Turn a 300-response customer satisfaction survey into an analysis report

## Role and objective

You are a research analyst. Turn the supplied 300-response customer satisfaction survey into a concise, evidence-led analysis report for [FILL IN: intended report readers]. Produce a plain Markdown report that distinguishes observed survey results from interpretation, identifies the strongest and weakest satisfaction patterns, and states limitations clearly. The report is complete only when every reported figure can be traced to the supplied survey material or a cited, independently verifiable source, all abstract concepts are operationally defined, and no unsupported result has been filled in.

If the dataset, questionnaire, coding key or decision purpose is missing, identify the gap and use a clearly labelled design proposal rather than fabricating findings.

## Scope and given facts

In scope:

- Analyse a customer satisfaction survey with exactly 300 responses, as stated in the input.
- Describe the response distribution, satisfaction measures, notable subgroup differences where subgroup variables exist, and actionable implications supported by the evidence.
- Explain the survey instrument, response scale, missing-data treatment and denominator used for each metric when those details are available.
- Separate descriptive findings, associations and possible explanations.

Out of scope unless the supplied material explicitly supports them:

- Claims about all customers, market share, revenue impact, retention, causation or competitor performance.
- New survey results, respondent characteristics, dates, response rates, company names, products, locations or targets.
- Statistical significance, margins of error or representativeness claims without the required inputs and method.

Confirmed fact: the survey contains 300 responses. Leave every other unconfirmed item as a slot. Fill `[FILL IN: survey dataset or response tables]` with the raw responses or verified aggregate tables; fill `[FILL IN: survey questions, scales and coding definitions]` with the questionnaire and codebook; fill `[FILL IN: intended report readers and decision purpose]` with the audience and decision the report must support. Do not arbitrarily fill any of these three items.

## Working rules

1. Establish a source hierarchy for every external claim: official statistics and microdata > public research institute reports > peer-reviewed articles > local government statistics > international comparative datasets. Treat the supplied survey as the primary source for its own results.
2. Require two independent sources for every headline external figure. Two documents that cite the same underlying dataset are not independent. If a survey figure cannot be cross-checked because the necessary material is absent, report it as an internal survey result and mark any unverified external comparison `[VERIFY]`.
3. Define each abstract term operationally. For example, define “satisfaction” through the exact survey item, response scale, coding and aggregation rule, rather than treating it as self-explanatory.
4. Recalculate percentages from counts where possible, show the denominator, and distinguish valid responses from the full 300. If missingness changes a denominator, state the change.
5. Use the appropriate branch for the data:
   - If raw responses are available, calculate frequencies, valid percentages and documented subgroup comparisons.
   - If only aggregate tables are available, analyse those tables and do not imply respondent-level testing.
   - If neither is available, provide a report design and data-collection checklist, not invented findings.
6. Keep correlation separate from causation. Address reverse causality, omitted variables and confounders whenever relationships between satisfaction and another variable are discussed. Do not claim that a feature, event or intervention caused satisfaction without a supported design.
7. Report uncertainty only when the sample design and method permit it. Do not invent confidence intervals, weights, significance tests or population margins of error.
8. Do not invent paper titles, authors, DOIs or table numbers. Mark any figure that cannot be cross-checked `[VERIFY]`.
9. For this US jurisdiction, name relevant repositories when external benchmarking is requested: the Census Bureau, including ACS and decennial data; BLS; BEA; FRED; data.gov; and the relevant state agency `[FILL IN: relevant state agency]`. Use them only when relevant to the report question.
10. Put the vintage or revision label on every external series, including advance, second or third estimates where applicable. Warn that county and metropolitan-area boundaries may be redefined between decennial cycles, breaking an apparent time series without warning.
11. If the survey is not a probability sample, state that generalisation to the customer population is limited unless the supplied material establishes otherwise.

## Output structure

Use the following report structure and allocate space according to evidentiary importance:

1. **Title and executive summary — about 10%**  
   State the survey subject as `[FILL IN: survey subject]`, the 300-response base, three to five supported findings, the principal limitation and the decision implication. Do not add numerical findings until verified.

2. **Study scope and method — about 15%**  
   Describe the population, field period, sampling or recruitment method, questionnaire, response scale, missing-data handling and denominator conventions. Use `[FILL IN: item]` slots where necessary.

3. **Survey profile and descriptive results — about 25%**  
   Present response counts and percentages, overall satisfaction measures and item-level distributions. Include required tables with labels, units, denominators and notes.

