Share of voice that means something: beyond mention counts

Share of voice is your brand's portion of the total conversation in a category. Counted by mentions alone, it rewards noise: a giveaway or a complaint thread inflates it. Measured by theme, emotion and tension, it shows which parts of the conversation you actually own, which you are losing, and the posts that prove it.

Why does mention-based share of voice mislead?

The classic formula is simple: your mentions divided by all mentions of you and your competitors. It is easy to compute and easy to put on a slide. It is also easy to game, and easy to misread.

Consider three things that inflate mention volume without improving your position at all. A contest that asks people to tag your handle produces thousands of identical posts. A delivery failure produces a wave of angry replies. A celebrity mentions you in passing while talking about something else. All three push your share up. Only one of them is good news, and even that one tells you nothing about whether people now think differently about you.

At Somonitor we take a firm view: share of voice should describe the conversation you are winning, not the volume you are generating. That requires reading meaning, not counting words.

What should you measure instead?

We call the better measure share of conversation by concept. Instead of one number, you get a small grid. Along one axis sit the themes that matter in your category: price, reliability, service, design, values, whatever people actually talk about. Along the other sit you and your competitors. In each cell sits the share of that theme's conversation you hold, and the dominant emotion attached to it.

This is where the concept engine from SOMIN, an AI audience-research platform, does the heavy lifting inside Somonitor. It decodes posts into tones, emotions and tensions, and keeps each tag tied to the posts that evidence it. So when the grid says a competitor holds most of the conversation about reliability, you can open the cell and read why.

Definitions worth agreeing on

• Share of voice (SOV): your mentions as a proportion of all tracked brand mentions in the category.
• Share of conversation by concept: your proportion of the posts about a specific theme, such as reliability or price.
• Emotional share: the mix of emotions attached to your share of a theme, for example pride versus frustration.
• Owned theme: a theme where you hold the largest share and the emotion is broadly positive.
• Contested theme: a theme where shares are close or the emotion attached to you is mixed.

How do you build a share-of-conversation grid?

You do not need a data science team. You need a disciplined process and a tool that keeps the evidence attached.

1. Agree the competitive set. Three to six brands. Include the challenger your sales team keeps losing to, not just the incumbents.
2. Let the themes emerge. Resist writing the theme list from your brand guidelines. Run the listening first and see which concepts cluster. The themes your audience uses are rarely the ones in your positioning deck.
3. Strip the noise. Remove contest entries, bot-like duplicates and posts where your name appears incidentally. Keep a log of what you removed so the number stays honest.
4. Compute share per theme. For each theme, what proportion of on-topic posts mention each brand?
5. Attach emotion. For each cell, what is the dominant emotion, and what is the second?
6. Sample the evidence. Read ten posts from every cell that will drive a decision. If the posts do not support the label, the label is wrong.

A worked example

Imagine a mid-sized running-shoe brand in Southeast Asia. Its raw share of voice looks healthy, roughly level with the category leader, mainly thanks to an influencer campaign. The concept grid tells a different story.

On design, the brand owns the conversation and the emotion is delight. On durability, the category leader dominates, and where our brand appears it is mostly attached to disappointment: soles wearing out early. On price, a newer challenger is gaining, attached to a mix of curiosity and scepticism.

The headline number said "we are doing fine". The grid says "people love how we look and doubt we last, and a cheaper rival is getting a hearing". That is a brief a product team and a content team can both act on. The brief would be very different if you only read the volume chart.

For a view of how this kind of evidence-first reading plays out with a real brand, the Exostride case study on the SOMIN site is a useful companion read.

How often should you look?

Share of conversation moves slowly on most themes and quickly on a few. A monthly grid is enough to steer positioning. A weekly check on contested themes catches shifts early. And an always-on alert on sensitive themes, the ones that can turn into a crisis, is what Somonitor is built for. Our piece on reading crisis signals before they trend covers that layer.

Where does share of voice fit with paid media?

Teams often use share of voice to justify spend: if we buy more reach, our share will go up. It will, briefly, and that is the problem. Paid reach raises volume but does little to change which themes you own. If you want to understand whether a campaign is actually profitable rather than just loud, our sister app FMedia looks at true-profit returns on paid traffic. Pair the two and you can see whether spend moved both the conversation and the margin.

Share of voice tells you how loud you are. Share of conversation tells you what you are known for. Only one of those is a strategy.

A short checklist before you report share of voice

• Have contests, bots and incidental mentions been removed, with a log?
• Is the number broken down by theme, not just total?
• Is an emotion attached to every theme you report?
• Can every claim in the report be traced to posts a colleague can read?
• Have you flagged contested themes, not only owned ones?

If the answer to any of these is no, the number on the slide is a volume metric dressed up as a strategy metric. Fix that first. Then share of voice becomes what it should have been all along: a map of the conversation and your place in it.

Frequently asked questions

What is a good share of voice?

There is no universal benchmark. A useful rule is to compare your share of voice with your market share: if conversation share is higher, you are likely to grow attention; if lower, you may be under-discussed. More important is which themes your share sits in and the emotion attached.

Does paid media count towards share of voice?

Paid impressions are not conversation. Somonitor measures organic public posts. Paid activity can raise conversation indirectly, but reporting ad reach as share of voice confuses buying attention with earning it.

How does Somonitor calculate share of conversation by theme?

SOMIN's concept engine tags each post with the themes, tones and emotions it expresses. Somonitor then compares brands within each theme and links every figure to the underlying posts, so the numbers can be checked.