How to Calculate Share of Voice, With Formulas

Share of voice calculation divides your brand's measured presence (mentions, impressions, media spend or search visibility) by the total presence of every tracked brand in your market, then multiplies the result by 100. The formula is simple. The hard part is choosing the right numerator and denominator for each channel, cleaning the data, and reading the result without fooling yourself. Buyers now form opinions in Reddit threads, Google AI Overviews, peer reviews and ad feeds, so in October 2026 a single social mention count no longer describes your share of the conversation. This guide covers the core formula, a formula for each channel, worked examples with real arithmetic, a spreadsheet layout you can copy, and the rules that keep a monthly SOV report honest.

Share of Voice Definition

Share of voice (SOV) is the percentage of a market's total visibility that belongs to your brand, measured against a defined set of competitors. In its original advertising sense, share of voice measures a company's media spending as a percentage of total media expenditure in its category. If your brand spent $2 million on TV in a category where everyone together spent $10 million, your SOV was 20%.

Digital channels turned spend into one input among several. Marketers now calculate SOV from:

• Social and forum mentions, including Reddit threads and product communities
• Organic search rankings weighted by search volume
• Paid search impression share from Google Ads or Microsoft Advertising
• Earned media coverage in news and trade publications
• Citations in AI answers from ChatGPT, Perplexity and Google AI Overviews

SOV is a competitive metric. It tells you whether you are gaining or losing attention against the brands buyers compare you with, which raw reach or follower counts cannot. It also puts campaigns in context. A launch that doubles your mentions looks strong until you see that a competitor tripled theirs in the same month.

Share of Voice Formulas for Every Channel

Every share of voice formula follows one structure: your brand's value for a metric divided by the sum of that metric across all tracked brands, times 100.

The denominator does most of the work. "All tracked brands" means your brand plus a fixed competitor list, not the whole internet. Add a competitor halfway through a quarter and every past number changes. Keep the same brands, the same channel, the same time window and the same query or keyword set.

Formula table by channel

Paid search impression share is a different calculation

Paid search impression share has no competitors in the denominator. Google Ads divides the impressions you received by the impressions you were eligible for, based on targeting, budget and Ad Rank. Your 60% and a competitor's 45% in Auction Insights will not sum to 100%, because each brand has its own eligible pool. Report it next to other SOV figures, but never add it into a cross-brand total.

Weighted vs. unweighted SOV

Unweighted SOV treats a reply in a 40-member forum and a post in a 2-million-member subreddit as equal. Weighted SOV multiplies each mention by a factor before summing. Common weights:

• Audience size: followers, subreddit members or outlet circulation
• Engagement: likes, comments, upvotes or shares on the mention
• Estimated reach: the impression estimate your listening tool assigns
• Authority: domain authority or an editorial tier (tier 1 outlet = 3, trade = 2, blog = 1)

Raw follower counts let one celebrity post swamp a month of data. A log weight, such as weight = log10(audience), compresses the range. A 1,000-follower account scores 3 and a 1,000,000-follower account scores 6, instead of a 1,000x gap.

Sentiment-adjusted and positive SOV

Positive SOV counts only favorable mentions in both numerator and denominator. Your sentiment ratio (your positive mentions ÷ all your mentions) describes your brand alone; positive SOV describes your share of all favorable talk in the category. A net version subtracts negatives: Net SOV = (brand positive − brand negative) ÷ total mentions of all tracked brands × 100. Net SOV can go below zero, which makes it a useful alarm during a crisis.

Share of Voice Calculation Example

Each example below takes your brand's count for one metric, divides it by the category total, and multiplies by 100.

Simple example: one keyword, impression-based

Say your brand gets 5,000 impressions for a keyword and the total market generates 25,000:

1. Numerator: 5,000 impressions
2. Denominator: 25,000 impressions
3. 5,000 ÷ 25,000 = 0.20
4. 0.20 × 100 = 20% share of voice

Multi-brand example: four brands over 30 days

Northwind, a cloud security brand, tracks three competitors from 1 to 30 September, after removing duplicates and bot posts. Clean mentions total 3,300; positive mentions total 1,340.

