Share of voice (SOV) is a brand's percentage of total market visibility against competitors, calculated as brand metric divided by total market metric times 100, for one channel and time period. Compare that percentage with market share to see whether visibility is high enough to hold or grow share, or low enough that share is at risk.
Share of voice at a glance
The formula is consistent across channels, but the metric inside it changes each time:
| Core idea | What it means and why | What to do or check | Evidence and limits |
|---|---|---|---|
| Definition | Share of voice started as a brand's media spending as a percentage of all category media spending in a market, channel, and period. Teams now apply the same ratio to mentions, impressions, estimated traffic, and AI answers. | Name the channel before you quote a percentage. A social SOV and a spend SOV are different measurements. | The original meaning was advertising spend. Wider channel uses are later extensions, so an unlabelled SOV number has no fixed meaning. |
| Formula | SOV = (your brand metric ÷ total market metric) × 100. The units must match on both sides. | Use the same unit, period, geography, and language for every brand in the calculation. | Vendor formulas use this ratio, but vendors choose different inputs. Two tools can produce different SOV values from the same market. |
| Denominator | The denominator is the total for your defined market, not every brand that exists. It may be all category spend, all tracked keyword traffic, or mentions for a named competitor set. | Write the denominator next to the result, for example: tracked mentions for five named competitors, 30 days, United States. | Change the competitor set and the percentage changes without any change in your own activity. |
| Channel types | Paid spend, paid search impressions, earned coverage, social mentions, organic search visibility, share of search, and AI answers each capture a different form of visibility. | Measure the channels where buyers encounter the category. Do not average channel percentages into one total unless you state a weighting rule. | Digital walled gardens make total category spend harder to observe, so a cross-channel total is usually an estimate. |
| SOV vs market share | Market share is your share of category sales. SOV is your share of category visibility or spend. The gap between them is the planning signal. | Plot both series over time. Act on a persistent gap, not a one-week movement. | The association is well documented for paid media, but it does not prove that raising any visibility metric will raise sales. |
| ESOV | Excess share of voice, or ESOV, is SOV minus share of market. Positive ESOV means SOV sits above market share. | Set SOV relative to your current market share, then check price, distribution, product, and creative quality before committing budget. | ESOV research links positive ESOV with growth on average. Effects vary by category, brand size, and creative quality. |
| Sentiment | Raw SOV counts negative, neutral, and positive visibility together unless you separate them. | Report volume SOV, positive SOV, negative SOV, and net sentiment side by side. | A crisis can raise SOV while damaging demand. Sentiment classifiers also differ, so do not compare net scores built with different rules. |
| Benchmarks | No universal good percentage exists. A useful target depends on market share, category size, competitor count, and whether you lead or challenge. | Benchmark against your own market share and against the same competitor set over time. | Published ESOV rules are averages, mainly from paid media studies. No analyst benchmark in the research sets a good SOV for AI, social, or earned visibility. |
| AI answers | AI SOV counts brand mentions or citations across a fixed prompt set, engines, and window. A mention is your brand named in answer text. A citation is your website linked as a source. | Fix the prompts, engines, region, run count, and counting rule before you track change. Report a range when repeated runs vary. | Answers vary between runs of the same prompt. One reading from one engine and one day is a sample, not a stable ranking. |
How share of voice works
Share of voice works by comparing your presence with the total presence available in a defined market. The mechanism is arithmetic, but the meaning comes from the choices behind the arithmetic:
