MethodologyVersion 1

How Drifter measures what AI tells travelers

Drifter measures AI visibility with repeatable traveler questions, a consistent four-system panel, source evidence, and a receipt behind every number. You can inspect what was asked, what the evidence supports, and what the result cannot prove.

Skip to what we check
Every engine’s saved answer to question 50 of La Jolla Travel’s blind traveler questions, with the websites the four panel answers cited.
Evidence receiptLa Jolla Travel · Question 50

The question

“Which coastal towns in the Southern California Coast are best for outdoor trips?”

The question

Questions that never name you.

Every report starts from a working model of your place: the areas travelers weigh it in, and the cities they plan from. The 50 blind questions in La Jolla’s report come from it, and not one names La Jolla.

The Place ModelLa Jolla

A scenic coastal playground for beach walks, viewpoints, and easy outdoor time

Areas
La Jolla, San Diego · the Southern California Coast · the California Coast
Feeder markets
Los Angeles · Orange · Phoenix · Las Vegas · San Francisco Bay
Question 43
“top coastal towns in the Southern California Coast for travelers from Phoenix seeking food and drink trips”
Blind traveler questions
Only questions that never name you count toward Answer Share. Questions that name you show how AI describes you, and questions that weigh you against other places show how AI compares you.

The answers

One engine is an anecdote. A panel is a measurement.

Asked the same question, ChatGPT, Claude, Gemini, and Perplexity sent the traveler first to four different towns. Three of the four named La Jolla.

The answers

La Jolla’s place in each answer

  • ChatGPT named ten towns, Laguna Beach first and La Jolla fourth. In order: Laguna Beach, Dana Point, Huntington Beach, La Jolla, Encinitas, Oceanside, Ventura, San Clemente, Carlsbad, Coronado.
  • Claude named ten towns, Santa Barbara first and La Jolla seventh. In order: Santa Barbara, Malibu, Encinitas, Oceanside, Del Mar, Solana Beach, La Jolla, San Clemente, Coronado, Laguna Beach.
  • Gemini named ten towns, La Jolla first. In order: La Jolla, Laguna Beach, Santa Barbara, Encinitas, Malibu, Avalon, Coronado, Carlsbad, Ventura, San Clemente.

    “Renowned for its dramatic ocean cliffs, marine ecological reserves, sea caves, and exceptional outdoor activities like kayaking, snorkeling, and tidepooling.”

  • Perplexity named ten towns, San Clemente first. It did not mention La Jolla. In order: San Clemente, Laguna Beach, Ventura, Dana Point, Encinitas, Carlsbad, Huntington Beach, Imperial Beach, Avalon, Newport Beach.
  • Beside the panel
  • Google AI Overview named three towns, Laguna Beach first and La Jolla second. In order: Laguna Beach, La Jolla, Malibu.
Answer Share
Answer Share is a provider-weighted average. Within each provider, Drifter deduplicates eligible blind-discovery questions and divides the questions that mention the target place by all eligible questions sampled for that provider. It then combines those rates using the current configured mix: OpenAI 35%, Anthropic 25%, Gemini 25%, and Perplexity 15%, renormalized when a provider is absent. These are product aggregation weights, not observed market-share measurements.
Google AI Overviews
Paid plans also read Google AI Overviews, the AI answer at the top of Google Search, as a distinct lane, reported beside the panel rather than blended into its weighted averages.

Sources in the answers

An answer is only as good as its sources.

Every citation in every answer is traced to its website and sorted by who publishes it.

Sources in the answers

Citations

visitlagunabeach.com
9
visitcalifornia.com
7
ca.gov
3
thetravel.com
3
travelandleisure.com
3
purewow.com
2
15 more websites
15
21 websites in the four panel answers
42
Authority Share
Authority Share is the number assigned to owned domains, divided by all classified citation occurrences in the eligible sample set. The report opens with a bounded summary for a fast first read, then loads every cited source and page behind the score on demand, and splits the same evidence by AI assistant.

Evidence receipt

Every number has receipts.

Drifter keeps each answer with its question, its engine, when it was asked, the places it named in order, and its citations. Nothing is filled in.

Before and after
Completed work sits beside later answers, so your team can compare what was asked, what changed on the site, and what the answers show next. A before-and-after sequence is descriptive unless the measurement design supports a causal conclusion.

Technical audit

What we check and what it proves

Technical-audit findings are limited to the signals observed during the run. Here is each check, what it can establish, and what it cannot.

