Overview
The Full Report is the deep diagnostic layer behind MakeMeRank. In the app, it runs as a multi-step pipeline with visible progress and combines three analysis strands into one overall picture.
The three analysis strands
1. Technical audit
This strand is active only for domain scans. It checks how readable, structured, and semantically understandable your website is for AI systems.
Signals include:
- crawl data such as
title, meta description, H1, H2, excerpts, and canonical
- structured data such as
Schema.org / JSON-LD
llms.txt, robots.txt, and AI bot access- Open Graph data, language, HTTPS, response time, and
meta robots
- GEO and content signals such as answer blocks, Q&A patterns, lists, hidden content, and snippet risks
The result is the clarity score.
2. Visibility
This strand checks whether you are recommended even when your brand name is not explicitly mentioned in the prompt.
It does this by:
- generating neutral search intents
- testing different query types such as navigational, commercial, comparative, and optionally a custom search focus
- checking whether search-oriented models mention you and how prominently you appear
- evaluating positioning signals such as the competitive cohort AI seems to prefer
The result is the visibility score.
3. Recognition and risk
This strand measures how stably different models recognize you and how high the risk is for confusion, ambiguity, or weak classification.
MakeMeRank looks at:
- Knowledge: what models know without web access
- Web Search: what models can find through current sources
- confidence, external mentions, and ambiguities per model
- aggregated recognition rate and external authority
The result is the recognition score and risk score.
The five pipeline steps in the app
While the report is running, MakeMeRank shows these steps (same order as in the UI):
- Context analysis
- Presence analysis (neutral / unseeded queries—the same strand as Visibility in the methodology above)
- Recognition analysis
- Scoring
- Action items
The second step corresponds to the visibility strand (unseeded recommendation path); the UI emphasizes presence in neutral search results. Some steps can be skipped or weighted differently depending on the scan type—for example, person scans do not include the technical audit.
How the scores are weighted
| Dimension | What it is based on |
|---|
| Recognition | Model recognition, confidence, and external authority |
| Visibility | Unseeded search queries and placements |
| Clarity | Technical and crawl signals for domains |
| Risk | Uncertainty, ambiguity, and contradictions |
Weighting for domains
- Recognition: 25%
- Visibility: 35%
- Clarity: 30%
- Risk: 10%
Weighting for persons
- Recognition: 60%
- Visibility: 15%
- Clarity: 0%
- Risk: 25%
For person scans, the topic is critical. If you enter only a name, the visibility analysis becomes less precise and MakeMeRank has to rely more heavily on fallbacks.
How recommendations are created
Recommendations are not invented independently. They are derived directly from the analysis strands. Each recommendation includes:
- a finding source, such as the technical audit or visibility analysis
- a priority
- a rough effort level
- an expected impact
- an E-E-A-T classification
This makes it clear why a given action appears in the report.
LinkedIn and extra context
For person scans, LinkedIn can make the report more precise. MakeMeRank uses that context to improve search focus, classification, and prompt specificity around your real positioning.
The LinkedIn connection is read-only. It is used for context, not for posting.
Why results can vary slightly
LLMs do not respond deterministically. Small differences between runs are normal. What matters more than a few points is the direction:
- are you recognized more consistently?
- do you appear more often in recommendations?
- are your signals becoming clearer and more consistent?
If you want to understand how to turn the report directly into topics and drafts, see Content Generation.