A CMO asks one question
"What share of consumer attention in our category goes to our brand?"
Three different tools answer three different versions of this question. None of them answer the original.
- Social listening answers: "What share of social mentions are about your brand?" (Vanity metric. Inflated by complaint volume.)
- SEO tools answer: "What share of category SERPs include your URL?" (Ignores that position 1 captures 14× more clicks than position 10.)
- PR analytics answer: "What share of category press mentions are about your brand?" (Inflated by paid distribution.)
None of these are wrong as measurements. They're just measurements of different things, presented as if they answered the original question.
What the right answer requires
A click-weighted, volume-weighted, surface-aware share of voice computation.
For each keyword k in your category:
brand_clicks_k = Σ over URLs owned by brand [ CTR(position_k) ]
Aggregate across keywords:
SoV_brand = Σ_k ( brand_clicks_k × volume_k ) / Σ_k ( total_clicks_k × volume_k )
Three properties matter:
- Click-weighted: position 1 is worth ~14× position 10
- Volume-weighted: a high-volume keyword counts more than a low-volume one
- Surface-aware: Google organic, Amazon, AI Overview each have their own CTR curve
This is the measurement that actually answers the CMO's question.
The CTR curve is everything
Most "SoV" calculations don't even apply a CTR curve. They count appearances. This is mathematically equivalent to claiming a #1 ranking and a #20 ranking are equally valuable. Nobody believes that — but the measurements ship anyway.
Even tools that do apply a curve often use the wrong one. Google CTR has changed materially since 2024 with AI Overview rollout — position-1 CTR roughly halves on AIO-present queries. Running a 2022 CTR curve in 2026 overstates organic visibility by ~2×.
The Theia CTR curves blend public CTR studies, observed DataForSEO CTRs, and client GSC calibration, refreshed quarterly. It's the boring, unsexy work that determines whether every downstream SoV number is right or wrong.
Three surfaces, three SoVs
A complete share of voice in 2026 is multi-surface:
| Surface | Position 1 CTR | Why it matters |
|---|---|---|
| Google organic (no AIO) | ~28% | Classic SEO baseline |
| Google organic (with AIO) | ~14% | The big shift of 2024-2025 |
| Amazon organic | ~35% | For any brand that sells via Amazon |
| AI Overview citation | ~40% | The new battleground |
Single-surface SoV is a 2019 metric. Multi-surface SoV is the new minimum.
The Canon UK cameras analysis we ran in early 2026 showed:
- Google organic SoV: 19% (click-weighted) vs 24% (mention-counted)
- Amazon organic SoV: 31% (click-weighted) vs 28% (mention-counted)
- AI Overview citation share: 12% (the surprise — half of where the dashboard suggested)
The AI Overview number completely reframed the strategy. The brand was assuming citation share roughly tracked organic position. It didn't. Editorial properties (TechRadar, DPReview, Photography Blog) were winning citations Canon had no presence on.
The Bose Germany pattern
This is the single most strategic SoV finding we've published:
Bose converts generic-keyword traffic 10× better than Soundcore — but captures 13× less of it.
Decomposed:
- Soundcore wins 31% of generic-keyword clicks at ~0.3% conversion → big sales contribution
- Bose wins 7% of generic-keyword clicks at ~3.2% conversion → smaller sales contribution
The brand-equity story is real. Bose IS the premium choice. When consumers see Bose at the top of generic search, they buy at 10× the rate of competitors. But because Bose isn't winning generic clicks, the brand-equity advantage compounds into nothing.
This finding is invisible to mention-count SoV. It's invisible to single-surface SoV. It requires click-weighted SoV with conversion attribution, on the right CTR curve, for the right surface.
That's a €1.8M growth opportunity that traditional dashboards literally cannot surface.
Three rules for measuring SoV properly
01 — Use a CTR curve. And make sure it's the current curve for the current surface, not a 2022 baseline.
02 — Multi-surface from day one. Google, Amazon, AI Overview at minimum. Single-surface SoV will mislead you systematically.
03 — Volume-weight. A 50% SoV on a 100-volume keyword is less valuable than a 20% SoV on a 10,000-volume keyword. Most tools that compute "average position" silently violate this.
Why most tools don't fix this
Three structural reasons:
01 — CTR curves are unsexy. Vendors don't publish their CTR curve. Buyers don't ask. The number ships looking like a metric.
02 — Multi-surface integration is hard. Crawling Google SERPs and Amazon SERPs and AI Overview responses is three different problems. Most vendors solve one.
03 — The wrong number sells. A 24% mention-counted SoV sells better than a 19% click-weighted SoV. The customer feels better. The strategy that flows from the higher number is wrong.
The fix requires unbundling the customer's instinct ("higher is better") from the diagnostic value of the metric ("right is better"). That's a hard sell. It's also what the brands building durable category positions are quietly buying.
Coming next
Next post: The 11 shapes of a market — search-based segmentation finds the segments traditional research can't.
Theia computes click-weighted, multi-surface SoV continuously across Google, Amazon, and AI Overview surfaces. See share of voice methodology or CTR curves for the deeper dive.