Share of Voice Is Not Brand Strength: Use It to Track Competitive Attention, Not Positioning Quality

Share of voice is a brand’s share of a defined pool of competitive attention during a stated period. The pool may be media spend, eligible ad impressions, search visibility, or measured mentions. SOV can show whether attention is shifting within that measured universe. It cannot show, by itself, whether the audience understands the brand, values its difference, believes its claims, or prefers it.

Nielsen’s current explanation frames share of voice as a brand’s media activity relative to category activity and stresses category, market, channel, and time boundaries. Google Ads uses a narrower impression-share measure: impressions received divided by estimated impressions the account was eligible to receive. Both are useful when named correctly; neither is a complete measure of brand strength.

The formula is simple; the denominator is not#

Share of voice = brand's qualifying measure
                 ÷ total qualifying measure for the competitive set
                 × 100

Every term needs a definition:

• Measure: spend, impressions, visibility-weighted rankings, qualifying mentions, or another observable unit.
• Competitive set: the brands included and the rule for adding or removing them.
• Category: the product or problem space being measured.
• Market: country, language, customer segment, or another relevant boundary.
• Channel: television, paid search, organic search, news, social, or another source.
• Window: the exact start and end dates and, for volatile channels, collection cadence.
• Eligibility and exclusions: which ads, queries, sources, duplicates, bots, and ambiguous mentions count.

Without that contract, the denominator can silently change. Adding a competitor lowers every existing brand’s share even if no brand’s absolute attention changed. Removing a source can raise SOV while reducing observed mentions. A percentage may be arithmetically correct and still fail the decision.

One label, several non-equivalent measures#

Google Ads explicitly describes impression share as an estimate based on eligibility, with factors such as targeting, approval status, quality, and auction conditions. It is therefore an account-and-platform diagnostic. Calling it total paid-search SOV would overstate its coverage.

Worked example: attention is not meaning#

The following calculation is synthetic, not market data.

A team monitors five named brands across a fixed set of English-language publications for one calendar month. After applying its duplicate and ambiguity rules, the set contains 2,000 qualifying mentions. One brand appears in 320.

Mention share of voice = 320 ÷ 2,000 = 16%

The result supports one statement: the brand generated 16% of qualifying mentions within that monitored source set and period. It does not establish that the mentions were favorable, reached the intended buyers, expressed the desired association, or caused demand.

To interpret the movement, retain the absolute numerator and denominator. If the brand remains at 320 mentions while total category mentions fall from 2,000 to 1,600, its SOV rises from 16% to 20% without gaining one additional mention. That is a denominator effect, not necessarily stronger attention.

SOV is not market share#

Share of market describes a commercial outcome within a defined market, usually revenue, units, customers, or transactions. Share of voice describes communications or attention. Their units differ.

Excess share of voice is commonly expressed as:

Excess share of voice = share of voice - share of market

Nielsen’s older analysis reports an average relationship between excess SOV and later share change in a bounded fast-moving consumer goods dataset, while also showing large variation. That evidence is useful as a historical hypothesis, not a universal rule for budgets or growth.

SOV is not positioning quality#

Positioning quality concerns what a chosen audience understands and believes about a brand relative to alternatives. It requires evidence about associations, relevance, differentiation, credibility, and choice. SOV measures how much of a selected attention pool the brand occupies.

A brand can therefore have:

• high SOV and unclear meaning because attention is broad but inconsistent;
• high SOV and unfavorable attention because the measure does not encode sentiment;
• low SOV but strong positioning inside a narrow, valuable segment; or
• rising SOV caused by a shrinking denominator rather than increased brand activity.

Awareness and salience are also separate constructs. A mention can contribute to exposure, but counting monitored mentions does not measure what an individual remembers or retrieves in a buying situation. Those questions need audience research under their own design.

Diagnose an SOV change before acting#

A practical SOV reporting contract#

A decision-ready SOV report should show the percentage, absolute brand measure, absolute category total, source coverage, competitor set, collection failures, and definition version. If automated mention classification is used, disclose how ambiguous names, aliases, reposts, and false positives are reviewed.

Trend comparisons require a stable panel. If the source list, keyword set, platform interface, or competitor set changes, mark a break in the series instead of presenting the new percentage as continuous with the old one.

There is no universal good SOV level in the evidence reviewed. A useful target begins with the business question. A challenger testing whether paid presence is increasing needs a channel-specific trend. A communications team diagnosing a launch needs source and message detail. A positioning team needs audience evidence beyond SOV altogether.

What to do with the result#

If SOV falls while the denominator and measurement contract are stable, locate the loss by source, topic, placement, or competitor. If SOV rises but quality measures fall, do not celebrate the aggregate. If SOV is stable and business outcomes change, investigate offer, distribution quality, conversion, and market conditions rather than forcing an attention explanation.