Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.getcited.in/llms.txt

Use this file to discover all available pages before exploring further.

Share of voice is the percentage of total brand mentions in a category that go to a specific brand. It is the most important competitive metric in AI visibility — mention rate tells you whether you appear at all, share of voice tells you how much of the category conversation you own.

​
Why it matters

Mention rate alone is insufficient for competitive benchmarking. A 15% mention rate might be category-leading in one vertical and below average in another. Share of voice normalizes for category density — it measures your presence as a fraction of all competitive presence, not as an absolute number. If your category has five major brands and you account for 40% of all AI mentions, you are the category leader even if your absolute mention rate is only 20%. Conversely, a 50% mention rate in a category with twenty competitors where the top brand captures 85% of mentions means you are not actually winning.

​
Formula

share_of_voice = (brand_mentions / total_category_mentions) × 100
Inputs:
  • brand_mentions — total count of mentions for your brand across all category queries
  • total_category_mentions — sum of mentions for all brands (including yours) across the same queries
Share of voice is computed within-category, not across categories. A brand in the Travel category is compared only to other Travel brands.

​
Example calculation

A skincare brand tracks 50 category queries. Across all runs, AI platforms mention skincare brands a total of 420 times in their answers. Your brand is mentioned 84 of those times. 
share_of_voice = (84 / 420) × 100 = 20%
In this category, your brand owns 20% of the total AI conversation about skincare.

​
Cited Index benchmark — share of voice

The following percentiles are computed from the Cited Index, March 2026 edition — a sample of 253 Indian brands across 8 categories, measured using non-branded category queries. See the benchmarks methodology for details on what these numbers measure and how they differ from customer dashboards.

​
Overall distribution

Across 281 brand-category observations:
PercentileShare of voice
p100.43%
p250.86%
p50 (median)1.67%
p753.66%
p907.40%
The median Indian brand captures just 1.67% of its category’s AI conversation. Even the 90th percentile brand owns only 7.40% — meaning the top 10% of brands in a category still leave more than 92% of the conversation to competitors combined. Category leadership in AI search is not winner-take-all; it is a distributed share.

​
By category

Categories with fewer brands show higher per-brand share of voice because the denominator is smaller — Audio & Wearables brands capture a median 4.04% share compared to just 0.78% in the 55-brand CRM & Sales category.
CategoryBrandsp25p50p75
Audio & Wearables222.46%4.04%5.79%
Travel & Luggage211.86%3.53%5.94%
Skincare & Beauty212.02%2.69%6.73%
Online Learning341.64%2.09%3.60%
Credit Cards421.27%1.90%2.54%
HR & Payroll310.82%1.64%4.60%
Digital Payments550.43%0.86%1.28%
CRM & Sales550.39%0.78%1.55%
Categories with fewer brands naturally have higher share-of-voice percentiles — in a category with 20 brands, the median brand owns 1/20 of mentions at minimum, compared to 1/55 in a large category. Per-category comparisons are more meaningful than cross-category ones.

​
What changes this metric

  • Relative mention volume vs. competitors — if you gain mentions faster than the category average, your share of voice grows even if competitor mentions are also rising
  • Category concentration — if one dominant brand captures most mentions, everyone else’s share of voice stays low regardless of individual mention rate
  • New competitor entries — when a new brand is added to the competitive set, the denominator expands and everyone’s share drops proportionally
  • Query coverage drift — share of voice is sensitive to which queries are tracked; adding queries that favor competitors lowers your share

​
How Cited measures it

Cited runs each prompt in your prompt library and extracts all brand mentions from every response. For your tracked brand and each configured competitor, Cited counts the total mentions across all queries. Share of voice is computed as your brand’s count divided by the sum of all tracked brands’ counts (including yours), then expressed as a percentage. For the Cited Index, all brands in a category are part of the competitive denominator — there is no pre-selected competitive set. Each brand’s share is calculated as its mentions divided by the total category mentions across all brands in that category.

​
Frequently asked questions

Mention rate is an absolute measure — the percentage of queries where your brand appears. Share of voice is a relative measure — your mentions as a percentage of all brand mentions in the category. A 30% mention rate could be high share of voice in a small category or low share of voice in a crowded one.
In most Indian categories, AI mentions are distributed across a long tail of brands rather than concentrated in a few leaders. The median brand captures only 1.67% of category mentions because there are typically 20-55 brands competing for mentions in any given category. Category leaders capture 5-10% share of voice, not 50-80%.
Yes. Share of voice is computed against the total mentions in the category, which includes your own brand. A brand with 100% share of voice would mean only your brand is ever mentioned — which is virtually impossible in any real category.
In the Cited Index benchmark, no — only non-branded category queries are used. In a customer dashboard, share of voice is computed across the full custom prompt library, which typically includes a mix of branded and non-branded queries. This is one reason dashboard share of voice numbers are often higher than benchmark numbers.
Monthly is sufficient for most brands. Share of voice moves slowly — it requires either gaining mentions or competitors losing them, both of which happen over weeks, not days. Checking daily will mostly surface noise from non-determinism in LLM responses.