Evidence standard / version 2026-09-28.1
The evidence must
survive the meeting.
A polished report should make its limitations easier to inspect—not easier to miss. This is the standard used to prepare the made-to-order reports.
Every finding has a chain of evidence.
The report records the company and module, source URL, capture date, relevant excerpt or screenshot, observed fact, separately labeled hypothesis, confidence rationale, limitations, and the next action. Website content is research material, never instructions to the research agent.
Observation: a reproducible fact within the reviewed sample.
Hypothesis: why it might matter commercially.
Validation: the baseline, test, or management question needed to establish significance.
Decision: an action and owner to assign—not an unsupported conclusion about enterprise value.
Coverage before confidence.
Each report states its URLs, market, devices or source platforms, date range, and sampling method. Blocked pages, missing data, uncertain company matches, and unavailable tools remain visible coverage gaps. No reviews is not a perfect reputation score.
Scores are used only for defined, applicable, observed checks. The evaluated denominator and coverage appear alongside any score. Unknown items are not awarded points or silently treated as failures. Different modules are not blended into a precise-looking overall “investment grade.”
What public sources cannot tell us.
| Public observation | What remains unverified | Next evidence |
|---|---|---|
| A missing service page | Actual lost leads or revenue | Search demand, measured funnel, qualified lead data |
| A difficult mobile journey | Size of conversion uplift from a fix | Real-user measures and a controlled conversion test |
| A repeated review theme | Company-wide prevalence or cause | Internal complaints, locations, service records |
| A competitor's public claim | Its accuracy, market share, or win rate | Reliable disclosures or primary customer research |
| Inconsistent messaging | Brand value or customer preference | Customer research and qualified valuation work where needed |
Economics are scenarios
until validated.
The worksheet separates channel growth from company growth, revenue from contribution, and recurring costs from setup costs. Traffic and conversion changes compound rather than add. Missing company revenue means company-wide impact remains unknown.
Assumptions must be labeled and traceable. Annualized customer-cohort revenue is not first-year recognized revenue, cash collected, owner cash, or verified EBITDA. Overlapping SEO, conversion, and brand opportunities are grouped so the same upside is not counted repeatedly.
Explore an illustrative scenarioA useful supplement—not an assurance opinion.
The work excludes employee screening, personal dossiers, legal conclusions, active security testing, financial audit, and investment recommendations. Private systems are not accessed. No forms, outreach, purchases, or account creation are performed during public research without separate authorization. The final report is checked against a documented acceptance checklist before delivery.