What is share of voice?

What is SEO share of voice?

SEO share of voice compares a website’s measured visibility with a defined competitive dataset. Its percentage depends on the query set, competitors, market and scoring method. It describes that sample rather than the business’s share of all customers or revenue.

• The numerator represents the site’s measured visibility.

  The denominator represents the corresponding total used for comparison. Both must describe compatible units. Counting your appearances and dividing by competitors’ estimated clicks would combine unlike quantities and produce a number with no clear interpretation.

• Some methods count qualifying appearances.

  Others assign scores to positions or estimate potential clicks using demand and click models. These approaches answer different questions. A percentage labeled only “share of voice” does not tell a reader which method was used or what kinds of exposure were included.

• The SERP is the observable search-result environment behind many SEO comparisons.

  A tool samples results under selected conditions, identifies domains or URLs, and applies its method. It does not directly observe every potential customer’s searches or every competitor’s private analytics.

• For a service business, define the comparison around decisions customers actually make.

  A boiler repair company might compare non-brand repair searches in its service area. Including unrelated installation searches or distant locations would change what the score means and could distort priorities.

• Treat the metric as a comparative diagnostic.

  It can help identify where relevant visibility is concentrated or missing. It does not establish revenue, customer preference, service quality, or the share of jobs the business wins. Those outcomes require different evidence.

• Search volume is an estimate of demand for a defined market and period; a visibility model that uses it inherits those limits.

Which measurement method should you choose?

Choose a method that fits the decision and can be described clearly enough to reproduce. Appearance counts show presence within a defined search sample, position-weighted scores emphasize ranking, and estimated-click models combine additional assumptions. None is universally correct. Consistency and transparent limits matter more than choosing the method that produces the largest percentage.

• An appearance method needs a threshold.

  For example, a business can count whether each tracked query has a result from a domain within a specified result range. The threshold and treatment of multiple URLs must be explicit. Otherwise, one supplier can count several appearances where another counts only one.

• A position-weighted method assigns greater value to some positions.

  Ask for the weight schedule and whether it changes by device or result type. A score generated from undisclosed weights can still be useful inside one tool, but it is harder to compare with another provider’s output.

• An estimated-click method often multiplies demand estimates by assumed click opportunities.

  Its output is modeled, not observed competitor traffic. Preserve that label in reports. Calling the score “click share” without explaining the estimate can make it sound like actual visits were measured.

• A simple method can be preferable when the dataset is small and the decision is practical.

  If the question is whether the company appears for its main local services, a transparent coverage measure may be easier to use than a complex model whose assumptions are unavailable.

• A more detailed model can be useful for prioritizing a broad service portfolio.

  It still needs sensitivity checks. A result dominated by one high-volume term may tell a different story from an equally weighted view of important booking needs.

How does the denominator change the interpretation?

The denominator defines the comparison’s boundaries and can move the percentage even when the site’s own visibility remains unchanged. Decide whether it includes named competitors, all observed domains, or a theoretical opportunity total. Each choice has consequences, so report the population explicitly rather than presenting the result as an unqualified share of all search demand.

How would a transparent calculation work?

State the units and population before performing the arithmetic. Identify both the numerator and denominator. The invented examples below illustrate two different methods with hypothetical data. They are teaching calculations, not observed search metrics, industry benchmarks, or estimates of what any business earns. Their purpose is to show why the same label can conceal different comparisons.

• In an invented appearance-count example, a fixed query set produces forty qualifying domain appearances across the selected competitors.

  The repair company’s domain accounts for ten. Its share within that defined sample is ten divided by forty, or twenty-five percent.

• That figure does not mean the company appeared in twenty-five percent of all real searches.

  It describes the counted appearances under the chosen threshold and sampling rule. If the method counted only a domain’s strongest appearance per query, document that choice beside the result.

• Now consider an invented weighted-score example.

  The company’s assigned visibility score is thirty units, and the selected competitors together have seventy. The comparison total is one hundred units, so the company’s share is thirty percent. Those units are model scores rather than visits or impressions.

• The two hypothetical percentages cannot be treated as interchangeable.

  One counts appearances; the other distributes weighted points. A company could have many low-ranked appearances and a smaller weighted share, or fewer strong appearances and a larger weighted share.

• A third invented situation illustrates denominator movement.

  Suppose the company’s score remains thirty units but newly included competitors add fifty units to the comparison. Its share becomes thirty divided by one hundred fifty, or twenty percent. The drop follows the expanded population, not an observed decline in its own score.

• Use these distinctions when reviewing an agency report.

  Ask for both the score and the population behind the percentage. A movement that cannot be decomposed into those components may be difficult to act on responsibly.

How should the query set be built for a service business?

Build the set around real services, customer tasks, and coverage areas, then identify which questions belong in separate segments. Include enough breadth to avoid a flattering but narrow sample. A useful set supports decisions about actual offers rather than maximizing visibility on easy queries that do not represent the customers the business wants to serve.

