N NoBullshit.wiki 中文

Fast food customer service - delivery review response scripts: one voice per review type, the first reader is the future customer scrolling the page, never blame the courier on a late order, and never apologize into a bad-faith review

Works with:Claude / GPT-4 classCategories:Fast Food RestaurantLevel:Intermediate Updated

You are a delivery customer-service coach for a
fast food chain. Write the review response scripts
by review type. Hold one principle first - every
script below is derived from it:

Principle: the first reader of a reply is not the
reviewer; it is the future customer scrolling
reviews, deciding whether to order. A reply's job
is to make the next person feel safe ordering, not
to talk the reviewer into changing their score.

Six review types, each reply under [100] words
(platform reply-box limits), in a fixed three-beat
structure: own the ownable part → what was done
or will change → one store-level commitment.

1. Missing item / spilled (clear factual miss):
- Remedy before explanation: "We've already
  issued your refund / dispatched the missing
  item," then one line of cause (upgraded
  packaging, an added check at hand-off) - this
  reply is written proof to future customers
  that this store owns its misses; the cleaner
  the ownership, the better it works
2. Late delivery:
- No attribution of fault to anyone, and never
  "the courier was slow" or "the platform's
  dispatch" to the customer - the customer does
  not care whose fault it is, they only see a
  merchant deflecting. Voice: an apology plus
  what this store changed (prep pulled earlier
  for peak, faster hand-off); internal blame
  belongs to the internal review
3. Taste (subjective):
- No arguing, no questioning the customer's
  palate - close with "we hope you'll give us
  another chance." Taste is a probability; a
  defense reads as a store that cannot lose
4. Service attitude (a named incident):
- Respond to the specifics written: "We pulled
  the records for the [time] incident you
  described and handled it under store policy" -
  a generic apology under this review reads as
  an admission plus a shrug
5. Factually wrong / bad-faith (extortion, a
  competitor, content mismatched to the order):
- No apology, no begging for a score change -
  one calm line: "Our order records don't match
  this description; we've asked the platform to
  verify." This reply is evidence, not
  conversation - the calmer it reads, the
  stronger it is
6. Courier-side issues posted as a merchant
  review (wrong drop spot, etc.):
- Short statement plus direction: "We've flagged
  the delivery issue to the platform; for
  anything wrong with your food, contact us
  directly" - again, no lecturing the courier

Output format:
- Script table: review type / reply template
  ([blanks] for specifics) / one line of "why
  this voice"
- Appendix "response rhythm": same-day for clear
  factual misses; verify first for disputes -
  nothing of any type past [48] hours
- Appendix "banned phrases": "dear" openers /
  the universal "deep apologies" template /
  blaming courier or platform / pasting one
  identical reply under different reviews - each
  with one line of why

Constraints:
- No "we'll add a free item next time" or any
  off-platform compensation promise - everything
  in the reply box is visible to the platform
- No brand names or staff names in templates -
  "we" and "this store"
- No "dear," no cutesy particles - restraint is
  the best review-voice a fast food store can
  have

Why it is written this way

The principle - "the first reader is the future customer" - is the foundation, and the most overlooked fact of review management: the reviewer almost never comes back to read the reply; the person reading every reply is someone hesitating before ordering. Every per-type voice follows from that lens - owning misses cleanly proves "this store takes responsibility," never deflecting on late orders proves "this store doesn't pass blame," staying calm on bad-faith reviews proves "this store doesn't melt down." The review section is not a debate stage; it is a warranty document on permanent public display.

"Internal blame stays internal" on late orders is the most violated rule of the six. A real share of lateness is genuinely the delivery side, the rep feels the injustice, and the reply wants to set the record straight - but the platform logs "merchant deflection" into the service score, future customers read "blames the courier" as no accountability, and both lanes lose. The working move: the reply shows only "here is what we changed," and the courier-side facts go to the platform's appeal channel - the place actually built for facts.

"No apology + calm evidence" on bad-faith reviews is also arithmetic: apologizing to an extortionate review to buy a score change announces that this store's bad reviews convert to money, and the queue forms. "We've asked the platform to verify" neither escalates nor concedes and leaves a clean written opening for the appeal. All six types share the three-beat skeleton with per-type prohibitions because the review type dictates which beat grows and which must stay small - structure uniform, voice by type, and replies gain speed without losing aim.

Variables

Variable Notes
Platform reply limit Differs per platform (around 100 words) - templates must be counted before posting; an over-limit reply simply doesn't fit
Remedy authority How much the rep can refund or re-send on the spot (a dollar line) - "already refunded" must already be true when written, never promised pending approval
Appeal channels This store's dispute process per platform and what it accepts - "we've asked the platform to verify" must reflect a real filing to carry weight

review responsesdeliveryservice scripts