Docs
Concepts

Visibility score

What your visibility score means, how it's calculated, and how to read it

Your visibility score is the primary metric in Heading. It tells you how often your brand appears when people ask AI assistants questions relevant to your business.

How it works

Heading monitors a set of prompts across the platforms the property tracks. Each response is analysed for two signals:

  1. Citations: the AI links directly to your website as a source
  2. Mentions: the AI names your brand in the response text

Visibility measures detected mentions in answer text. Heading matches the brand name, configured aliases and domain. Citations are recorded separately; a source link alone does not increase visibility.

Calculation

Daily visibility = (responses with a detected mention) / (recorded responses) × 100

For example, if one platform records 100 responses and 42 contain a matching mention, daily visibility is 42%. A recorded result with no AI answer counts in the denominator. A failed fetch with no stored response does not. Check collection coverage when comparing periods.

The daily cross-platform aggregate pools recorded responses. Period summaries average daily values. Rolling averages use days with data; missing days are not filled with zero.

See the measurement methodology for a worked example and limitations.

What the pipeline records

Per prompt, per tracked platform, once per UTC day:

  1. The prompt is sent to the platform
  2. The response is analysed for the brand name and the domain
  3. Citations (source links) and mentions (brand name in text) are recorded
  4. On paid plans, the day's results are aggregated into the time series

Sub-metrics

Citation rate

The percentage of monitored prompts where your brand is cited with a source link. A citation provides a link to inspect; it does not prove a visit or conversion.

Mention rate

The percentage of monitored prompts where your brand name appears in the response, even without a link. Mentions indicate brand awareness but don't directly drive traffic.

A brand can be mentioned without being cited (the AI talks about you but doesn't link to you), or cited without being mentioned (the AI links to your page but describes it generically).

The full distinction, with examples, is in Citations vs mentions.

Free and paid data

Free propertyPaid property
ShapeOne point-in-time snapshotA daily time series
DatingLabelled "As of" its snapshot dateOne value per UTC day
Trends and deltasNone. There is no previous period to compare againstChange against the previous period
Comparison controlsNoneDate range and comparison window

See Your free snapshot report and Plans.

Reading the dashboard

  • Trend arrows: change against the previous period. Paid properties only
  • Green and red: improvement and decline
  • Platform breakdown: how often a matching mention appears on each tracked platform

Comparing scores

There is no universal benchmark. A score is a share of the prompts you chose, on the platforms you track, in your category, so a bare percentage carries no meaning outside that set.

Two comparisons do hold:

  • Against yourself over time. The daily series on a paid property is measured over the same prompt set from one day to the next, and the prompt changelog records the days the set itself moved. See How prompts work.
  • Against the competitors tracked on the same prompts. The leaderboard ranks you and each tracked competitor over one prompt set, so the gap between the rows is measured rather than inferred. See Competitor tracking.

Don't compare scores across different properties or industries. A 30% score for a local plumber in a specific city is a different thing from a 30% score for a global SaaS product.

Improving your score

Common improvements include:

  • Adding structured data to your website
  • Creating content that directly answers common AI prompts
  • Building authoritative backlinks that AI platforms trust as sources
  • Ensuring your brand name is consistently used across your web presence