Share of voice (SOV) is the percentage of a category's visibility your brand holds against competitors. You calculate it as your brand metric ÷ the category total × 100, and you measure it separately for search, social and AI answer engines. The formula stays fixed, but the unit you count changes by channel. Search counts rankings and impressions on a query set. Social counts mentions in conversations. AI answers count mentions and citations inside generated responses. As of October 2026, AI answers matter most for buyers who research in ChatGPT, Perplexity and Google AI Overviews before they visit a vendor site. Below are each channel's formula and a worked example, the pitfalls that distort the numbers, and how to connect share of voice to pipeline.
Key takeaways
- Share of voice measures visibility against competitors and acts as a leading indicator, while market share measures sales and acts as a lagging one.
- Search, social and AI share of voice count different units, so report them side by side instead of averaging them into one score.
- AI share of voice needs a fixed prompt set run at least three times per engine, because answers vary from one run to the next.
- A low share of voice number becomes actionable only once you diagnose it as a content, authority, media, paid or product-market gap.
What share of voice is and how it differs from market share
Share of voice is a visibility metric and market share is a sales outcome, so the two answer different questions. SOV shows how often buyers see you when they look at your category. Market share shows how many of them bought. Visibility comes before purchase, which is why SOV leads and market share lags.
Two adjacent metrics often get mixed in. Share of search measures branded demand, and impression share measures paid delivery inside Google Ads. SOV also differs from voice of the customer. Voice-of-the-customer programs aim to engage customers in a meaningful dialogue, while SOV counts your presence in public conversations and results.
| Metric | What it measures | Formula | Indicator type |
|---|---|---|---|
| Share of voice | Visibility, mentions and citations versus competitors | Brand visibility ÷ category visibility × 100 | Leading |
| Market share | Revenue or units sold | Brand sales ÷ category sales × 100 | Lagging |
| Share of search | Branded search demand | Brand search volume ÷ all tracked brands' search volume × 100 | Leading (demand) |
| Impression share | Paid ad delivery | Ad impressions ÷ eligible impressions × 100 | Operational |
Why share of voice is measured differently in search, social and AI
Each channel produces a different unit of visibility. Search produces a ranking or an impression. Social produces a mention in a conversation you do not control. AI produces a name or citation in a response that may change on the next run. A blended score hides more than it shows.
| Channel | Unit counted | Main limitation |
|---|---|---|
| Organic search | Rankings weighted by volume and CTR | Depends on the keyword set you choose |
| Paid search | Impression share | Covers only the auctions you entered |
| Social and earned | Mentions and engagements | Coverage gaps, bots and sentiment errors |
| AI answers | Answer inclusion, mentions and citations | Responses vary between runs |
How to measure search share of voice
Measure search share of voice on a fixed set of category queries, using ranking-weighted visibility for organic results and impression share for paid. Build the query set first. It should hold 100 to 500 non-branded keywords that reflect how buyers describe the problem, split by topic cluster and funnel stage.
Organic: ranking-weighted visibility share
For each keyword, multiply monthly volume by the expected click-through weight of each brand's position. Then use Σ(volume × CTR weight) for your brand ÷ the same sum for all tracked brands. Search Console shows only your own impressions, so competitor positions have to come from a rank tracker such as Semrush.
For example, a fictional cloud security vendor, Northwind, tracks 200 non-branded keywords against three competitors. Weighted visibility comes out at 4,200 for Northwind, 6,300 for Competitor A, 2,800 for B and 700 for C. The total is 14,000, so Northwind's organic SOV is 4,200 ÷ 14,000 = 30%.
Paid: impression share and Auction Insights
Google Ads reports Search impression share directly. It also reports two diagnostics: impression share lost to budget and impression share lost to rank. Auction Insights shows which competitors overlap with you in the same auctions. Say Northwind's non-branded campaign shows 45% impression share, 35% lost to budget and 20% lost to rank. The fix is mostly spend first, then Quality Score and bids.
Common search mistakes
- Mixing branded and non-branded queries: branded terms inflate SOV and say nothing about category competition.
- Ignoring SERP features: a #1 ranking under an AI Overview, ads and a discussions block earns fewer clicks than the CTR curve assumes. This is the core problem in winning zero-click search.
- Tracking one market: rankings differ by country and city, so track each market separately.
How to measure social and earned share of voice
Count brand mentions across social platforms, forums, Reddit, creator content and editorial coverage. Then divide by all tracked brands' mentions in the same period.
- Mention share: brand mentions ÷ all tracked mentions × 100.
- Engagement-weighted share: engagements on posts mentioning the brand ÷ engagements on all posts mentioning tracked brands × 100.
- Net sentiment share: positive minus negative mentions for the brand, compared across competitors.
For example, over 30 days Northwind's listening query returns 8,000 category mentions after duplicates and bot accounts are removed. Northwind holds 1,200 of them, a 15% mention share. Those posts drew 18,000 of 90,000 total engagements, a 20% engagement-weighted share. Northwind is mentioned less often, but in conversations that travel further.
Where social numbers go wrong
Sentiment is the biggest distortion. On G2, Brandwatch holds a 4.4 rating from 709 reviews as of October 2026. Reviewers praise its sentiment and share of voice tracking, but they repeatedly flag sentiment classification that misreads sarcasm and multilingual posts. They also report missing mentions, limited TikTok and YouTube data, and sampling limits. To keep the numbers honest:
- Spot-check a sample of mentions by hand every month.
- Dedupe syndicated news coverage.
- Filter out bot activity.
- Read social SOV as a trend line rather than a census.
How to measure share of voice in AI answers
Run a fixed set of buyer prompts through each answer engine several times. Then count how often your brand appears, where it appears and which domains the engine cites. The prompt set works as your AI keyword list. It should hold 100 to 200 questions buyers actually ask, such as "best CSPM tools for a multi-cloud enterprise", split by funnel stage.