| Variable | Mechanism | Consequence | Implication |
|---|---|---|---|
| Brand metric | You count one thing: spend, impressions, mentions, estimated clicks, searches, or AI mentions. | You choose which kind of visibility to measure. With spend, you compare budgets. With mentions, you compare conversation volume. | Pick the metric that matches the decision. Do not use mention SOV to judge media efficiency. |
| Total market metric | You count the same metric for every competitor in the set. | Your percentage rises when you grow, when competitors shrink, or when the set gets smaller. | Read your absolute count beside the percentage. A rising SOV can hide falling category volume. |
| Market share gap | You compare SOV with share of market over time. | Positive ESOV has been associated with later share growth in paid media studies. This is an established association, not a proven cause for every channel. | Use ESOV for budget direction. Test the result with sales, search demand, and pipeline data before you scale spend. |
| Sentiment and quality | You split visibility by tone, prominence, and source quality. | Equal SOV can carry opposite meanings when one brand's coverage is positive and prominent and another's is negative or buried. This follows from the counting rule rather than from a growth study. | Never set a volume target without a sentiment and prominence check. |
| Time | You repeat the same measurement across periods. | Direction becomes visible, and one-off spikes become easier to spot. | Keep the set and rules stable. If you must change them, restart the baseline and mark the break. |
Calculate share of voice with the core formula
Use this formula for one channel at a time:
SOV (%) = (your brand metric ÷ total market metric for the same channel and period) × 100
State four things with every result: the metric, the competitor set, the geography and language, and the time period. A 30-day window suits social, earned, and AI monitoring when you need regular decisions. A quarter suits spend planning. Share of search work in the IPA research used a 6–12 month rolling period, which smooths weekly noise.
A single-channel example makes the units plain. Your brand earns 2,500 tracked social mentions in 30 days. You and three competitors earn 10,000 tracked mentions in total. Your social SOV is 25%. That figure says nothing about reach, sentiment, ad spend, or sales until you measure those separately.
Define the competitor set and the total market
Start with the brands buyers actually name as alternatives, then add the brands that appear beside you in search results, AI answers, and category coverage. Sales input matters here. A competitor set built only from your keyword tool will miss substitutes that buyers find through other routes.
Choose one denominator rule and keep it:
- Named competitor set: Your brand plus a fixed list of competitors. This suits monthly reporting because the set stays stable and the work stays manageable.
- Discovered market: Every brand a tool finds for your tracked keywords or prompts. This suits market mapping, but the denominator moves as the tool discovers new brands.
- Category total: Total category spend or search demand from an external dataset. This suits budget planning when a credible category total exists.
Deduplicate before you calculate. Group syndicated copies of one article under your stated rule, merge brand name variants and misspellings, and remove spam and duplicate posts. For AI answers, decide whether a brand named three times in one answer counts once or three times, then keep that rule. State the rule in the report, because a deduped count and a raw count give different percentages.
What 100 percent share of voice means
A 100% SOV means your brand is the only brand visible in the measured universe, a statement limited to that universe and silent on how much of the whole market you hold.
Treat 100% as a warning sign first. The usual causes are a prompt set that names only your brand, a keyword set with no competitor rankings, a competitor list with one entry, or a category defined so narrowly that rivals fall outside it. Widen the set, rerun the measurement, and only trust the percentage when competitors had a real chance to appear.
