Visibility and citation signals
Platform or third-party observations, with their own sampling, coverage, definitions, and collection windows.
Research · Measurement · Analysis
AI Search Data
Original research, measurement frameworks, benchmarks and analysis on how AI search influences website discovery, traffic, conversions and customer journeys.
A research property built around evidence, defined methods, and visible uncertainty—not unsupported numbers.
What this property covers
Separate the signals before connecting them.
These terms describe different stages and sources. They should not be treated as interchangeable metrics.
- 01AI Search Visibility
- Evidence that a brand, page, or citation appeared in an AI search environment, as defined by the source that collected it.
- 02AI Search Traffic
- Website visits with a source or referrer signal that can be validated as AI-related. This is not a full count of AI influence.
- 03AI Search Engagement
- On-site actions and sessions recorded after a measurable visit, using documented definitions.
- 04AI-Assisted Conversion
- A conversion associated with an observed or modeled AI touch under a stated attribution scope—not proof of causation.
- 05CRM-Confirmed Outcome
- A lead or later business stage recorded in the CRM, reconciled only where an approved and reliable connection exists.
Measurement hierarchy
Follow the journey; preserve the gaps.
Not every stage is directly observable, and attribution may not carry context from one system to the next.
- 01AI Search VisibilityExternal visibility signal
- 02AI Search ReferralValidated referral
- 03Website EngagementRecorded on-site activity
- 04ConversionDefined key event
- 05CRM LeadAccepted business record
- 06RevenueWhere measurable
AI Search Visibility → AI Search Referral → Website Engagement → Conversion → CRM Lead → Revenue
Types of evidence
Analyze the source in its own context.
Referral and landing-page signals
Source, medium, referrer, and destination information received by a website and its analytics implementation.
Engagement and conversion signals
On-site events recorded according to the site’s GA4 definitions, consent behavior, and collection setup.
CRM and revenue outcomes
Business-system records that may be reconciled where identifiers, permissions, and business definitions permit.
Research methodology
Make the method as visible as the finding.
- Define the question, population, and measurement unit before collecting or comparing data.
- Name each source and explain how it was collected, classified, and checked.
- Distinguish direct observation from attribution, modeling, and inference.
- Record dates, coverage, exclusions, known gaps, and any change in method.
- Avoid unsupported causal conclusions and do not turn missing signals into assumed activity.
Published assets
Available now
Frameworks and explainers are not benchmark datasets. No original statistical dataset or survey findings are currently claimed.
Framework · PublishedAI Search Measurement FrameworkDocument the journey across visibility, referral, engagement, conversion, CRM, and revenue.Flagship guideAI Search Analytics: How to Measure AI-Driven Customer JourneysLearn the implementation and reporting approach for observable signals and their limits.Measurement articleHow to Measure AI Search Referral TrafficValidate observed referral values and build a transparent source grouping.
Research areas
Questions for future analysis.
These are areas of interest, not studies or data assets currently available.
- How AI visibility signals relate to observable referral traffic under stated source definitions.
- How measurement coverage varies across referral, engagement, conversion, and CRM stages.
- How teams can document AI-referral definitions and keep comparisons consistent over time.
- What can and cannot be concluded when visibility evidence and web analytics use different methods.
Related work
From research framework to implementation.
Measurement support
Explore AI Search Analytics Need to connect the signals in your own measurement setup?
Start with the data you can validate, define what matters, and document what remains unknown.