Citation rate tells you how often you are cited. Citation Share of Voice tells you how often you are cited relative to everyone else on the same queries. Here is why the second number is the one that matters competitively.
Citation Share of Voice (C-SOV) is a GEO metric that measures your domain’s citations as a percentage of all citations received across the same query set on the same platform. Where citation rate measures your domain in isolation, C-SOV measures your domain relative to the competitive field. A site with a 30% citation rate and a 5% C-SOV is being cited regularly — but its competitors are being cited far more often on the same queries.
What Citation Share of Voice Measures
Citation Share of Voice (C-SOV) measures competitive position in AI search — not absolute citation frequency. The distinction matters because two sites can have identical citation rates on a query set while being in completely different competitive positions.
Consider a 10-query Perplexity run where your domain is cited 4 times. Citation rate: 40%. If the total citations across all domains on those 10 queries is 80, your C-SOV is 5%. That means competing domains collectively received 95% of the citations on the same queries where you appeared. A 40% citation rate is a reasonable result in isolation. A 5% C-SOV tells you there are between 10 and 20 other domains being cited on every query where you appear — and several queries where you do not appear at all.
C-SOV is the metric that reveals whether you are a participant in AI search on your topic or a dominant presence. Those are different positions with different interventions required to improve them.
C-SOV = your domain’s citations ÷ total citations across all domains × 100
Calculated per platform, per query set. Total citations = every domain cited across every response in the run, summed.
How to Calculate C-SOV: Step by Step
C-SOV requires recording competitor citations, not just your own. This is the step that most citation rate tracking skips — and the reason most practitioners do not know their competitive position in AI search.
Step 1: Run your standard query set
Use the same 10 fixed queries you run for citation rate measurement. Same platform, same session. Do not change queries between citation rate and C-SOV measurement — the denominator (total citations) must come from the same run as your numerator (your citations).
Step 2: Record every domain cited per response
For each query, open the source panel on Perplexity and record every domain that appears — not just whether your domain appears. A Perplexity response typically cites 5–10 sources per query. Record all of them. This takes approximately 3× longer than citation rate tracking alone but produces the data that makes the metric meaningful.
Step 3: Count total citations
Sum all domain citations across all 10 queries. If each query returns an average of 7 sources, your total citation pool is approximately 70 across the run. This is the denominator.
Step 4: Calculate your share
Divide your domain’s citations by the total and multiply by 100. If your domain appeared in 7 of those 70 citations, your C-SOV is 10%.
7 your citations ÷ 70 total citations × 100 = 10% C-SOV
This means your domain received 10% of all AI citations on your query set. The remaining 90% went to other domains.
Step 5: Identify the domains taking the remaining share
Sort the competitor citation list by frequency. The domains appearing most often across your query set are your primary C-SOV competitors in AI search — these are not necessarily your traditional SEO competitors. A domain with a DR of 30 that publishes structured, declarative content on your exact topic cluster may be taking 15% C-SOV while a DR 80 competitor takes 8%. C-SOV reveals the actual competitive landscape in AI retrieval, which can differ significantly from the organic search competitive landscape.
Going deeper? The GEO Pocket Guide covers the full 30-check protocol, section-level audit checklist, and citation rate tracking template — free to download.
Citation Rate vs C-SOV: When Each Metric Matters
Citation rate and C-SOV measure different things and answer different questions. Using only citation rate is like measuring your organic traffic without knowing your market share — useful, but incomplete.
| Question | Use citation rate | Use C-SOV |
|---|---|---|
| Is my content being cited at all? | ✓ Primary metric | Not needed |
| Is my citation rate improving month-on-month? | ✓ Track the trend | Not needed |
| Who is being cited instead of me? | Cannot answer this | ✓ Primary metric |
| Am I dominant on my topic or just present? | Cannot distinguish | ✓ Primary metric |
| Is my GEO gap an authority problem or a structure problem? | Cannot diagnose | ✓ Reveals competitor profile |
| Is my content working at the longitudinal tracking level? | ✓ Simpler to track | Use as supplement |
The most important use of C-SOV is diagnosing the type of gap you face. If the domains taking the remaining C-SOV share are DR 80+ authoritative publications (Semrush, Wikipedia, Search Engine Journal), the problem is domain authority — content structure changes alone will not close the gap quickly. If the domains taking share are DR 20–40 niche sites, the problem is content structure and extractability — GEO Stack interventions directly target the gap and can move C-SOV within weeks.
