# Customer Evidence-to-Campaign Desk

Local-first worksheet for market researchers and marketing specialists. It does not authorize data collection, tracking, profiling, outreach, publishing, advertising, spending, testimonials, or account changes. Use approved research methods, channels, consent, privacy controls, and review procedures. Never place private participant/customer data, credentials, tokens, browser profiles, payment data, or access-controlled material into an unapproved AI service.

## 1. Decision and research-question card

- Decision this research will inform: ____________________
- What is explicitly out of scope: ____________________
- Audience/population definition: ____________________
- Current hypothesis, labeled as hypothesis: ____________________
- Evidence needed to update the decision: ____________________
- Approved methods/channels: ____________________
- Consent, notice, privacy, and retention owner: ____________________
- Research/marketing reviewer: ____________________
- Publish/outreach/ads/spend authority: ____________________

## 2. Evidence and source ledger

| Evidence ID | Exact source/method | Capture date | Population/context | What it directly supports | Claim class | Limit/bias |
|---|---|---|---|---|---|---|
| E-01 |  |  |  |  | primary / participant statement / platform metric / vendor claim / inference / uncertain |  |
| E-02 |  |  |  |  |  |  |

Popularity, engagement, search rank, creator statements, and competitor copy may signal demand; they do not prove truth, causation, typical results, or customer outcomes.

## 3. Interview, survey, and observation log

Use only approved, consented, minimum-necessary data. Anonymize or code participants where appropriate.

| Participant/code | Method | Exact question/prompt | Verbatim or faithful note | Researcher observation | Consent/status | Do not generalize beyond |
|---|---|---|---|---|---|---|
| P-01 | interview / survey / usability / observation |  |  |  |  |  |

AI may organize supplied notes. It must not fabricate participants, quotes, responses, demographics, sample size, consent, or research sessions.

## 4. Insight synthesis matrix

| Theme | Supporting evidence IDs | Contradicting evidence IDs | Frequency in this sample | Confidence | Alternative explanation | Next evidence needed |
|---|---|---|---:|---|---|---|
|  |  |  |  | low / medium / high |  |  |

Do not convert a small or biased sample into a population percentage, causal claim, “customers want,” or “everyone struggles with” statement.

## 5. Audience and positioning hypothesis

- Audience/job/situation: ____________________
- Painful workflow in their language: ____________________
- Existing alternative/workaround: ____________________
- Proposed useful outcome: ____________________
- Proof currently available: ____________________
- Reason to believe: ____________________
- Claim we are **not** ready to make: ____________________
- Hypothesis test that could disconfirm this position: ____________________

This remains a hypothesis until approved evidence supports a marketing claim.

## 6. Claim, endorsement, and disclosure review

| Proposed claim/content element | Evidence IDs | Objective or subjective? | Typicality/results issue? | Material connection/disclosure? | Qualified review | Decision |
|---|---|---|---|---|---|---|
|  |  |  |  |  |  | approve / revise / reject |

Hard stops:

- fabricated or AI-generated customer reviews, testimonials, case studies, experts, citations, or results;
- unsupported performance, savings, revenue, health, legal, financial, safety, “best,” comparative, or AI-capability claims;
- hidden sponsorship, affiliate, employment, gift, or other material connection;
- paid/native content presented as independent editorial content;
- dark patterns, buried terms, deceptive scarcity, confusing cancellation, or tricking people into sharing data.

## 7. Campaign experiment brief

- One approved hypothesis: ____________________
- One audience/context: ____________________
- One useful asset/offer: ____________________
- Approved claim and disclosure: ____________________
- Channel/account/targeting approved by: ____________________
- Exact CTA/destination: ____________________
- Primary metric and decision rule: ____________________
- Guardrail metrics: ____________________
- Start/stop dates and owner: ____________________
- Privacy/consent/data handling: ____________________
- Budget/spend authorization: ____________________
- Human publish/send/launch approval: ____________________

This worksheet prepares an experiment. It does not publish, send, buy ads, change targeting, or spend money.

## 8. Measurement and learning log

| Experiment ID | Exact version/channel | Exposure/denominator | Primary result | Guardrail result | Data quality limits | Decision | Next test |
|---|---|---:|---:|---|---|---|---|
| X-01 |  |  |  |  |  | continue / revise / stop / inconclusive |  |

Separate observed metrics from causal interpretation. A conversion change is not proof of why it changed without an appropriate design and analysis.

## Search and AI-discovery quality check

- [ ] The page answers a real customer question with original evidence or useful analysis.
- [ ] It is not a thin variation created mainly to manipulate rankings or generative search responses.
- [ ] Headline, metadata, structured data, visible content, and destination agree.
- [ ] Sources, authorship/reviewer, dates, limitations, and material disclosures are visible.
- [ ] No scraped or stitched content was republished without substantial original value and permitted use.
- [ ] Search/indexing status is measured separately from business outcomes.

## Safe prompt

> Act as a customer-evidence research clerk, not a participant, customer, expert, analyst of record, advertiser, publisher, media buyer, or compliance approver. Using only the supplied approved evidence, produce: (1) a source ledger, (2) themes with supporting and contradicting evidence IDs, (3) alternative explanations and limitations, (4) audience/positioning hypotheses labeled as hypotheses, (5) claims that are and are not currently supportable, and (6) a draft experiment brief that stops before publishing, outreach, targeting, ads, or spend. Do not invent people, quotes, sessions, survey responses, sample sizes, metrics, citations, testimonials, results, typicality, consent, or causation. Preserve uncertainty and private-data boundaries.