Share of voice changes with the channel you measure
Each channel counts a different event, so channel SOV figures do not substitute for each other:
| Channel | Metric used | Data source | Free and paid tools | Caveat |
|---|---|---|---|---|
| Social | Brand mentions, tags, and hashtag volume for your set | Social listening platforms | Mention and Brand24 offer free SOV calculators. Brandwatch, Sprout Social, Talkwalker, and Meltwater sell listening products. | Mention counts reward volume, including negative posts. Check sentiment, unique authors, reach, and duplicates before you act. |
| Earned media | Articles or mentions for your brand divided by articles or mentions for the set | Media monitoring databases | Muck Rack, Cision, and Meltwater sell coverage reporting with SOV widgets or benchmarks. | Syndication can inflate counts. Decide whether each syndicated copy counts, and separate prominence from a passing name-check. |
| PPC | Search Impression Share, or impressions divided by eligible impressions | Google Ads and Auction Insights | Google Ads reports Impression Share, Search top Impression Share, Search absolute top Impression Share, overlap rate, and outranking share at no extra tool cost. | Impression Share measures auctions you were eligible for. It does not measure total category demand, organic results, or auctions outside your targeting and quality eligibility. |
| SEO | Estimated organic traffic share for tracked keywords, built from ranking position, search volume, and expected click-through rate | Rank tracking and clickstream estimates | Semrush, Ahrefs, SE Ranking, Moz, and Conductor sell rank tracking with SOV or visibility reports. Trials and limited free plans change, so check the current plan before you rely on one. | Estimated traffic is a model. Keyword set, click-through curve, search volume source, and AI Overviews can move the estimate without a real traffic change. |
| Paid spend | Brand ad spend divided by total category ad spend | Competitive ad intelligence datasets, including Nielsen Ad Intel, MediaRadar, and Pathmatics by Sensor Tower | Paid datasets dominate this channel. Google Ads gives your own spend and auction data, not competitor spend. | Spend datasets estimate competitor budgets from observed placements. Coverage differs by medium and market, and digital walled gardens leave gaps. |
| Share of search | Brand searches divided by searches for all brands in the set | Google Trends and search volume data | Google Trends is free. SEO suites sell search volume reporting. | Search demand reflects interest after exposure, price moves, news, and distribution. It is a demand signal, not proof that advertising caused the searches. |
| AI answers | Brand mentions or citations across a fixed prompt set and engine set | Repeated AI answer collection by engine | CheckThat publishes a public AI visibility index. AirOps, Profound, Semrush, Ahrefs Brand Radar, and Conductor sell AI visibility tracking. | Denominator rules differ by vendor. One tool may divide by total brand mentions, another by responses, and another may weight mentions by search volume. |
Share of voice in AI answers
AI answer SOV measures how often your brand appears when buyers ask category questions in ChatGPT, Claude, Perplexity, and Google AI Overviews. A mention is your brand named in the answer text. A citation is your website linked as a source for part of the answer. A brand can earn citations without being named in the text, and it can be named with no citation to its own website. Track mention SOV and citation SOV as separate lines.
Build the denominator from unbranded prompts that reflect real buying questions, such as category, comparison, problem, and fit prompts. Branded prompts test how engines describe you, so keep them out of the competitive denominator or report them separately. Run each prompt more than once, record the engine, region, language, date, and whether the run used a fresh session, and report the trend as a range when results vary.
As of September 2026, CheckThat measures AI visibility across 5,800+ brands, nearly 200 B2B software categories, and 2.6M+ AI responses.
Share of voice compared with related metrics
Related metrics answer different questions, and swapping one for another changes the decision:
| Metric | What it explains | How it connects | Decision it changes | Limit |
|---|---|---|---|---|
| Share of voice | Your share of visibility or spend in one channel | Compare it with market share to find ESOV | Whether to raise, hold, or cut channel investment | Channel-bound. A high percentage in a narrow set can mean little. |
| Share of market | Your share of category sales, in units or revenue | The outcome SOV is often compared against | Whether visibility is translating into commercial position over time | Sales also move with price, distribution, product, stock, and seasonality. |
| ESOV | SOV minus share of market | Positive values mean visibility share runs ahead of sales share | Budget setting for growth, especially in paid media | The growth evidence is strongest for advertising spend. It does not transfer automatically to mentions or AI answers. |
| Share of search | Your share of brand searches in the category | Share of search research from the IPA reports that share of search represented 83% of share of market on average across 30 cases, 12 categories, and seven countries. | Whether demand is moving before sales data confirms it | The IPA work reports correlation, not causation. Category definition and price still distort the signal. |
| Share of spend | Your share of category advertising spend | The classic spend-based form of SOV | How your budget compares with competitor budgets | Spend says nothing about creative quality, attention, or placement quality on its own. |
| Share of audience | Your share of people reached in a defined audience | Links visibility to people rather than placements, mentions, or spend | Whether you reach enough of the buying audience, and how often | Reach data depends on panels, platform reporting, and identity matching. Duplicated reach across platforms is hard to remove. |
Binet and Field also examined the relationship between ESOV and growth in B2B. The 5 Principles of Growth in B2B Marketing reports a strong correlation between ESOV and market share growth for B2B brands in IPA Databank cases. Treat that as support for testing ESOV in B2B budget planning, not as a promise for any single brand.
