Share of voice is a brand's portion of all the advertising, or all the visibility, in its category over a set period, expressed as a percentage. A company spending 2 million dollars in a category where total advertising, its own included, reaches 20 million holds a 10% share of voice. The metric exists because a brand's own figures say nothing about the competition: impressions and clicks can rise while a rival triples its budget.

The term carries two meanings. For advertisers it is a planning ratio, set against sales share to decide spend. For publishers it is an inventory allocation, the percentage of ad requests promised to one buyer: the same arithmetic, sold in advance rather than measured afterwards.

How the ratio is built

The calculation divides one brand's activity by the combined activity of a defined competitive set. John Philip Jones of Syracuse University framed the denominator in 1990 as the total value of the main media exposure in a product category, according to Harvard Business Review's summary of his article.

Planners rarely read the number alone. Share of market (SOM), a brand's percentage of category sales, sits beside it, and the gap between them is excess share of voice (ESOV): SOV minus SOM. A brand with 20% of sales and 30% of category advertising runs an ESOV of plus 10. Nielsen, analysing 123 brands across 30 categories in 2009, found that 10 points of ESOV added about 0.5 percentage points of market share over a year, taking its worked example from 20.5% to 21%.

Inputs vary by medium, and most are estimates. Television and print figures come from monitoring services that log rivals' placements and put a cost on them; paid search substitutes impressions, while social and AI tools count mentions.

The denominator decides the result. A category drawn around three rivals flatters a brand that the wider market would dwarf. Guidance from the Interactive Advertising Bureau (IAB), published in August 2026, made the competitive set a mandatory disclosure for AI share of voice, along with how the pool of mentions is sized.

Share of voice as a unit of sale

Sellers build the second meaning into their ad servers. Google Ad Manager's sponsorship line items deliver "a percentage of available matching requests", a type Google says "is sometimes called 'share-of-voice'". A 100% goal serves exclusively unless another sponsorship targets the same inventory. Price plays no part in splitting traffic between competing sponsorships, which are sold per thousand impressions (CPM), per click or per day, and share of voice is one of the dimensions sellers price on their rate cards.

Google extended percentage sponsorships to programmatic guaranteed deals in March 2022, letting buyers reserve a share of a publisher's daily impressions or a daypart through a programmatic guaranteed contract. The Google Ads API lists FIXED_SHARE_OF_VOICE as "a manual bidding strategy for YouTube Sponsorships that bills either on CPM or a fixed cost per day". Version 22, released on October 15, 2025, added support for it, though the value appears in reference documentation from at least version 13.

Retail media followed. Amazon Ads introduced reserve share of voice in October 2025, selling Sponsored Brands top-of-search placements on branded keywords at a fixed price Amazon calculates, with a documentation example of 4,914 dollars for three months. In November 2025 it split its homepage hero placement into 5%, 10%, 25%, 50% and 100% packages.

Lufthansa Group's 2027 media guides price seatback pre-roll the same way: a guaranteed 20% tier reaches 1.9 million contacts a month at an opportunity to see of 1.53, while a 10% tier delivers roughly half of each. The figures are self-reported.

Origin and evolution

The empirical case began in the grocery aisle. James O. Peckham studied A.C. Nielsen store-audit data for grocery, toiletry and drug products between 1960 and 1978, according to a Kendall Hunt marketing textbook, and concluded that a new brand should set its share of voice at about one and a half times the market share it wanted after two years. His work appeared in The Wheel of Marketing, which academic citations date to both 1972 and 1981.

Jones generalised the pattern in "Ad Spending: Maintaining Market Share", published in Harvard Business Review in January 1990, separating small "investing" brands that spend above their market share from larger "profit-taking" brands that spend below it. Les Binet and Peter Field extended the evidence with Institute of Practitioners in Advertising (IPA) Effectiveness Awards data, in books published in 2007 and 2013. Nielsen's 2009 study found leaders gaining 1.4 points of share per 10 points of ESOV against 0.4 for challengers, according to WARC.

Search auctions produced a measured cousin. Google Ads has reported impression share for Search and Display since October 2012 and for Shopping since September 2013, according to its help centre. Auction insights, added to Report Editor in January 2022, set it beside rivals' overlap and outranking rates.

