DAVI
Last refresh · Apr 15 · Next · Jun 15
3
Reference

About the DAVI Score

A two-layer measurement of how AI engines see your destination. The Website Score measures whether AI can read your web presence and the third-party sources that cover you. The Perception Score measures what AI says about you when asked. Both feed the composite Destination Health Index.

01Layer one

DAVI Website Score

76

Measures whether AI engines can find, parse, and extract content from your destination's web presence. Five sub-components: Discoverability (crawler access across 10+ AI bots, sitemap, JavaScript rendering), Site Architecture (download time, JSON-LD structured data, semantic HTML, readability), Editorial Surface (text density, identifiable prose), External & Earned Readability (which third-party sources covering your destination are themselves AI-readable), and the Zertura Destination Layer (structured destination schema, canonical entity definitions, and the Content Deployment Program output that lands on partner sites). Scored 0 to 100. Re-audited monthly.

02Layer two

DAVI Perception Score

92

Measures what AI engines actually say about your destination when prompted. Five dimensions scored equally (Presence, Accuracy, Distinctiveness, Recency, Sentiment) plus a Brand Pillar Alignment sub-component that grades each of your defined brand pillars STRONG / PARTIAL / GAP based on direct LLM querying. Run across ChatGPT, Claude, Gemini, Perplexity, and Grok against 18 standardized visitor-intent prompts per entity, with multiple runs per quarter to capture sampling variance. Scored 0 to 100.

03The composite

Composite formula

Destination Health Index = (0.40 × Website Score) + (0.60 × Perception Score). The 40/60 weighting reflects the priority most destinations place on what AI says over whether AI can read them, but the weighting is configurable per client to reflect local priorities.

Destination Health Index = (0.4 × 76) + (0.6 × 92) = 86
04Methodology note

On measurement and AI's non-deterministic nature.

AI responses are non-deterministic. Any individual response is one draw from a distribution and cannot be predicted exactly. DAVI does not claim to predict individual responses.

DAVI measures the directional distribution of AI behavior across 5 models, 18 prompts per entity, and multiple runs per quarter, with reported variance. This is the same statistical posture used in election polling, clinical trials, and brand tracking studies.

"Stochastic" doesn't mean "unmeasurable." It means "sample-based with confidence intervals." The Website Score answers a determinate question about HTML readability. The Perception Score answers a probabilistic question about AI behavior. Both are operating metrics.

Score bands
01Excellent85 to 100

85 to 100. Best in class for this destination size and category.

02Strong70 to 84

70 to 84. Healthy performance. Protect the gains and look for selective lift.

03Watch55 to 69

55 to 69. Below where it should be and likely to slide further without action.

04Critical0 to 54

Below 55. The dimension is actively pulling the composite down and needs immediate attention.

Powered by Zertura.
DAVI V10.1