Measuring brand share of voice: search SOV vs. AI SOV

SOV (share of voice) is a relative metric showing how often, and how prominently, your brand is mentioned compared with competitors in a given market or keyword pool. In the AI search era, you get an accurate picture only by splitting this metric into two branches and looking at them together: SEO SOV based on search rankings, and AI Citation SOV based on citations in AI answers. Look at only one, and you miss the retreat happening in the other channel.

SOV as a concept has long been used in marketing. It measures "our slice of the whole pie," like share of advertising or share of media coverage. In search marketing, SOV has long been calculated as the share of top positions you hold across a target keyword set. But as AI search spreads, the very idea of "top positions" has started to wobble. Once AI synthesizes a single answer, whether you rank 3rd or 7th, your share is zero if the AI answer doesn't mention you.

SEO SOV: familiar, but still a valid axis

SEO SOV calculates the share of top positions your pages hold across a target keyword set, or your estimated click share. Using impression and click data from Google Search Console, you can calculate it close to measured values^[3]. This metric is still valid. Search traffic hasn't disappeared overnight, and links are still often shown below AI Overviews^[2].

AI Citation SOV: the axis you need to add

AI Citation SOV calculates the share of a target prompt set in which AI mentions your brand, compared with competitors. For example, you write 20 prompts asking for "recommended products in this category" and count how many times your brand and each competitor appear in each engine's answers. This metric often moves independently of search rankings^[1], so it reveals signals you can never see from SEO SOV alone.

An integrated SOV framework

If you manage the two metrics separately, teams split and reports split, and it's easy to miss the big picture. In practice, we recommend comparing them side by side in one dashboard, as below.

The weights for integrated SOV vary by industry. In a category with a high share of traffic from AI search, raise the weight of AI Citation SOV; in a category still dominated by traditional search, keep the weight on SEO SOV.

A practical guide to measuring SEO SOV

1. Choose 20-30 core keywords that represent your category.
2. Collect each keyword's ranking weekly with a rank tracking tool.
3. Plug rankings into an expected click-through rate (CTR) curve to convert them into estimated click share.
4. Collect competitor domains' rankings for the same keywords and calculate the relative share.

A practical guide to measuring AI Citation SOV

1. Build a set of question-style prompts that real users would plausibly ask while buying or comparing.
2. Query each prompt repeatedly across several AI engines (ChatGPT, Perplexity, Google AI Overviews, and others).
3. Record how many times your brand and competitors are mentioned in the answers, and where (in the first paragraph or further down).
4. Repeat the same method weekly and monthly to watch the trend.

With a diagnosis tool such as TRAIL Search, you can automate this repeated querying and counting and see SOV changes by prompt and by engine as a time series.

An SOV benchmark guide by industry

The baseline for a "healthy SOV" should be viewed differently by industry. In industries with a clear category leader (for example, markets dominated by a few large brands), it's natural for late entrants to show low SOV. Conversely, in fragmented markets the gap between the first and second SOV positions is often small, so even small changes can be meaningful signals. The key is tracking "changes in our relative position within the category" rather than your absolute value.

A diagnosis process for when SOV drops

When SOV falls, narrow down the cause in order.

1. Did the keyword or prompt set itself get wider? As the market grows, new keywords may have increased and naturally diluted your relative share.
2. Did a new competitor enter? Even if the pie stays the same, more players splitting it means a smaller share.
3. Has your content aged? Older content can lose out to competing content with fresher information^[4].
4. Was there an algorithm or model update? Check whether the timing of the drop overlaps with a search engine update or an AI model change.

Integrated SOV through a worked example

Walking through the numbers once makes the concept much clearer. Suppose the target prompt and keyword set has 30 items, and your brand and competitors A and B were mentioned as follows.

This table alone shows an interesting signal. You're first in SEO SOV but far behind competitor A in AI Citation SOV. Had you looked only at SEO metrics, you'd have mistakenly believed "we're leading the market." When you find a gap like this, the next step is to reverse-engineer competitor A's content structure and figure out which elements are giving it the edge in AI citations.

Common mistakes

Mistake 1: Calculating SOV by keyword count alone

If you simply count how many keywords you rank at the top for, high-volume and low-volume keywords are treated the same. Weight them by search volume and prompt frequency.

Mistake 2: Tracking only SEO SOV and ignoring AI Citation SOV

This is the most common and most dangerous mistake. Even when SEO SOV is stable, AI Citation SOV can be quietly collapsing.

Mistake 3: One-off measurement

SOV is not a metric you measure once and forget. The competitive landscape keeps changing, so without periodic tracking you can't tell which way things are heading.

Mistake 4: Lumping all keywords into one SOV

Brand keywords, category keywords, and product keywords behave differently. Calculate by segment to see clearly where you're strong and where you're weak.

In the end, SOV measures "how much of the market pie we hold." Now that the pie is split into two channels, search and AI, only teams that measure both slices separately and then look at them together know their real share.

For related reading, see What is query fan-out? Observing directly how AI splits one question into many searches, Why GEO became a brand reputation problem: when AI describes your brand wrong, and What is the AVI score? A guide to brand visibility metrics in the AI search era.

You can continue with Finding ChatGPT traffic in GA4.