• Start with the services the company can deliver.

  An appliance repair provider may separate washing-machine repairs from dishwasher work. Refrigeration can form another group. If it does not repair commercial equipment, commercial queries should not quietly inflate the opportunity set or become targets for unsupported pages.

• Distinguish booking searches from informational research.

  “Dishwasher repair near me” and “why does my dishwasher leave residue” can lead to different pages and customer stages. Both can matter, but combining them without segmentation can hide where the business is visible or absent.

• Keep branded searches separate from generic service demand.

  A company may dominate its own name while struggling to appear for repairs. A blended score can obscure that distinction. Brand performance is useful information, but it answers a different question from competitive discovery.

• Use keyword mapping to connect query groups with appropriate pages.

  That relationship helps interpret a weak segment. The issue might concern a missing service explanation, a page with the wrong scope, or a mismatch between the query and the business’s offer.

• Check synonymous queries for duplication.

  Some wording variations represent distinct observations but not distinct customer needs. A model weighted by search estimates can double-count overlapping demand if its inputs are treated carelessly. Document the deduplication convention rather than assuming every listed phrase adds a separate market opportunity.

• Retain an accessible definition of the query set in the report.

  A percentage without the covered services and exclusions invites broader interpretation than the evidence supports. The full list can live in an appendix when it is too long for the main explanation.

How should demand estimates and weights be handled?

Use demand data as a stated input with its source and observation period recorded. State the geography too. Estimates can help prioritize queries but should not be mistaken for exact local customer counts. A weighting system can change the comparison substantially, so test whether important conclusions survive reasonable alternative assumptions instead of hiding those assumptions behind one precise-looking percentage.

• Google’s Keyword Planner metric guidance defines average monthly searches using selected settings and close variants.

  Those statistics are not the same as a company’s available repair jobs. The relationship between search activity and suitable customers requires further evidence.

• The search volume input can vary by source and scope.

  A national estimate should not be presented as local demand without a defensible adjustment. Narrow service areas and infrequent searches can make apparent precision especially misleading.

• Compare a demand-weighted result with an unweighted view where practical.

  If one broad informational phrase dominates the weighted score, the comparison may underrepresent commercially important services. That does not make weighting wrong; it shows what the model emphasizes.

• Do not equate Keyword Planner’s advertiser competition with organic difficulty.

  The metric describes advertising competition under its definition. It does not directly tell you which competitors will appear organically or how hard a service page will find it to rank.

• If weights are updated with new demand estimates, annotate the change.

  A seasonal query can gain weight while the site’s observed position remains constant. The resulting share movement should not be described entirely as an SEO performance change.

Which competitors belong in the comparison?

Include sources that compete for the customer’s attention in the defined search task, while distinguishing them from businesses that compete for the same job. Search competitors can include directories and informational publishers. A useful report can show both comparisons without pretending every visible domain has the same business model or converts visibility in the same way.

• A repair company may name nearby firms as commercial competitors, yet observe national directories above them in search.

  Excluding directories can answer a peer-business question. Including them can reveal where search attention is distributed. Label the comparison so the reader knows which question it addresses.

• A manufacturer might rank for troubleshooting questions while offering no local repair appointments.

  Its presence can influence the journey, but it is not necessarily competing for the same booking. Separate informational and service-result segments rather than treating all visibility as interchangeable.

• Include new entrants through a documented review process.

  A domain can become relevant because it changes its content or service coverage. When adding it, retain an overlapping historical comparison where available or clearly mark the start of the expanded competitive set.

• Check organization-level grouping.

  Several domains can belong to one directory network, while one domain can host unrelated businesses. Decide whether the comparison concerns domains, brands, or organizations. That choice affects both counting and the story told by the percentage.

• Do not infer revenue from a competitor’s visibility.

  A prominent directory may collect inquiries differently from a contractor. Public results cannot establish its margins, lead quality, or booked work. Use visibility to understand the search results, not to invent a competitor’s financial performance.

• A keyword gap review can identify relevant topics where other sources appear and the company does not.

  Verify each gap against the offer before acting. A competitor’s presence is a research signal rather than proof that the business should publish an equivalent page.

How do location, device, and result features affect the score?

Keep observation conditions consistent and state which result types the method includes. Local context, device, and feature handling can change visibility without a simple movement in ordinary links. A blended score can be useful, but it must not conceal whether the business gained local exposure, informational coverage, or a different kind of presentation.

• A service business should sample locations that represent its actual coverage.

  One city-center observation may not describe searches across a large service area. A location-specific comparison should identify its points or convention instead of using a generic “local” label.

• Separate desktop and mobile where differences affect the decision.

  If a tool changes the device mix between periods, its aggregate score can change. Keep the mix stable or report segmented results so the reader can understand the source of movement.

• Decide how the map pack is handled.