Benchmarks for a good share of voice
Benchmark against market share first. Paid media researchers associate sustained SOV below market share with a risk of share loss. If it sits above, you have the condition those studies associate with growth, provided the visibility is positive, prominent, and backed by product, price, and distribution.
The ESOV rule of thumb needs careful handling because the sources give different figures:
- Peter Field's 2009 WARC report on Nielsen research covered 123 FMCG brands. It found that 10 points of ESOV produced 0.5 share points of growth per year, and that bigger brands gained more from the same ESOV.
- The B2B Institute report gives 0.6 percentage points per year for consumer brands and 0.7 for B2B brands per 10 points of ESOV. The report and web chart cite IPA Databank B2B cases but give conflicting period labels of 1998–2008 and 1998–2018. Confirm the period before using the figure in a board deck.
- Nielsen figures reported by the IPA give a market-position split for FMCG: leaders gained 1.4% market share per 10% ESOV on average, against 0.4% for challengers. The IPA page does not state the sample or period for that split, so treat it as a historical indication of brand-size differences, not a precise forecast.
- An IPA post gives a conflicting figure of 1% market share growth for a 10-point gap, citing Binet and Field from 2007. The sources do not reconcile these figures. Use a range, name the source behind the number you choose, and state the population and period.
Peckham's new-brand rule belongs in launch planning. It sets a new brand's SOV at 1.5 times the target market share it wants by the end of year two. The underlying Nielsen work covered grocery, toiletry, and drug products, so apply the rule outside repeat-purchase consumer goods with caution.
No credible benchmark sets a good absolute SOV for AI answers, social mentions, or earned coverage. A leader with 40% mention share in a three-brand social set and a challenger with 12% AI mention share in a 40-brand prompt set are not comparable cases. Set targets from your market share, your prior trend, and the cost of the next point of visibility.
A framework for measuring and acting on share of voice
Use this working method for channel budget and content decisions. Work through the decisions in order, because each one changes the meaning of the next:
| Situation or signal | What to examine | Conditional action | Expected output and check |
|---|---|---|---|
| You have no current SOV baseline | Channels where buyers research the category, plus the competitor set buyers name | Measure paid spend, search, and AI answers first if those drive shortlisting in your category. Add social and earned when conversation shapes the shortlist. | A channel list with one metric per channel. Check that each metric matches a budget or content decision. |
| Two tools give different SOV values | Denominator, competitor set, dedupe rule, geography, language, and window in each tool | Pick one reporting standard. Keep the second tool for diagnosis if it adds source or prompt detail. | One reported figure per channel with the rules written beside it. Recalculate one month by hand to confirm the tool follows them. |
| SOV is below market share | Whether the gap persists for two or more periods, and whether sentiment and prominence are healthy | Raise investment in the channel with the clearest buyer influence. Fix weak creative, content, or source coverage before you buy more volume. | ESOV moves toward zero or positive. Check market share, share of search, and qualified demand over the following periods. |
| SOV is above market share | Creative quality, price position, distribution, and whether visibility reaches in-market buyers | Hold the investment long enough to test the effect. Do not chase a higher percentage if reach quality falls. | A test window with sales and demand checks set in advance. Growth is a possible outcome, not a guaranteed one. |
| SOV rises fast | Sentiment split, source quality, duplicate coverage, and whether a crisis, launch, or campaign drove the rise | Separate crisis volume from planned visibility. Report net sentiment and positive SOV beside total SOV. | A corrected reading that a leader can act on. Check whether branded search and demand moved in the same direction. |
