The BeyondSEO Reputation Score answers a specific question: how much reputation evidence does this observed sample support? Its 0–100 scale is our own transparent measure. It is not a search-engine ranking factor, a probability of being cited, or a replica of another company's authority metric.
- Discover candidate pages and record the query, dates and actual search coverage.
- Fetch each candidate and verify target links or brand mentions in captured content.
- Review topical relevance, publisher relationship and placement context with attributable evidence.
- Group repeated publishers and distinguish owned profiles from independent editorial proof.
- Calculate supported points, reduce them for the weakest measured evidence factor and show the result with its limits.
Unknown source qualities earn no supported points. Search snippets are discovery clues, not verified backlinks. Five search pages cannot represent the whole web. The published model 1.1 formula defines every weight, factor and reporting rule.
| Evidence state | Report behavior |
|---|---|
| No conclusive link checks | Withhold the score; retain any observed partial mentions |
| Partial or weakly reviewed sample | Mark provisional / low confidence; headline and adjusted range cannot reach 50 |
| Adequate defined sample coverage and review | Mark moderate confidence within the sample; still no claim of complete coverage |
| Uneven client/competitor evidence | Show the differences; do not declare a numerical winner |
The ceiling is 49/100. It is a conservative product rule, not a statistically fitted boundary. The adjusted range describes unknown rubric choices within this sample, not all possible unseen links.
Large comparative authority models depend on their provider's crawled link graph, model training, ranking observations and refresh cycle. Published descriptions of relevant signals do not supply a reproducible trained model. Applying a logarithm to a small backlink count would not recreate those systems.
BeyondSEO follows useful evidence principles: verified links, publisher diversity, source review, consistent competitor comparisons and dated repeat measurements. It does not rename its score “Domain Authority,” manufacture a provider's number or claim statistical equivalence. Keep technical readiness, actual visibility and reputation evidence as separate measurements.
In a fictional example, supported sample points are 80, with 10 conclusive checks among 40 candidates, 90% source review and full breadth/discovery coverage. Verification is the weakest factor:
80 × (10 / 40) = 20 / 100
Report 20/100, provisional, low confidence. A higher diagnostic midpoint from source-quality assumptions is not the headline. The low result describes limited supported evidence; it does not prove the unread pages are poor or that the business has little reputation everywhere.
Resolve unverified sources, establish real relationships, publish useful original work and earn relevant independent recognition. Review the same discovery budget and market context on the next run. A score can change because evidence coverage or the scoring model changed; explain that before attributing the movement to business performance.
The same rules apply to every client and competitor. That consistency reduces favoritism, but ranked search sampling and human review still have biases. Business outcomes and verified evidence matter more than optimizing the number.