Loading…
Pillar — full guide
A framework for GEO competitor analysis using four independent scores — SOV, Position, Sentiment, and Diversity — to pinpoint your weaknesses and decode competitor moves from observation data.
"Raise our AI search visibility" sounds like one goal, but it actually splits into several different issues.
These are separate problems with separate remedies. Judging from a single metric (e.g., raw mention count) means missing critical weaknesses.
GEO Meter expresses each as an independent score on a 0-100 scale.
| Axis | Short name | What it measures |
|---|---|---|
| Share of Voice | SOV | Industry share (relative mention count) |
| Position | Position | Quality of ranking when cited (average position) |
| Sentiment | Sentiment | Favorability of the citation context |
| Diversity | Diversity | Diversity of citing source domains |
Organizing the four axes: SOV and Position represent "exposure volume," while Sentiment and Diversity represent "exposure quality." Each is independent and the fixes differ.
What it measures: across Topics in the same industry, mentions of your company ÷ total mentions across all companies. Your "share of voice" inside the industry.
Typical causes when the score is low:
llms.txt)Direction of the fix:
What it measures: when AI offers multiple candidates, where does your company appear? Rank 1 ≈ 100, rank 10 ≈ 30.
Typical causes when the score is low:
Direction of the fix:
What it measures: whether the citation context is positive, neutral, or negative. Is the company being described as "recommended" or "be careful with"?
Typical causes when the score is low:
Direction of the fix:
What it measures: the number of distinct citing source domains. Are you over-reliant on one source like your own site or only comparison sites?
Typical causes when the score is low:
Direction of the fix:
In real competitor analysis, the combinations of the four axes let you read off strategy.
= the industry leader. Top of every axis.
= volume but not ranking. Exposure opportunities exist, but you are not the front-runner.
= locally strong. Highly rated in a specific domain (e.g., your own site) but underexposed externally.
= a hidden gem. Quality is there, volume is not.
= a negative-exposure problem. Volume exists but the perception is bad.
GEO competitors do not necessarily match SEO competitors.
| SEO competitors | GEO competitors | |
|---|---|---|
| Definition | Compete for the same keyword rankings | Get mentioned by AI on the same Topic |
| How to find | Look at the top of Google search results | Look at companies named together in AI search answers |
| Overlap | — | Some overlap, but there is a set of top-cited companies specific to AI search |
| Likely strategy | Win the keyword rankings | Become the entity AI "tends to cite" |
In GEO Meter's initial observations (the first ranking in the tax accountant industry), generic-sounding firm names (e.g., "○○ Startup Financing Center" — names that directly include area + domain) were observed ranking near the top of the GEO ranking. This suggests that firms which are not strong from an SEO perspective can still land at the top of AI citations. We will continue verifying the cross-industry trend as observation data accumulates.
In GEO Meter, the industry-specific rankings automatically visualize the competitor list.
In practice, monthly operations consist of three cycles: start of month, mid-month, and end of month. A rough time allocation looks like this:
Total: about 75 minutes per month. Consistency matters.
If your SOV is below the industry average, prioritize SOV (it is improvable with foundational tactics). If your SOV is around the industry average, Position or Diversity tends to be the next move.
GEO Meter's Service Report displays the citations underlying the Sentiment score. When Sentiment is low, read through those citations and identify what is being quoted in a negative context.
Covering every competitor is impossible. Focus on the top 3-5; for the rest, a once-a-quarter check is enough.
Set the industry median as the floor and the top 25% of the industry as the 12-month target. For concrete numbers, see the industry benchmarks (refer to the complete guide to GEO management).
By viewing GEO competitor analysis through the four-axis score, "what is the problem" and "what should we change" become clear.
What matters is tracking per-axis change monthly. Rather than concluding from a single snapshot, watching change over 3-6 months lets you decompose the effect of your moves and the actions of competitors.
→ Complete guide to GEO management → Complete guide to GEO implementation → Industry-specific rankings → Measure your 4-axis scores with the free diagnostic
Once you have read the article, check your own status. GEO Meter observes how Claude / ChatGPT / Gemini cite you in one pass.