| AI mention SOV lags citation SOV | Pages cited, brands named in text, third-party sources that shape answers, and prompt coverage | Improve citable pages and earn mentions in the sources engines cite. Rerun the same prompt set before you judge the change. | Mention and citation trend lines from repeated runs. Check movement across engines, not only in the aggregate. |
| Competitor SOV rises | New competitor spend, content, PR, product news, and changes in the discovered denominator | Confirm the rise with absolute counts. Respond in the channel where the competitor gained, not across every channel. | A named cause and a targeted response. Check your absolute visibility to see whether you fell or only lost share of a growing total. |
Worked example across three channels
This illustrative example uses hypothetical Brand A and two competitors for a social comparison, alongside category spend and campaign eligibility totals over one 30-day period. It covers three conventional channels, plus AI mentions as a separate comparison:
| Channel | Brand A input | Total market input | Brand A SOV | Reading |
|---|---|---|---|---|
| Social mentions | 2,500 mentions | 10,000 mentions across three brands | 2,500 ÷ 10,000 × 100 = 25.0% | Brand A has a quarter of tracked mentions. Competitor counts are needed to establish its rank, and sentiment is needed to interpret the volume. |
| Ad spend | $80,000 | $1,000,000 category spend | $80,000 ÷ $1,000,000 × 100 = 8.0% | Brand A under-spends relative to an assumed 15% market share, giving ESOV of minus 7 points on spend. |
| PPC impressions | 60,000 impressions | 400,000 eligible impressions in tracked campaigns | 60,000 ÷ 400,000 × 100 = 15.0% | Paid search Impression Share numerically matches the assumed market share, but uses a different denominator. Check campaign eligibility before interpreting the match. |
| AI mentions | 180 brand mentions | 900 brand mentions across tracked prompts and engines | 180 ÷ 900 × 100 = 20.0% | AI mention share runs ahead of the assumed market share. Check citation share, sentiment, and prompt coverage before you call it an advantage. |
Brand A should review the sentiment split before buying more social volume. Brand A should examine why total ad spend sits at 8% against an assumed 15% market share. If the spend gap remains after checking category coverage, test whether added spend in the highest-intent channels improves share of search and demand. Check citation share, sentiment, and prompt coverage before deciding whether to protect or expand the coverage behind the 20% AI mention share. Raising every percentage at once would spend money without naming the constraint.
Where share of voice misleads
Use this diagnostic table as a working synthesis: its possible explanations are hypotheses to check, not confirmed causes. Compare absolute counts and measurement rules before attributing a change to marketing activity:
| Sign or mistaken assumption | Possible explanations | What distinguishes them | Correction | Limit |
|---|---|---|---|---|
| High SOV during a crisis | Negative coverage, social backlash, product failure, or a competitor's crisis that mentions your brand for comparison | Sentiment split, source quality, branded search direction, and sales or pipeline movement | Report positive SOV and negative SOV separately, each using all tracked brands' mentions of the same tone as its denominator. Report net sentiment as (your positive minus negative mentions) ÷ all your mentions × 100. Exclude crisis volume from planned visibility targets. | Sentiment tools misclassify irony, mixed reviews, and industry terms. Sample the underlying items by hand. |
| Mentions rise but buyers do not act | Low-reach posts, duplicate posts, bot activity, passing name-checks, or coverage in publications buyers do not read | Unique authors, reach, prominence, placement in the article or answer, and movement in search and demand | Weight or segment by reach and prominence. Keep the unweighted count as a secondary line. | Reach estimates differ by platform and vendor. A weighted score is only as good as its stated weights. |
| SOV jumps after a tool or set change | New competitors added, a prompt set expanded, dedupe rules changed, or a data source added | Recalculate the prior period under the new rules and compare absolute counts | Mark the break in the series. Publish old and new baselines for the transition period. | You may not be able to rebuild history if the tool did not store the underlying items. |