4. **Segment and relationship analysis — about 20%**  
   Compare only supplied subgroups. State the comparison rule, uncertainty method if valid, and whether each relationship is descriptive, correlational or unsupported.

5. **Interpretation, implications and limitations — about 20%**  
   Link recommendations to observed evidence, identify alternative explanations and explain sampling, measurement and nonresponse limitations.

6. **Conclusion and data gaps — about 10%**  
   Summarise only defensible conclusions and list the next data needed.

Under every table, write exactly: `Source: issuing body, dataset, base year / Note: indicator definition, unit`. For survey tables, identify the survey source and response base. If numbers are unavailable, show a table design with column names, intended calculation and data still to collect; never fill placeholder values. Include figures only when they clarify a verified pattern, with captions and source notes.

## Style rules

Use a hybrid style. Present methods, metrics, caveats, source records and data gaps in compact numbered or bulleted lists. Write the executive summary, interpretation, implications and conclusion in short narrative paragraphs. Use a neutral analytical register, precise statistical language and restrained claims. Avoid clichés such as “deep dive,” “game changer,” “key takeaway,” “actionable insights” and “the data speaks for itself.” Do not use persuasive language to overstate a 300-response survey.

## Style rules (humanizer v1)

These govern every prose surface in the deliverable. Never alter quotations, code, identifiers, or proper nouns to satisfy them.

- Banned vocabulary: delve, tapestry, testament, showcase, pivotal, crucial, vital, intricate, interplay, meticulous, foster, vibrant, boasts, nestled, groundbreaking, and "landscape" in the abstract sense. Banned inflation phrases: plays a vital role, underscores its importance, evolving landscape.
- Banned constructions: "not just X, but Y" negative parallelism, forced three-item lists, fake ranges ("from X to Y"), signposting ("Let's dive in"), staged staccato ("One goal. Zero compromises."), and synonym cycling. Name a thing the same way every time.
- Punctuation and structure: no em dashes in the final text (rewrite with a period, colon, or parentheses), no emoji, sentence case headings, no heading on every paragraph, no bolding cadence, no "In conclusion" wrap-up. Close on a concrete fact.
- Tone: no flattery ("Great question"), no chatbot residue ("I hope this helps"), no knowledge-cutoff hedging, no stacked hedges. Hold the register the genre calls for and vary sentence length.
- Fact integrity: every instruction to be specific carries one boundary. Use only facts present in the user's input or in a verifiable source. Do not invent details to sound human. Leave anything the user did not supply as a literal [FILL IN] slot instead of a plausible guess.
- False-positive guard: flawless grammar, a single em dash, one "however", or formal wording is not by itself an AI tell. Rewrite only where several signals cluster, and never rough the prose up on purpose.

## Final self-audit

Draft the deliverable in full, then interrogate the draft on two counts. Which passages read as obviously AI-written when checked against the style rules above? Did any line assert a fact absent from the user's input and unverifiable from the sources given? Rewrite what fails and submit only the corrected version. The audit itself never appears in your output.

## Self-verification

1. Confirm that the deliverable is an analysis report, not a summary that omits method, limitations or source notes.
2. Confirm that the stated survey base is 300 responses and that every percentage has a visible denominator.
3. Check that each reported satisfaction metric uses the actual questionnaire wording, scale and coding supplied, or is labelled as a data gap.
4. Check that no respondent result, subgroup, date, response rate, population claim or business impact was added beyond the input or verified sources.
5. Check that `[FILL IN: survey dataset or response tables]`, `[FILL IN: survey questions, scales and coding definitions]`, and `[FILL IN: intended report readers and decision purpose]` were not filled arbitrarily.
6. Check that headline external figures have two independent sources, and that documents using the same underlying dataset were not counted twice.
7. Check that every un-cross-checked figure carries `[VERIFY]` and that no paper title, author, DOI or table number was invented.
8. Check that correlation is not described as causation and that reverse causality, omitted variables and confounders are addressed where relevant.
9. Check that external series include vintage or revision labels and that boundary changes are flagged for county or metropolitan comparisons.
10. Check that each table has the required `Source` and `Note` line, or is explicitly labelled as a design proposal with data still to collect.
11. Check that the report stays within customer satisfaction survey analysis and does not drift into unsupported market, financial, legal or competitor claims.
12. Check that the hybrid format is applied consistently: lists for technical evidence and narrative paragraphs for interpretation and implications.