Northwind's mention SOV is 1,240 ÷ 3,300 × 100 = 37.6%, the category lead. Its positive SOV is 410 ÷ 1,340 × 100 = 30.6%, second place. Competitor B gets less attention but more of the favorable attention. A raw-only report would hide that buyers speak better of Competitor B.

Weighted example: when reach flips the result

Brand A earns three mentions on accounts with audiences of 2,000, 50,000 and 500. Brand B earns two mentions with audiences of 400,000 and 1,000.

• Unweighted: A = 3 ÷ 5 = 60%, B = 2 ÷ 5 = 40%
• Reach-weighted totals: A = 52,500, B = 401,000, combined = 453,500
• Weighted SOV: A = 52,500 ÷ 453,500 × 100 = 11.6%; B = 401,000 ÷ 453,500 × 100 = 88.4%

Brand A wins the count and loses the audience. Report both figures, and say which one drives decisions.

Spreadsheet layout for manual SOV

Use one row per brand, channel and period, with raw and cleaned counts side by side so anyone can audit the cleaning.

The SUMIFS denominator sums clean mentions only for rows with the same period and channel, so a new channel or month never contaminates the total. Format column I as a percentage, and add column J for positive SOV with the same SUMIFS pattern on column H.

How Share of Voice Calculation Differs by Channel

Social SOV counts conversations, SEO SOV estimates clicks from rankings, PPC SOV reports auction eligibility, and AI SOV counts citations. For the full playbook on measuring share of voice across search, social and AI, see our dedicated guide; the calculation rules are below.

Social media and Reddit

Build Boolean queries for each brand, including product names, common misspellings and ticker symbols, and exclude homonyms (a brand called "Falcon" needs exclusions for sports teams and birds). On Reddit, decide whether a thread with 200 comments mentioning your brand counts as one mention or 200, document the rule, and apply it to every brand. Engagement-weighted SOV usually suits Reddit best, because upvotes signal which comments readers actually saw.

SEO and organic search

For each tracked keyword, multiply monthly search volume by the click-through rate for each domain's position, then sum by domain. If a 10,000-search keyword gives position 2 an assumed 15% CTR, that domain earns 1,500 estimated clicks. Build the CTR curve from your own Search Console data where possible, since AI Overviews and SERP features push organic results down and lower real CTR. Sample at least a few hundred keywords across awareness, comparison and branded intent; a 20-keyword set produces volatile SOV.

Paid search and display

Pair Google Ads impression share with lost impression share due to budget and lost due to rank. Those splits tell you whether more budget or better Ad Rank would raise your share. For display and social ads, spend- or impression-based SOV relies on ad intelligence estimates, so treat competitor figures as directional.

PR and earned media

Count articles mentioning each brand in a defined outlet list. A syndicated wire story counts once, not once per republishing site. Weight by outlet tier when one feature in a major publication matters more than ten minor pickups.

AI answers

Count how often each brand is named or cited across a fixed set of prompts, run on the same schedule, because answers vary between runs. Citation share (which domains an answer links to) and mention share (which brands it names) are separate metrics; track both. Our comparison of tools for share of answer in AI engines covers how trackers sample prompts.

Regions and languages

Calculate SOV per market first, with translated queries, local brand nicknames and local competitors; a regional challenger may outrank you in Germany and not exist in the US. For a global figure, weight each market's SOV by its revenue or pipeline importance instead of summing raw mentions, or English-language volume will dominate.

Share of voice tools by SOV type

Brandwatch holds a 4.4 rating on G2 from 709 reviews as of October 2026. What reviewers say about its SOV data:

• Reviewers use it daily to track sentiment and share of voice
• Sentiment classification misreads sarcasm and multilingual posts, so verify a sample before trusting positive SOV
• Sampling limits and missing mentions mean its counts may not be a complete denominator

Semrush holds a 4.5 rating on G2 from 3,945 reviews, and reviewers single out its AI visibility tracking; its AI Visibility Toolkit refreshes competitor share of voice weekly and prompt rankings daily.