| Spend SOV rises and growth does not follow | Weak creative, poor distribution, high price, low attention placements, or a shrinking category | Creative testing results, price and distribution checks, attention or prominence measures, and category sales trend | Treat SOV as one input. Fix the limiting factor before adding spend. | The ESOV studies are observational and based largely on case data. They cannot isolate spend from every other change a brand made. |
| AI SOV moves sharply week to week | Normal answer variation, a model update, prompt changes, region or session differences | Repeated runs of the same prompts, a fixed engine set, and the spread across runs | Report a range across runs and a rolling trend. Change prompts only at a marked baseline reset. | Repeated runs are not fully independent samples. A range understates uncertainty if the prompt set itself misses buyer language. |
| A leader assumes low SOV is safe | Large brands often run SOV below market share for a period, but rivals can still take demand at the margin | Share of search trend, challenger ESOV, new-buyer demand, and AI shortlist presence | Set a floor tied to share defence. Monitor challenger movement even when current sales hold. | Jones's 1990 profit-taking observation described popular brands with market share above SOV. It does not prove that under-spending caused their position or remains safe. |
Paid media research supports an association between sustained positive ESOV and market share growth, and going-dark research finds share losses after advertising stops. The IPA also challenges ESOV as a single-factor model for a multi-factor outcome and notes that ESOV models lack the rigour finance teams now expect. Use SOV to set direction and size a test. Do not use it as the sole forecast for revenue.
Apply share of voice to your market
Copy this worksheet into your reporting doc and complete one block per channel:
| Input | Your entry |
|---|---|
| Channel and decision it informs | Example: AI answers, for content and digital PR priorities |
| Metric and unit | Mentions, citations, impressions, spend, estimated clicks, or brand searches |
| Competitor set | Named brands, plus the rule for adding or removing one |
| Denominator | Named set total, discovered market total, or external category total |
| Geography, language, and period | Market, language, start date, end date, and rolling window if used |
| Dedupe and counting rule | Syndication handling, name variants, one count per answer or raw count |
| Formula | (Brand metric ÷ total market metric) × 100 |
| Sentiment and prominence split | Positive, negative, neutral, net sentiment, and prominence rule |
| Market share comparator | Current share, source, unit or revenue basis, and period |
| ESOV | SOV minus market share, for paid spend first |
| Check and next decision | The sales, search, demand, or pipeline check that confirms or rejects action |
Deliverable: one page per channel with the SOV trend, absolute counts, sentiment split, ESOV where relevant, and the decision taken. Check the first calculation by hand from the underlying items. If the hand count and tool count differ, fix the rule before you publish the trend.
Your next decision is which channel gets the test budget for the next period. If maintaining this workflow across separate tools is the bottleneck, the GrowthX platform runs it in one workspace, and you can book a demo to see it. For AI answers, the platform tracks AI visibility on the pages and segments a team chooses, powered by CheckThat. Its segments screen covers ChatGPT, Claude, Perplexity, and Google AI Overviews, and the AI Engine Visibility report adds Google AI Mode.
Selected sources and related reading
These sources carry the load-bearing claims in this article:
- Need to know: What is share of voice? Nielsen's spend-based definition, with market, channel, and time period stated.
- Binet and Field, Marketing in the Era of Accountability (IPA, 2007), The Long and the Short of It (IPA, 2013), and Media in Focus (IPA, 2017). The ESOV research line, drawn from IPA Databank case studies. Read the figures with their population and period, and note that later summaries give conflicting growth-per-10-points values.
- About impression share Google's Impression Share definition and eligibility rules. Search top and absolute top variants have separate definitions.
- Share of search think tank findings IPA report on share of search as a proxy for market share across 30 cases, 12 categories, and seven countries, stated as correlation.
- Comprehensive Advertising Intelligence Nielsen Ad Intel product scope for competitive spend tracking by category, platform, and media type.
