholders.

Background (put it in the SOP's opening paragraph): the
espresso machine fails during the morning rush (no flow,
abnormal pressure, tripped breaker, dead steam). The
manager must run four tracks - equipment, menu,
customers, online - inside the first fifteen minutes.
The SOP runs in five stages: diagnose, downgrade,
communicate, call, restart.

1. Diagnosis (minutes 0-5)
- Safety first: tripped breaker, burning smell, or water
  near live parts - cut power and tag the machine. No
  troubleshooting at all; every stage below is void,
  execute only the service call and communication
- Diagnose the blast radius with three questions: is it
  one group head or the whole machine? Does the steam
  wand still produce steam? Is the grinder fine? The
  three answers decide which downgrade rung applies in
  stage 2
- The barista on shift performs the diagnosis. The
  manager does not repair; the manager's post is between
  the register and the phone

2. Menu downgrade (minutes 5-15, pick the rung the
   diagnosis points to)
- Whole machine down: pull every espresso-based drink;
  brew bar, cold brew (check what's steeping), and
  non-coffee drinks stay; cover the menu board with a
  temporary menu sign
- Steam system only: milk drinks go iced-only (cold milk,
  no steaming), hot drinks narrow to black coffee, or
  pause milk drinks - whichever fits our menu
- One group alive: limited list - keep only the 2-3
  signature drinks, pull the rest, and tell customers
  the wait time is longer
- Count servable brew-bar and cold-brew cups on the spot
  and estimate how many 【】 hours they cover at current
  traffic; drop another rung if they won't hold

3. Customer communication (one script, no failure
   details)
- In store: sign at the register plus one line at
  ordering - "our espresso machine is down, today it's
  brew bar and cold brew only; no recovery estimate yet,
  but order now and we'll call you"
- Ordered but not served: offer exactly three options -
  swap for an equivalent, wait for the repair, full
  refund. Every staffer uses the same script; nobody
  improvises for their regulars
- Online channels: pause or delist affected items
  immediately - no courier should discover the shortage
  at the counter

4. Service call (say everything in one call)
- Report: machine model 【】, symptom (tripped breaker /
  no flow / no steam / pressure fault), time it started,
  what has already been done
- Log the ticket number and the ETA, sync both to the
  shift handover book

5. Restart and review
- After repair: flush the lines, dump the first 【2】
  test shots, confirm extraction is back in the normal
  range before selling again
- Execute in order: restore menu, remove signs, relist
  online - tick each item, rely on no one's memory
- Before close, the manager writes the incident review:
  timeline, tickets affected, refunds issued, recovery
  time; file it

Constraints:
- No machine-model-specific steps (they differ wildly
  across brands); anything touching the machine reads
  "per our equipment manual, page 【】"
- No repair-time promises to customers - give a time
  only when it can honestly be estimated, otherwise
  "we're on it"
- The SOP is an action document: no paragraphs about
  purpose, meaning, or team spirit

Why it is written this way

"Safety first, then scope" leads the document because every later fork depends on the diagnosis: a live steam wand means iced milk drinks survive; one working group head means a limited list outperforms a full delist through half the rush. Downgrading before diagnosing pulls drinks you could still sell - the largest chunk of a downtime loss is usually self-inflicted.

Customers who already ordered get one fixed script with three options (swap, wait, refund) because that is exactly where the room loses control: when every staffer explains the failure differently, complaints start there. Keeping the script free of failure detail is deliberate - the more mechanical detail staff volunteer, the more customers conclude "this has been broken for ages." The script protects more than the ticket; it protects the shop's credibility.

The restart stage forces dumping the first two shots because lines hold standing water and cleaning residue after a repair, and that does not go into a customer's cup. Writing it into the SOP is what stops a rushed opener from skipping it. Restart steps get ticked rather than remembered because menu, signage, and online relisting are three separate recovery points - miss one and the downtime's tail drags into the afternoon.

Pointing all machine operations at a manual placeholder is the SOP's shelf-life design: hard-code one brand's steps and the SOP becomes a wrong document the day the equipment changes. The SOP owns the flow, the manual owns the machine - each fails on its own schedule.

Variables