Then the vocabulary travelled. Binet presented share of search at IPA EffWorks Global on October 14, 2020. "By tracking Share of Search we have a powerful, not to mention cheap, metric to measure what people are actually doing online, rather than what they say they are doing," he said. From 2025, AI visibility vendors applied "share of voice" to brand mentions in chatbot answers.

Why marketers track it

Share of voice turns a growth target into a budget question. If the ESOV relationship holds, a brand seeking another point of share can estimate the spend required, and one cutting spend can see the erosion coming. Television still reports this way: iSpot ranked advertisers across more than 25 billion World Cup ad impressions in July 2026, with beer and hard seltzer taking 8.01%.

Automated campaigns, which disclose little else, now carry it. Microsoft Advertising added share of voice metrics to Performance Max in January 2026, with data back to November 10, 2025.

The largest shift is in AI answers, where no ranking report exists. Google added share of voice to Merchant Center's AI performance insights on May 27, 2026, benchmarked against similar brands across AI Mode, AI Overviews and the Gemini app. Semrush reported lifting its own AI share of voice from 13% to 32% in a month, disclosed on October 17, 2025, under a definition that weights position: a brand named third of five scores below one named first.

Where it breaks down

Estimation comes first. Monitoring covers television and print well but sees little of search, social and retail media budgets inside walled gardens, so cross-channel share of voice is often a modelled figure presented as a measured one.

Generalisation is the second problem. Peckham's research covered packaged goods sold through supermarkets and drugstores, and the same textbook cautions that it may not transfer to other products. Nielsen recorded large variances across categories and brands in its own study.

Correlation is not causation. The IPA's share of search think tank, reporting 30 case studies across 12 categories and seven countries in October 2021, found share of search averaging 83% of share of market but described the link as correlated rather than causal, according to The Media Leader.

Sold share of voice carries caveats. Amazon's reserve product delivers top-of-search placement "most of the time". "They are not guaranteeing 100% impression share," said Alexander Swade, an Amazon advertising specialist. Because prices draw on historical branded volume, a brand whose searches fall mid-reservation overpays. The product arrived weeks after the Federal Trade Commission opened inquiries into search ad pricing disclosures at Amazon and Google on September 12, 2025.

AI share of voice is the least stable variant. Output is non-deterministic, and the IAB warns that retraining alone can move a brand's share five points overnight. Search consultant Lily Ray documented in November 2025 how ChatGPT personalisation rewrites prompts per user, so trackers rarely see what customers see. Cloudflare's dashboard, released on August 6, 2026, draws on a benchmark panel built once per category and probes only Claude and GPT models.

Not the same as

Impression share is the percentage of eligible auctions in which an advertiser's ad appeared. It measures presence against potential rather than against rivals.

Share of market counts sales, not communication. The two are compared, never interchangeable.

Share of search counts searches for a brand as a proportion of searches for all brands in its category. It tracks demand that advertising may have created, and Binet framed it as a leading indicator of market share.

Mention rate, in the IAB's AI vocabulary, is the proportion of answers naming a brand at all. Share of voice is relative: mentions can rise while share falls if rivals rise faster. Neither shows whether mentions flatter the brand, a gap addressed by sentiment analysis.

Recent developments

Share of voice is moving into application programming interfaces (APIs). Google Ads API version 25.1, released on August 19, 2026, surfaced share_of_voice through the BenchmarksService. Version 25.2 followed on September 23, 2026 with percentile tiers for competitive standing, though share metrics are withheld for quarters still open, according to Google's release notes.

Standards are catching up. The IAB published Measuring Visibility in the AI Era on August 3, 2026, defining share of voice as brand mentions divided by all brand mentions in a competitive category and favouring ranges such as 22%, plus or minus four points, over bare figures. Paid AI placements sit outside it, and PPC Land noted in September 2026 that no independently audited share-of-voice standard yet exists for ChatGPT advertising.

Sellers keep packaging exclusivity: Spotify offers 100% share of voice on editorial playlists, and TikTok pitches its Category Search Hub as "exclusive share of voice" on generic terms.

