Clearline · evidence-based audits
A report should show its work.
Our audit process is built to make the boundary of each conclusion visible. We preserve source evidence, distinguish observations from recommendations, and leave missing coverage plainly labeled.
Evidence policy
Every finding has a status
Verified
The observation is supported by captured evidence and is suitable for a customer-facing recommendation.
Needs review
The signal may matter, but it needs human or source-level confirmation before it becomes a recommendation.
Not tested
The required source, access, environment, or safe test was unavailable. This is not evidence of a failure.
Unsupported
The system cannot substantiate the claim. It is not presented as a finding or recommendation.
Workflow
Observe, verify, prioritize
- Establish scope. Normalize the domain, document crawl limits, and identify the data connections that are actually available.
- Collect evidence. Check the technical, on-page, AI-access, and public local (on-site NAP + Maps listing) lanes the purchased tier covers today, without treating a tool warning as a conclusion. Field performance (Core Web Vitals) and owner-linked GBP Insights stay disclosed as not tested until those data sources are connected.
- Reject weak claims. High-impact signals are verified, conflicts are flagged, and unsupported claims are removed. Crawler folklore never reaches the customer.
- Make the work actionable. The delivered plan ranks a limited number of dual-voice cards — plain terms for the owner, technical detail for the implementer — by impact, confidence, and effort.
Coverage
The report is specific about its limits
Today we note the URLs and crawl coverage actually audited. Public local listing checks run without your login. Owner Insights (GSC, GA4, linked GBP metrics), field Core Web Vitals, prompt sampling, and provider-backed competitive intelligence stay not tested until those connections are available. Missing access is disclosed; it is never quietly treated as observed evidence.
Review
Where a person stays in the loop
Growth audits receive human QA. Dangerous findings, uncertainty, and any proposed publication action require appropriate review. The system creates a structured report; it does not invent certainty or publish changes unattended.