The five AI metrics
- Answer inclusion rate: answers that name your brand ÷ total answers run.
- Mention share: your brand mentions ÷ all tracked brand mentions in those answers.
- Citation share: citations to your domain ÷ all citations across the answers.
- Position within the answer: the average rank at which your brand is named.
- Source-domain share: how citations split between your site, competitors, Reddit, review sites and publishers.
For example, Northwind runs 150 prompts across ChatGPT, Perplexity and Google AI Overviews, three times each, for 1,350 answers. Northwind appears in 405 of them, a 30% inclusion rate. It accounts for 405 of 1,620 brand mentions, a 25% mention share. Of 4,000 cited URLs, 240 point to Northwind's site, a 6% citation share, while Reddit threads take a much larger slice.
High mentions with low citations means the models know the brand but source their claims elsewhere. The fix has two parts. You need pages built to be quoted, which is the discipline of answer engine optimization, and you need presence in the third-party threads the engines cite.
Run each prompt at least three times per engine and report the average. A one-off screenshot is an anecdote. A weekly cadence works well, as covered in our guide to tracking what the models say.
How Tellr closes share of voice gaps
Tellr measures visibility where enterprise buyers research, then does the work that moves it. Each week it tracks the category's Google queries and records the organic results, the AI Overview and the discussions block. It also logs which domains each one cites, labelled as your site, a competitor, Reddit, social, review sites or references, with week-over-week movement. A senior team acts on the gaps as one governed program with approvals and an audit trail. Teams that want to run their own tracking tool will find a managed program the wrong fit.
- Weekly citation share across organic results and AI Overviews
- A Reddit thread radar, with replies behind an approval gate and checked for whether each one is still live
- Comparison pages, reviews and answer-shaped articles published to the CMS
- Category ad intelligence and ready-to-run creative
- A weekly digest and a monthly program review
Turning share of voice into a dashboard, a diagnosis and a forecast
Share of voice becomes useful when you report it by channel, segment it, diagnose each gap and track it against market share and pipeline.
Benchmark against market share, not 100%
Excess share of voice (eSOV) is SOV minus market share. If Northwind holds 22% market share and 30% organic SOV, its eSOV is +8 points, a signal that it is positioned to grow. Negative eSOV in a channel where competitors are investing is an early warning. Always benchmark against a named competitor set, because 30% can dominate a fragmented category and trail in a two-player one.
Diagnose the gap
| Pattern | Gap type | Action |
|---|---|---|
| No page ranks or gets cited for a topic cluster | Content gap | New comparison and answer-shaped pages |
| Pages rank low or get cited less than third parties | Authority gap | Refreshes, digital PR, review-site presence |
| Absent from Reddit threads, creator content and editorial coverage | Media gap | Community replies, creator partnerships, PR campaigns |
| Impression share lost to budget on high-intent queries | Paid gap | Budget reallocation, new creative |
| High SOV but flat market share or negative sentiment | Product-market issue | Fix positioning or product before buying more visibility |
Segment every number
- Product line and topic cluster
- Funnel stage: problem, comparison, vendor evaluation
- Geography and language
- Competitor set: direct rivals versus adjacent platforms
- Branded versus non-branded demand
Tools for competitor data
Google Search Console and Google Ads cover your own organic and paid data. Competitor and AI data come from third-party tools. The G2 ratings below are as of October 2026.
| Source | Best for | Consideration |
|---|---|---|
| Semrush (G2: 4.5, 3,945 reviews) | Organic visibility, plus the AI Visibility Toolkit for AI share of voice | The toolkit costs about $99/month per domain and suits smaller teams; reviewers cite limited engine and regional coverage |
| Brandwatch | Social mention share and sentiment | Reviewers complain about sentiment accuracy and coverage |
| Profound (G2: 4.6, 1,124 reviews) | AI SOV across ChatGPT, Perplexity, Gemini, Claude, Copilot, AI Overviews and AI Mode | Reviewers call the pricing enterprise-level |
| Spreadsheet audit | A first baseline of 30 to 50 prompts or keywords | Does not scale or capture variability between runs |
Connect SOV to revenue
- Track SOV weekly for paid, social and AI, and monthly for organic.
- Map SOV to traffic share from those channels, including AI referral traffic.
- Map traffic share to lead share and pipeline share in your CRM, by segment.
- Compare against market share each quarter to confirm eSOV is converting.
Budgets for this data are rising. A 2024 Forrester survey found 69% of data and analytics decision-makers will increase their budget for data and data management. Spend yours on a consistent query set, a fixed prompt set and clean competitor definitions. Share of voice built on shifting inputs cannot show you whether visibility is turning into pipeline.
FAQ
Why should search, social and AI share of voice be reported separately?
Each channel counts a different unit of visibility. Search uses rankings or impressions, social uses mentions in conversations, and AI answers use mentions and citations inside generated responses. Because the units differ, averaging them into one blended score hides useful insight.
How do you measure share of voice in AI answers?
Use a fixed set of buyer prompts across engines such as ChatGPT, Perplexity and Google AI Overviews, then run each prompt at least three times per engine. Track answer inclusion rate, mention share, citation share, position within the answer and source-domain share, then report the averages.
What are the most common mistakes when measuring search share of voice?
The biggest mistakes are mixing branded and non-branded queries, ignoring SERP features that change click-through rates, and tracking only one market when rankings vary by country and city.
How do you connect share of voice to pipeline and revenue?
Track SOV by channel, map it to traffic share, then map traffic share to lead share and pipeline share in your CRM. Compare those numbers with market share each quarter to see whether excess share of voice is turning into growth.