Timeline

  • 1960 to 1978: James O. Peckham analyses A.C. Nielsen store-audit data on grocery, toiletry and drug brands, the basis for the rule that new brands need one and a half times their target share in advertising
  • 1972 and 1981: Editions of Peckham's The Wheel of Marketing appear, as dated in academic citations
  • January 1990: John Philip Jones publishes "Ad Spending: Maintaining Market Share" in Harvard Business Review, splitting brands into investing and profit-taking groups
  • 2007: Les Binet and Peter Field publish Marketing in the Era of Accountability using IPA Effectiveness Awards data
  • August 6, 2009: Nielsen publishes its 123-brand analysis finding 0.5 points of share growth per 10 points of ESOV
  • October 2012: Google Ads impression share data begins for Search and Display campaigns
  • September 2013: Impression share reporting extends to Shopping campaigns
  • 2013: Binet and Field publish The Long and the Short of It
  • October 14, 2020: Les Binet presents share of search at IPA EffWorks Global
  • October 12, 2021: IPA think tank reports share of search at 83% of share of market across 30 cases
  • January 2022: Google moves auction insights into Report Editor
  • March 2022: Google Ad Manager allows programmatic guaranteed deals priced by percent of share of voice or daypart
  • September 12, 2025: Federal Trade Commission opens inquiries into search ad pricing disclosures at Amazon and Google
  • October 2025: Amazon introduces reserve share of voice for branded Sponsored Brands placements
  • October 15, 2025: Google Ads API version 22 adds support for fixed share of voice bidding for YouTube Sponsorships
  • October 17, 2025: Semrush discloses raising its AI share of voice from 13% to 32% in one month
  • November 3, 2025: Lily Ray documents how ChatGPT personalisation undermines LLM tracking tools
  • November 10, 2025: Start date of Microsoft Advertising's Performance Max share of voice history
  • November 12, 2025: Amazon introduces fractional share of voice packages for homepage hero ads
  • January 2026: Microsoft Advertising makes share of voice metrics available for Performance Max
  • May 27, 2026: Google introduces AI performance insights in Merchant Center, including share of voice
  • July 27, 2026: iSpot publishes World Cup advertising data, with beer and hard seltzer at 8.01% share of voice
  • August 3, 2026: IAB publishes Measuring Visibility in the AI Era, defining AI share of voice and its required disclosures
  • August 6, 2026: Cloudflare releases its AEO Visibility Dashboard in early access
  • August 19, 2026: Google Ads API version 25.1 surfaces share of voice through the BenchmarksService
  • September 23, 2026: Google Ads API version 25.2 adds percentile competitive tiers and withholds share metrics for open quarters

Summary

Who: Advertisers, agencies and media planners use share of voice to set budgets; publishers and platforms including Google, Amazon, Spotify, TikTok and Lufthansa Group sell it as a percentage of inventory; measurement companies such as Nielsen, iSpot, Semrush and Cloudflare report it, and the IAB has defined its AI variant.

What: A brand's share of total advertising or visibility within a defined competitive category, expressed as a percentage and usually read against share of market through excess share of voice. On the sell side, the same term names a percentage of ad requests guaranteed to one buyer.

When: Peckham's analysis of Nielsen data from 1960 to 1978 grounded the concept, John Philip Jones formalised it in January 1990, and Binet and Field extended it in 2007 and 2013. Google Ads began reporting impression share in October 2012, programmatic guaranteed sponsorships by share of voice arrived in March 2022, and AI share of voice metrics emerged during 2025.

Where: In media planning for television, print and digital channels; in publisher ad servers such as Google Ad Manager; in search and retail media platforms including Google Ads, Microsoft Advertising and Amazon Ads; and in AI visibility tools monitoring ChatGPT, Claude, Gemini, AI Mode and AI Overviews.

Why: Absolute spend and performance figures cannot show whether a brand is being outspent. Share of voice supplies that relative view, though estimated denominators, packaged-goods origins, correlation rather than causation, "most of the time" guarantees and non-deterministic AI answers all limit how far a single figure can be trusted.