Share of Placement.
A new metric for how much of the AI ad layer a brand owns.
Why existing metrics don’t fit.
Brand teams already have a stack of presence metrics. Share of Voice tracks mention volume across press and social. Share of Search tracks query volume on Google. Share of Answer tracks how often a brand surfaces in organic AI citations. Each one was designed for a layer that existed before the AI ad layer did. None of them captures sponsored presence inside AI outputs.
Share of Voice counts mentions. It treats a press hit, a social post, and an earned citation as roughly equivalent weights of attention. That worked when the category being measured was media coverage. It does not work for sponsored AI Surfaces, which are bought, not earned, and which carry a structural disclosure. A brand can have zero Share of Voice in a given week and still own a quarter of the paid AI placements the same week.
Share of Search measures search query volume for the brand against its category. It is a clean number on a clean surface, but it describes intent at the top of the funnel on Google. It tells you nothing about what is happening inside a ChatGPT answer, a Claude response, or a voice agent reply. Agent-layer activity is invisible to it.
Share of Answer is the closest adjacent metric. It measures how often a brand appears in organic AI citations. It is a useful diagnostic, but it has two problems. First, the citation surface is not inventory the brand bought, so the number reflects SEO and freshness more than marketing investment. Second, it is manipulable in ways paid inventory is not, through prompt stuffing, synthetic content, and reputation farming. Share of Placement is the metric that sits cleanly next to Share of Answer: one measures earned AI presence, the other measures paid.
The definition.
“Share of Placement is the percentage of disclosed sponsored surfaces, across a defined AI app network and time period, that a given brand occupies.”
Each clause of that definition carries weight. Defined AI app networkmeans the measurement is scoped to a specific set of apps whose inventory is known. Surfacedd reports it against Surfacedd’s network. A different network would report it against its own. The number is always relative to a named set of Surfaces, never a claim about the entire AI internet.
Disclosed means the metric only counts surfaces that are marked as sponsored. Organic AI citations, retrieval results, and model mentions do not enter the denominator or the numerator. The metric is a paid-presence measure, not a blended one.
Time period means the number is always reported against a window. Weekly and monthly are the default rollups. A Share of Placement number with no period attached is not a complete number.
Brand means the advertiser, as identified in the ad account. Sub-brands and product lines can be rolled up or split out depending on how the advertiser configures the hierarchy.
Why it matters.
Share of Placement is useful because it has four properties that AI-brand monitoring metrics lack. It is buyable: a brand can raise its Share of Placement by spending more, just as it can raise Share of Search by bidding more. The lever is direct.
It is measurable: impressions are counted at the point of render by the ad server, not inferred from scraping or sampling. The denominator is the full set of disclosed sponsored impressions in the network, which the network already knows because it served them.
It is auditable: because the network publishes impression counts, any brand can check its own numerator against reported totals. Third parties can be granted log access for verification. The number is not a vendor-defined score with opaque inputs.
It is disclosed: every surface counted carries a structural sponsored label inside the unit. Share of Placement does not reward brands for hiding. It rewards brands for buying clearly labeled inventory at scale.
Contrast that with the current crop of AI-brand monitoring tools that try to guess how often ChatGPT mentions you. Those tools probe the model with a scripted set of prompts, count mentions, and extrapolate. The output is a directional signal at best. The inputs are not the same prompts real users run, the sample is not the full population, and the mentions are organic, so they shift with model retraining. Share of Placement is not an estimate. It is the network’s own record of what it served.
For the broader context on how brands are thinking about this layer, see reach AI users and AI brand placement.
How to measure it.
The formula is intentionally simple:
Share of Placement = (brand’s sponsored impressions in the network) / (total disclosed sponsored impressions in the network, same period)
A worked example. In a given month, Surfacedd’s network serves 100 million disclosed sponsored impressions across text, image, voice, and code Surfaces. A travel brand running across text and voice receives 7 million of those impressions. Its Share of Placement for the month is 7 percent. The same brand, the following month, raises its budget and captures 9 million impressions against a network total that also grew to 110 million. Its Share of Placement is 8.2 percent. The absolute impression count rose, and so did the share.
The same brand, a month later, keeps its budget flat while two new entrants enter the category. Network totals rise to 125 million. The brand’s impressions stay near 9 million. Its Share of Placement drops to 7.2 percent. The brand spent the same and got less share, because the competitive set grew. That is the metric working as designed.
Share of Placement is reported inside the Surfacedd dashboard at campaign, brand, and category roll-ups. It is available via API for brands that want to pipe it into their own BI stack. For the vocabulary comparison, see Share of Voice vs Share of Placement. For the paid-versus-earned lens, see AEO vs paid placement.
What Share of Placement is not.
Not Share of Voice. Share of Voice counts mentions across earned media. Share of Placement counts paid, disclosed surfaces inside AI outputs. Different input, different layer, different meaning.
Not Share of Search. Share of Search tracks query volume for a brand against its category on Google. Share of Placement tracks sponsored presence inside AI outputs, which is a different surface with a different buying mechanism.
Not Share of Answer. Share of Answer measures organic AI citations. It sits next to Share of Placement in a dashboard, but the two are distinct: one is earned, one is paid. A brand can have high Share of Answer and zero Share of Placement, or vice versa.
Not “how often ChatGPT mentions you.” That phrase describes an organic-mention diagnostic, usually produced by a scripted probe. Share of Placement is about inventory the brand bought in a network that reports impressions.
Not a vanity score. Share of Placement is grounded in served impression counts the network already reports for billing. If the billing number and the Share of Placement number disagree, something is wrong with one of them. That tight coupling is the point.
Frequently asked questions.
Who invented Share of Placement?
Can I measure Share of Placement outside Surfacedd’s network?
How often does it refresh?
Does it apply to voice Surfaces?
What’s a good Share of Placement for a brand?
Reach users inside the AI tools they already use.
CPC from $0.50. CPM from $5. Text, image, voice, and code placements across independent AI apps.