Share of voice: how to measure it without fooling yourself
Four ways to measure share of voice, which one fits which question, and the mistakes that make the number meaningless.
Share of voice is a useful idea that becomes useless the moment you stop defining the space you are measuring. Most reported share of voice numbers are just a tool's default setting, and nobody can say what they are a share of.
Four kinds, four questions
1. Search share of voice
Your visibility across a defined keyword set, weighted by position and search volume, expressed as a percentage of the total available.
The question it answers. Are we winning the terms we care about.
How to do it honestly. Define the keyword set yourself, deliberately, and freeze it. Twenty to fifty commercial terms you genuinely want. Then measure your weighted visibility against your named competitors on exactly those terms, every month.
The mistake. Letting a tool pick the keyword set. Tool defaults include thousands of terms you do not care about, and the resulting percentage moves for reasons that have nothing to do with your business.
2. Share of search
Brand search volume for you as a percentage of brand search volume across your category.
The question it answers. Is demand for us growing relative to them.
How to do it. Google Trends comparison of brand names, or Google Ads keyword volumes for each brand term. Free either way.
Why it matters. This is one of the few genuinely leading indicators in marketing. Brand search tends to move before revenue does, which makes it valuable despite being crude.
The mistake. Brand names that are also ordinary words. If your competitor is called "Apex" you are measuring a lot of things that are not them.
3. Social share of voice
Mentions of you as a percentage of mentions of the set.
The question it answers. Are people talking about us.
How to do it. A listening tool. Brand24 from $199 a month, BuzzSumo from $199, or Google Alerts for a free approximation limited to news and blogs.
The mistake. Counting all mentions equally. One post from an account with reach beats two hundred from nobody, and negative mentions count the same as positive in a raw total.
4. AI answer share
Of the buyer questions you ask assistants, what percentage of answers name you.
The question it answers. When a buyer asks an AI for a recommendation, are we on the list.
How to do it. Twenty questions, no brand names in them, asked across ChatGPT, Claude, Gemini and Perplexity, weekly. Record who is named. Your share is your count over the total.
Why it is the most useful of the four right now. It is closest to a buying decision, and most companies have never measured it, which means the gaps are large and the fixes are cheap.
The mistake. Measuring yourself alone. A 12% share means nothing without knowing that your main competitor is at 40%.
The rules that make any of them honest
Freeze the set. Keywords, questions, competitors. If the set changes, the number is not comparable to last month and the trend is fiction.
Always include competitors. Share is a relative measure. Your number alone is not share of anything.
Report the space. Never "our share of voice is 18%". Always "18% across our 40 target commercial keywords, against these four competitors".
Watch direction, not level. The absolute number depends on arbitrary choices in how you defined it. The direction does not, and the direction is what you act on.
Expect category wide moves. If your share drops and every competitor's drops, something changed in the measurement, not in your marketing. You cannot see this without tracking competitors in the same run, which is the single most common omission.
Which one to pick
If you report one number, pick by what you are trying to influence.
- You are doing SEO work: search share of voice on a frozen keyword set.
- You are building a brand: share of search.
- You are running a PR or social programme: social share of voice.
- You are worried about AI search: AI answer share.
And pick one. Reporting all four monthly guarantees that none of them gets acted on.
Keeping it going
Any of these is easy to measure once and hard to measure for twelve months, which is the timescale where it becomes worth something.
Search share of voice through whatever SEO platform you already pay for. Share of search through Google Trends, five minutes a month. AI answer share by hand at roughly two hours a week, or automatically: Figo runs ten buyer questions across all four assistants weekly with your competitors in the same run, from $49 a month.
The measurement discipline matters more than the tool. A crude number measured the same way every month beats a sophisticated one measured twice.
Questions people ask
What is share of voice?
Your visibility as a percentage of the total visible in a defined space. The definition of that space is the whole argument, and a share of voice number without it means nothing.
What is a good share of voice?
There is no absolute answer, because it depends entirely on how many credible players exist. What matters is your share relative to the competitors you actually lose to, and its direction over time.
Is share of search useful?
Yes, and it is one of the better leading indicators available. Brand search volume relative to competitors tends to move before market share does.
How often should I measure it?
Monthly for search, weekly for AI answers, because AI answers move faster and are noisier.
See it on your own competitors
Figo checks their ads, pages, rankings, reviews and AI answers every week, then tells you what to do in plain words. Set up in two minutes.
Keep reading