What Your C-SOV Tells You About Your GEO Problem
The competitor profile revealed by C-SOV determines which layer of the GEO Stack to focus on. This is C-SOV’s primary diagnostic value — not the percentage itself, but who is generating the citations you are not.
Low C-SOV against high-authority competitors
If Semrush, Wikipedia, HubSpot, or Search Engine Journal are taking the bulk of citations on your target queries, the gap is primarily a Structural Authority problem (GEO Stack Layer 4). These domains have decades of backlink equity and training data presence. Content structure improvements alone — declarative openings, entity naming, self-contained sections — will move your citation rate but will not displace DR 90+ domains on category-level queries within months. The intervention here is long-cycle: authority accumulation, external references, consistent publishing.
Low C-SOV against comparable or weaker competitors
If domains with similar or lower domain authority are outperforming you on C-SOV, the problem is content structure — specifically Retrieval Probability (Layer 1) and Extractability (Layer 2). A DR 25 site taking 15% C-SOV against your DR 40 site’s 5% C-SOV means the weaker domain’s content is more structurally aligned with what the AI retrieval system selects. The fix is specific and fast: rewrite section openings to be declarative, make sections self-contained, name entities explicitly.
High C-SOV on concept queries, low on category queries
This is the Tier 1 / Tier 2 pattern documented in The GEO Lab’s E014 experiment. Proprietary concept queries (“What is the GEO Stack?”) generate high C-SOV because your terminology is the only source. Category queries (“What is GEO?”) generate low C-SOV because established authorities own the category definition. The intervention for the category gap is a different problem — it requires authority accumulation, not content structure fixes.
What Practitioners Say
The diagnostic use of C-SOV — checking whether your gap is against high-authority or comparable domains — changed how we prioritise GEO work. We were spending time on content structure when the real problem was that Wikipedia and Semrush owned every category query we were targeting. C-SOV made that visible in a way that citation rate alone never could.
Recording every cited domain per query rather than just your own is the step that practitioners skip because it takes longer. C-SOV is what you find when you do it. The competitive landscape in AI retrieval is fundamentally different from the organic search competitive landscape — you cannot assume the same domains are competing with you.
Frequently Asked Questions
What is Citation Share of Voice (C-SOV) in GEO?
Citation Share of Voice (C-SOV) is a GEO metric that measures your domain’s citations as a percentage of all citations received across the same query set. Formula: your domain’s citations ÷ total citations across all domains × 100. C-SOV reveals competitive position in AI search — a domain can have a high absolute citation rate but a low C-SOV if competitors are cited more frequently on the same queries.
How is C-SOV different from citation rate?
Citation rate measures your domain in isolation — what percentage of your test queries return a citation for your URL. C-SOV measures your domain relative to all domains cited on the same queries. Citation rate tells you how you are performing. C-SOV tells you how you are performing relative to the competitive field on the same queries.
How do you calculate Citation Share of Voice?
Run your standard query set. For each query, record every domain cited in the response — not just your own. Sum all domain citations across all queries (total citations). Divide your domain’s citations by the total and multiply by 100. If your domain was cited 7 times and total citations across all domains were 70, your C-SOV is 10%.
What is a good C-SOV benchmark?
There is no universal benchmark — it depends on query competition and citation concentration. On competitive queries where 10+ domains are regularly cited, a 10–15% C-SOV represents strong presence. On niche queries where your domain owns the terminology, above 40% is achievable. The most useful benchmark is your own month-on-month C-SOV trend and your share relative to specific named competitors.
How often should you measure C-SOV?
Measure C-SOV monthly alongside your standard citation rate check. Use the same 10-query set each time to produce a comparable trend line. Monthly tracking reveals whether your competitive position is improving, declining, or stable — and whether interventions are closing the gap against specific competitor domains.
Citation Share of Voice (C-SOV) measures what citation rate cannot: your competitive position in AI search. A domain can have a 40% citation rate and a 5% C-SOV simultaneously — cited regularly, but outcompeted heavily on the same queries. The diagnostic value of C-SOV is in the competitor profile it reveals: high-authority competitors signal an authority accumulation problem, comparable-authority competitors signal a content structure problem. Those require different interventions.
Ready to measure your C-SOV? Start with a citation check using the AI Citation Leaderboard — free first check. Then expand to full C-SOV tracking with the 30-check protocol.
Questions? Contact The GEO Lab.