How to Collect and Clean Share of Voice Data

Clean data comes from a fixed competitor list, a documented query set, deduplication, spam removal, verified sentiment and one consistent time window. Lock these decisions before the first report:

• Define competitors: the four to eight brands buyers actually shortlist against you, plus one adjacent player if buyers compare you to it.
• Choose terms: brand names, product names, abbreviations and misspellings; for SEO, category keywords by intent, not only branded ones.
• Deduplicate: collapse retweets, cross-posts, syndicated articles and quote posts according to a written rule.
• Remove spam and bots: drop giveaway posts, affiliate spam and accounts posting the same text at high frequency.
• Verify sentiment: hand-check a random sample of 100 to 200 mentions per brand each month and correct the automated labels.
• Fix the time window: the same dates for every brand, held constant across reports.

Choosing the measurement window

Search and AI citations move with algorithm and content changes, so measure them weekly. Social and PR are too noisy weekly; use 30-day or calendar-month windows for them. To compare SOV against share of market, use quarterly or annual windows. Compare seasonal categories year over year (October vs. last October), not month over month.

Monthly reporting workflow

1. Export raw data for every brand and channel for the closed period.
2. Apply deduplication and spam rules; log the counts removed in columns E and F.
3. Hand-verify the sentiment sample and adjust positive and negative counts.
4. Calculate raw, weighted and positive SOV per channel with the SUMIFS formula.
5. Compare against the previous period and the same period last year.
6. Annotate every move of more than a few points with its cause: a launch, a news story, a competitor campaign or a ranking change.

How Tellr Helps You Measure and Move Share of Voice

Tellr is a premium earned-visibility agency that measures share of voice in search and AI answers, then does the work that changes it. Each week, Tellr tracks the category's queries on Google, records the organic results, the AI Overview and the discussions block, and labels every cited domain as your site, a competitor, Reddit, social, review sites or references, with week-over-week movement. That gives you a ready-made citation share denominator. A senior team then acts on the gaps in one governed program:

• Reddit: subreddit mapping, a daily thread radar and guideline-checked replies behind an approval gate
• Content: comparison pages and answer-shaped articles written to be quoted by ChatGPT, Perplexity and Google AI Overviews
• Paid media: category ad intelligence and ready-to-run creative
• Reporting: a weekly digest, a monthly program review and API access

Tellr is built for marketing teams spending $10k+ a month at larger companies, so a small team that only needs a SOV tracker is better served by a self-serve tool.

Interpreting Share of Voice: Share of Market, Benchmarks and Mistakes

To read SOV correctly, compare it with share of market, track its direction over time, and check what caused each change. The classic rule in the share of voice literature is that brands whose SOV exceeds their share of market tend to grow, and brands below it tend to lose ground. That gap, called excess share of voice, is more useful than SOV alone.

Share of voice vs. share of market

What counts as a good SOV

No universal number exists, because SOV depends on how many competitors you track. With four brands, 25% is parity; with eight, 12.5% is. Benchmark against three references:

• Your share of market: SOV above it signals growth potential
• Your own trend: a steady rise over two or three quarters beats one spike
• The right level: brand-level SOV against category leaders, campaign-level SOV only against competitor campaigns in the same window

SOV measures attention, not preference. Forrester describes brand preference measurement as a step beyond awareness, scoring engagement, loyalty and advocacy. Pair SOV with the brand awareness metrics you report to the board rather than presenting it alone.

Common calculation mistakes

• Changing the competitor list mid-year without restating past periods
• Adding paid search impression share to a cross-brand total
• Counting syndicated articles and retweets as separate mentions
• Trusting automated sentiment without a manual sample
• Comparing a 7-day window for one brand with a 30-day window for another
• Summing raw global mentions so one language dominates the result

A reliable share of voice calculation comes down to discipline: one formula per channel, a fixed denominator, documented cleaning rules, and positive and weighted versions reported next to the raw figure. Get those right and SOV becomes a leading indicator you can act on. Skip them and it is a vanity number.

FAQ