How rank.ai calculates SoV
We compute SoV on two different surfaces, both as a simple ratio.
- Geo-grid SoV. Count the cells where the business appears in the top 3 of the local pack, divide by total cells in the grid. A 10×10 grid with the business in the pack in 40 cells is 40% SoV.
- AI SoV. Count the provider×prompt pairs where the brand is mentioned by name, divide by total pairs. With 9 LLM surfaces (ChatGPT, Claude, Gemini, Perplexity, Grok, Google AIO, DeepSeek, Meta AI, plus the Bing Copilot stub) and 20 tracked prompts, the denominator is 180.
Both surfaces feed a single SoV trend line in the product, so you can see “visibility share” as a unified number across local and AI surfaces — or drill into one side.
Why SoV beats single-position metrics
Average position lies when distribution is uneven. Two stores with the same mean rank can have very different SoV: one ranks #1 in 70% of cells and invisible in 30%, the other ranks #4 everywhere. Same average, completely different real-world visibility. The first store dominates a chunk of its service area and bleeds out near the edges; the second is on the page everywhere but never the first result anywhere.
SoV captures the truth that visibility is binary at each measurement point — you’re either in the visible result set or you’re not. The percentage of measurement points where you make the visible cut is the number that maps to actual click-through.
SoV vs SERP visibility (organic)
Organic SEO tools usually report “visibility” as a weighted score across a keyword set, factoring in search volume and estimated click-through curves by position. That’s a different metric — it’s trying to estimate share of organic traffic across an entire keyword universe.
Our SoV is scoped to local + AI: cells in a grid for local visibility, provider×prompt pairs for AI visibility. We don’t roll in organic keyword volume because the underlying signals are different enough that combining them would muddy both. Track organic visibility separately in whichever tool you use for that surface.
Common SoV benchmarks
Don’t shop for absolute targets. SoV depends on the competitive density of your service area, your category, and how aggressively your competitors operate on Google. A 30% SoV for a plumber in a major metro with 200 active competitors is a different signal than 30% for a rural dentist with 6 competitors in the county.
The benchmark that actually matters is your historical baseline. Track SoV weekly, watch the trend line, and diagnose drops the same way you’d diagnose any ranking drop — a sudden 10-point fall says something changed (a competitor opened a new location, Google rolled out an update, your GBP got temporarily disabled). The trend is the signal.