Help us improve RankPilot Local
Allow analytics to measure public website visits and acquisition. We do not send passwords, report contents, emails, or phone numbers to GA4.
Evidence and scoring methodology
RankPilot Local separates user-provided inputs, public website evidence, and verified external sources. Deterministic rules then turn those signals into scores, tasks, confidence labels, and reports without hiding uncertainty.
Information entered by an authorised user, such as categories, services, target locations, profile links, competitors, review figures, and operational notes. These inputs are useful context but are labelled self-reported until independently verified.
Signals detected during a dated safe scan of accessible public pages, including identity, contact details, services, locations, titles, headings, canonical URLs, sitemap availability, structured data, and public freshness.
Information confirmed through an authorised integration, approved profile source, delivery record, or other reliable system. RankPilot does not treat a supplied URL or manual checkbox as complete proof of every underlying field.
Verification lifecycle
Save accurate business information and identify which sources are available.
Review public pages using crawl, security, robots, and page-limit safeguards.
Apply deterministic rules, evidence confidence, warnings, failures, and score caps.
Create one canonical task for each unique issue and preserve the strongest evidence.
Complete the work manually or through an approved workflow.
Run the relevant check again before counting the improvement or closing the issue.
Methodology principles
Audit Score evaluates local SEO and business-readiness signals. AI Visibility Readiness evaluates source clarity, crawlability, structure, trust, citations, and freshness. They are not averaged into one hidden health score.
When review counts, ratings, photos, posts, hours, contact proof, schema, or competitor inputs cannot be verified, the model awards limited or zero credit and may cap the final result.
A schema draft, post draft, report, or recommendation does not improve a score merely because RankPilot generated it. The relevant public change must be implemented and detected where verification is possible.
Scanner, audit, AI, schema, and manual workflows can describe the same problem differently. Canonical issue keys merge equivalent work rather than creating repeated tasks with slightly different wording.
Marking a task complete records that the user believes the work is done. Verification checks whether the relevant public evidence changed. Scores increase only after the scoring source includes the verified change.
RankPilot does not guarantee Google rankings, profile performance, AI citations, leads, revenue, review growth, or time-to-result. Reports explain evidence confidence and important missing sources.
Score caps and confidence
A raw point total can be capped when essential evidence is absent. This prevents a business from receiving a high score based mainly on unchecked inputs or internal activity.
Use a score to locate evidence gaps and prioritise work—not as a guarantee of search performance. Always review the failed checks, caps, confidence, scan date, and recommended actions beside the number.
Use the audit guideRankPilot Local keeps recommendations connected to saved or public evidence, then separates implementation from verification so progress is not counted before it can be confirmed.
Explore implementation support