Crawl access3 checks

AI crawler access

What we observe
We fetch your robots.txt and read the published rules for each AI crawler we track.
What it proves
Whether your published rules allow or disallow those crawlers.
What it cannot prove
Whether a provider actually visits your site, or when.

Sitemap

What we observe
We request /sitemap.xml, follow any sitemap your robots.txt declares, and confirm the winning response is an XML sitemap.
What it proves
A sitemap is published and readable at a discoverable location.
What it cannot prove
Whether search or AI systems have fetched it.

Canonical URL

What we observe
We read the canonical link tag from your homepage HTML.
What it proves
The tag is present or missing.
What it cannot prove
How crawlers resolve duplicate URLs across the rest of your site.

Homepage reachable: a completed audit first verifies that the homepage responded with a readable page instead of an error or a security wall. This is a precondition, not a scored check.

AI metadata5 checks

Title tag

What we observe
We read the title tag from your homepage HTML.
What it proves
Presence and the exact text served.
What it cannot prove
Which title an engine chooses to display.

Meta description

What we observe
We read the meta description from your homepage HTML.
What it proves
Presence and the exact text served.
What it cannot prove
Which snippet an engine chooses to display.

Language attribute

What we observe
We read the lang attribute on your html element.
What it proves
The declared language.
What it cannot prove
How well systems handle your content's language.

Open Graph tags

What we observe
We read Open Graph and Twitter card tags from your homepage HTML.
What it proves
Presence and completeness of social metadata.
What it cannot prove
How each platform renders your preview.

llms.txt

What we observe
We request /llms.txt.
What it proves
Whether the file is published.
What it cannot prove
Whether any AI provider reads it. Major providers have not confirmed doing so.
Structured data2 checks

Structured data

What we observe
We parse JSON-LD from your homepage HTML, plus rendered evidence when available.
What it proves
Structured data exists and which types it declares.
What it cannot prove
That any engine uses it in answers.

Schema relevance

What we observe
We compare declared types against place and travel types.
What it proves
Whether types relevant to your entity are present.
What it cannot prove
That relevant types improve any specific answer.
Content clarity5 checks

Heading structure

What we observe
We count H1 and H2 headings in your homepage HTML.
What it proves
The heading outline as served.
What it cannot prove
Whether the writing under those headings is good.

Content depth

What we observe
We measure the visible copy on your homepage as it reads before JavaScript runs, with code and styling removed.
What it proves
How much readable text is available in the initial HTML before client-side rendering.
What it cannot prove
Content quality, or what appears after JavaScript runs.

Readability

What we observe
We measure sentence length of that same served copy.
What it proves
Average sentence length of the served copy.
What it cannot prove
Whether the writing is clear to a human reader.

Image alt text

What we observe
We check alt text and equivalent accessible names on homepage images and other image media: image inputs, image-map areas, SVGs marked as images, video posters, and embedded image files. An empty alt counts as an intentional choice for decorative images. Hidden or presentation-only assets are skipped.
What it proves
Presence of alt text or an equivalent accessible name on those homepage assets.
What it cannot prove
Whether the descriptions are accurate, images that load only after JavaScript runs, or sitewide coverage beyond the homepage.

Internal links

What we observe
We count links from your homepage to your own site.
What it proves
How many same-site paths the homepage offers.
What it cannot prove
Your full site architecture.
Performance2 checks

Mobile performance

What we observe
We request Google Lighthouse lab data through the PageSpeed Insights API.
What it proves
Google's lab measurement at audit time.
What it cannot prove
Real visitor experience. When Google returns no data, the check reads unknown and is not scored.

Desktop performance

What we observe
We request Google Lighthouse lab data through the PageSpeed Insights API.
What it proves
Google's lab measurement at audit time.
What it cannot prove
Real visitor experience. When Google returns no data, the check reads unknown and is not scored.

Provider statuses on the crawler matrix read your published robots.txt rules. Allowed means your published rules do not disallow that crawler. It is a policy observation, not a visit log.

A score does not decide the fix.

When the evidence supports it, Drifter turns a gap into a reviewable Action with a draft in your Website Voice, and your team decides what publishes.

See the review-first product workflowStart with the AI visibility guide

Your free AI Snapshot

What does AI tell travelers about your place?

  • Answer Share
  • Authority Share
  • AI Readiness
  • Sentiment themes
  • Peer validation
  • Prioritized fixes
Get your free AI SnapshotBook a demo

Destination teams usually start with a demo.