Share of voice measures one thing. How much of the category conversation belongs to your brand. The formula is simple. Divide your brand's mentions or impressions by the category total: you, plus every named competitor. Then multiply by 100. A brand with 200 media mentions in a 1,000-mention category holds a 20% share of voice. That's the math. Now there's a newer version of it. How often AI systems cite your brand versus competitors in the same space.
What Share of Voice Measures
Share of voice started as a media-buying metric: what percentage of total advertising spend or impressions in a category belonged to one advertiser. PR and communications teams adopted the same logic for earned and organic visibility. Instead of ad spend, the inputs become media mentions, share of search, social conversation volume, or AI citations.
The metric answers one question: out of everything people (or AI systems) say about your category, how much of it is about you? AMEC's guide to communication measurement traces the concept back to competitive media spend tracking, which is why the formula still works the same way whether you're counting print clippings or chatbot citations.
The Share of Voice Formula
The core formula never changes:
Share of Voice = (Your Brand's Mentions or Impressions ÷ Total Category Mentions or Impressions) × 100
"Total category" means your brand's count plus every named competitor's count for the same time period, topic, and channel. If you're measuring media coverage, the denominator is total media mentions across the competitive set. If you're measuring AI citations, the denominator is total citations across the same set of prompts.
Mentions-Based vs. Impressions-Based SOV
Two versions of this formula exist, and they answer different questions:
- Mentions-based SOV counts the number of times a brand is named, regardless of audience size. It answers: how often does the conversation include us?
- Impressions-based SOV weights each mention by the audience or reach behind it. It answers: how much of the total potential audience saw or heard about us?
A brand can win on mentions-based SOV (lots of small local hits) and lose on impressions-based SOV (a competitor lands one national placement that reaches more people than all those local hits combined). Report both when you can. Reporting only one creates a distorted read on competitive position, which is the same trap that shows up in Earned Media Value: How to Calculate It (and When Not to Trust It) when a single number gets treated as the whole story.
Worked Example: Calculating Share of Voice
The numbers below are illustrative placeholders built to show the math, not real client data.
Say a mid-size consumer brand tracks itself against four named competitors over one quarter. Media monitoring returns the following mention counts:
- Brand: 240 mentions
- Competitor A: 410 mentions
- Competitor B: 300 mentions
- Competitor C: 380 mentions
- Competitor D: 270 mentions
Total category mentions: 240 + 410 + 300 + 380 + 270 = 1,600.
Brand's share of voice: (240 ÷ 1,600) × 100 = 15%.
That 15% is only useful in context. If the brand's target SOV, based on its size relative to the category, is 20%, a 15% actual result signals underperformance worth a strategy conversation. If the brand is the smallest player in the set and 15% still beats its revenue share of the category, that's a sign PR output is punching above the brand's commercial weight. PR Measurement: The Metrics and KPIs That Actually Matter covers how to set that target before you ever run the calculation.
Traditional Media Share of Voice vs. AI Citation Share of Voice
Traditional share of voice counts mentions across news coverage, broadcast, trade press, and social platforms. AI citation share of voice counts something newer: how often a brand's name appears in the answers ChatGPT, Gemini, Perplexity, or Claude give when someone asks a category question.
The formula is identical. Run a fixed set of category-relevant prompts, count how many times each brand gets named across the answers, and divide your brand's count by the total:
AI Citation Share of Voice = (Your Brand's AI Citations ÷ Total Category AI Citations) × 100
Using illustrative numbers again: across 200 tracked prompts relevant to a category, a brand is cited in 38 answers. Competitors combined get cited 222 times across those same answers (multiple brands can appear in one answer, so the denominator isn't capped at 200). Total category citations: 38 + 222 = 260. AI citation SOV: (38 ÷ 260) × 100 = 14.6%.
The practical difference between the two metrics is what drives them. Traditional media SOV moves with pitching, press releases, and relationships with reporters. AI citation SOV moves with what large language models have indexed and trained on: your own site content, third-party coverage, review sites, forums, and structured data. A brand can hold a strong traditional media SOV and a weak AI citation SOV at the same time, because the two systems pull from overlapping but distinct source pools. Media Monitoring Tools vs. AI Citation Tracking breaks down how the tooling for each differs. I also wrote up How I Track My Own Citation Share to show the exact prompts and cadence I use to watch my own AI citation SOV move over time.
Data Sources for Each Type of Share of Voice
| Measurement Type | Common Data Sources |
|---|---|
| Traditional media SOV | Media monitoring platforms, press clipping services, broadcast transcript monitoring, trade publication archives |
| Social/online conversation SOV | Social listening tools, forum and review monitoring, comment and mention tracking dashboards |
| Search visibility SOV (share of search) | Search volume and ranking data, branded search query tracking |
| AI citation SOV | AI citation and brand-tracking platforms, manual prompt audits run on a fixed schedule, structured logging of chatbot answers across major AI assistants |
Why Share of Voice Matters for PR Strategy
Share of voice gives PR programs a competitive number, not just an internal one. A brand can grow mentions every quarter and still lose share of voice if competitors grow faster. Tracking the ratio, not just the raw count, keeps the metric honest.
It also creates a benchmark for setting targets. A challenger brand entering a category dominated by two large competitors should not expect 33% SOV in year one. A realistic target, tied to media budget, headcount, and newsworthiness, gives a PR team something concrete to report against instead of a vague growth mandate.
Common Mistakes When Calculating Share of Voice
- Leaving out a real competitor. Every competitor a customer would consider belongs in the denominator, even the ones a brand would rather not acknowledge.
- Mixing time periods. Comparing this month's mentions against last quarter's competitor totals produces a meaningless ratio.
- Treating every mention as equal. A brief namecheck and a full feature both count as "a mention" in raw counts, which is why impressions-based SOV exists as a check on mentions-based SOV.
- Running AI citation SOV on inconsistent prompts. If the prompt set changes every time you measure, the trend line isn't tracking share of voice; it's tracking prompt selection.
Frequently Asked Questions
What is a good share of voice percentage?
There's no universal good number. A useful target is your brand's fair share: roughly the percentage that matches your market share, ad spend, or category presence. Beating your fair share of voice signals PR output is outperforming your commercial size; falling short signals the opposite.
How is share of voice different from market share?
Market share measures revenue or unit sales relative to the category. Share of voice measures mentions, impressions, or citations relative to the category. The two often move together over time, but a brand can lead on one and lag on the other in any given period.
Can share of voice be negative or over 100%?
No. Share of voice is a percentage of a total, so it runs from 0% to 100% by definition. If a calculation produces a number outside that range, the denominator is missing a competitor or double-counting mentions.
How often should I calculate share of voice?
Monthly or quarterly works for most brands tracking traditional media SOV, since news cycles need time to accumulate enough data to be meaningful. AI citation SOV can be checked more frequently, since prompt-based measurement can be run on demand.
What is AI citation share of voice?
AI citation share of voice is the percentage of times a brand gets named in AI-generated answers to category-relevant questions, compared to the total citations all competing brands receive across the same set of questions. It uses the same formula as traditional share of voice, applied to chatbot answers instead of media coverage.