Write a Short Survey Customers Finish — and You Act On

You are a survey designer for an Australian small business. Write a short customer survey that people finish AND that changes a decision — every question must earn its place by feeding a named decision.

<context>
DECISIONS THIS MUST INFORM: [BE SPECIFIC — e.g. do we open Sundays? drop the kids' menu? raise the service fee?]
AUDIENCE + RELATIONSHIP: [e.g. email list of past customers, in-store regulars]
CHANNEL: [e.g. emailed link, QR on the counter, SMS]
INCENTIVE: [IF ANY]
PAST ATTEMPTS: [RESPONSE RATES OR WHY PREVIOUS SURVEYS FLOPPED]
TOOL: [e.g. Google Forms, Tally — or undecided]
</context>

Before writing questions, list the named decisions and strike anything from your drafting that doesn't feed one — 'nice to know' is the enemy of finished surveys.

<task>
1. Maximum 7 questions, completable in under 3 minutes. For each: the question wording, the answer format (choice/scale/short text), which DECISION it feeds, and a one-line wording rationale.
2. Wording rules, applied visibly: one thing per question (no double-barrels), neutral framing (no 'how much do you love'), answer options exhaustive and non-overlapping, scales labelled at both ends.
3. At most ONE open-text question, placed near the end, worded to invite critique safely ('what's one thing we should change?').
4. Only include a recommend-likelihood question if a decision actually hangs on it — otherwise omit and say why.
5. Order for momentum: easiest first, sensitive (price, complaints) later, demographics last and only if a decision needs them.
6. Write the invitation message for my channel: honest length promise, why their answer matters, incentive terms if any, and an anonymity statement that is TRUE for the stated tool setup (if responses attach to emails, don't claim anonymity — say 'linked to your email, seen only by us').
7. Pre-write the analysis plan: for each decision, which answer pattern triggers which action (e.g. 'if >60% of regulars pick Sunday morning, trial 4 Sundays'). Decisions without a trigger pattern get flagged — the survey can't settle them and shouldn't pretend to.
8. Pilot step: test on 3 people, watch where they hesitate, fix, then send.
</task>

<output_format>
Survey (question | format | decision fed | rationale) → invite copy → analysis plan table → pilot note.
</output_format>

Rules: no invented benchmarks; small-sample honesty — note the response count below which results are indicative only; en-AU spelling.

Copy the block above straight into Claude — anything in [BRACKETS] is yours to fill in.