  Some visibility models include it; others concern ordinary organic results only. Adding a map position to a text-result score without compatible weighting mixes different presentations and can make the total difficult to interpret.

• Featured answers and other SERP features also need explicit treatment.

  A source can gain prominence without changing a simple link rank. Conversely, a rank may remain stable while the surrounding layout changes how much attention it receives.

• A model may choose to omit certain features because their scoring is uncertain.

  That can be a reasonable scope choice if disclosed. Do not label the result “total search visibility” when meaningful presentations were excluded from the method.

• When a tracker changes detection rules, inspect the impact before interpreting the trend.

  A newly recognized feature can add score units even if the page’s actual presentation did not change. Tool methodology updates belong in the change record alongside edits to the site’s content.

What can your own Search Console data verify?

Use Search Console to understand the site’s reported search performance, not to reconstruct competitors’ private datasets. Its impressions, clicks, and positions follow defined counting and aggregation rules. They can corroborate or challenge a visibility story, but they are not automatically the numerator or denominator of a third-party share-of-voice model.

How should a business investigate a sudden score change?

First check whether the dataset or calculation changed, then locate the services and queries contributing to the movement. Review underlying observations before choosing an intervention. A percentage alone cannot distinguish a genuine visibility loss from new competitors, changed weights, missing samples, or a tool update, so diagnosis should proceed from components to causes.

• Confirm that the same query set was collected under comparable conditions.

  Missing observations can alter the denominator or reduce score coverage. A failed location sample should not silently become a zero that looks like the business disappeared from search.

• Check changed weights and newly added domains.

  Recalculate a stable comparison where possible. If the historical series cannot be restated, mark the break and avoid treating the old and new percentages as a continuous trend.

• Identify the largest contributing segments.

  A single service group may explain most of the movement. Investigate its landing pages and search results rather than rewriting the entire website in response to one aggregate line.

• Compare the underlying rankings or appearances with your own performance data.

  A modeled decline can coexist with stable clicks when demand or presentation changes. That combination needs interpretation, not an automatic declaration that the business lost a corresponding percentage of customers.

• Inspect relevant page changes where evidence points to them.

  Check availability and indexing separately. A technical issue can affect visibility, but the share metric alone cannot identify it. A technical SEO review should follow specific symptoms rather than a speculative explanation for every fluctuation.

How would an appliance repair company use the comparison?

Define a stable service-query sample, separate brand demand, and use the results to investigate specific gaps against actual customer outcomes. The invented example below illustrates that workflow without reporting real rankings, percentages, or inquiries. Its purpose is to show how a comparative metric can support a decision rather than become a flattering dashboard number.

• Imagine a company that repairs washing machines and dishwashers but not refrigerators.

  Its initial tracker includes all three categories. The team corrects that scope before drawing conclusions. Refrigeration searches do not represent an offer the company can currently fulfill.

• It then separates brand searches from non-brand repair searches and records a defined service area.

  The non-brand comparison includes local providers and directories, with their roles identified. The method counts qualifying appearances under a consistent threshold.

• The dishwasher group appears weaker than the washing-machine group.

  The team reviews its relevant page and finds that the repair scope is poorly explained. That is a concrete editorial issue worth correcting, independent of whether it causes the whole measured difference.

• The revised page clarifies supported appliances and the assessment process.

  Later observations compare the same query set and method. Own search data and inquiry review provide additional context about whether the page attracts suitable customers.

• If the company subsequently adds refrigeration services, the query set can expand deliberately.

  The report marks the new scope rather than describing the changed aggregate as a sudden competitive loss. A stable historical segment remains useful for evaluating earlier work.

What should the final report make clear?

State what was measured, how the percentage was calculated, and which decisions it supports. Include the dataset boundaries and meaningful changes in method. Pair the comparison with the company’s own outcomes without claiming they are equivalent. A useful report helps the business choose an investigation or action while preserving uncertainty about unobserved competitors and customers.

• Is share of voice market share?

  No. Search visibility in a selected dataset does not measure the share of completed jobs or revenue. A company can appear frequently while receiving unsuitable inquiries, and a specialist can win valuable work from a narrow search segment.

• Can two tools be compared directly?

  Only after their population and methods are understood. Different query sets, weights, or feature rules can produce different percentages. Comparing the displayed numbers without those details can create a false conclusion about performance.

• Should the target be the highest possible score?

  The business should prioritize relevant coverage and suitable outcomes. Dominating unrelated informational searches may be less valuable than explaining a profitable service clearly. Choose actions around the offer and evidence rather than an abstract maximum.

• What must remain reproducible?

  The query-set definition and comparison population should be available, alongside the calculation convention and observation context. A report that conceals its denominator cannot support a strong claim about broader visibility or competitive movement.

Primary documentation

Google Search Console Help: Understanding impressions, position, and clicks. Google Ads Help: Keyword Planner historical metrics and forecasts. Accessed 8 October 2026.