BASELINE
DankRank Launch AI Visibility Baseline
Most case studies begin after something has worked. With a new domain, we have the chance to document what was visible before there was a result to promote.
This baseline records where DankRank stood at launch across AI citations, organic search visibility, referral traffic and the other signals we can measure consistently over time.
Some of those starting numbers may be zero. We’ll record them exactly as they are. If ChatGPT isn’t citing the site, the count stays at zero. If DankRank has no meaningful non-brand visibility or AI referral traffic, that belongs in the baseline too.
Later, when something changes, we’ll have an honest starting point to compare it against.
- Your footprint and 2 competitors, recorded
- Anything after this, observed later
01
Baseline Status
Day 0 was recorded on September 10, 2026. At launch, DankRank had no observed citations, mentions or recommendations in either of the two AI visibility panels below. That zero is the starting point for the scheduled follow-up observations.
Batched Longitudinal AI Panel
The governed 24-prompt set was observed separately in each platform:
ChatGPT: 0 of 24 prompts with a DankRank citation, mention or recommendation.
Claude: 0 of 24.
Gemini: 0 of 24.
Perplexity: 0 of 24.
That produces 96 platform-prompt observations across the four platforms. We do not combine those observations into a proprietary or blended AI visibility score.
Separate Isolated Commercial Validation
Six commercial prompts were also tested individually in fresh chats as a separate validation method:
ChatGPT: 0 of 6.
Claude: 0 of 6.
Gemini: 0 of 6.
Perplexity: 0 of 6.
This second panel contains 24 platform-prompt observations. It is kept separate from the batched longitudinal panel because the testing method is different.
Perplexity required a signed-in environment. Several fresh-chat runs pulled in existing account context, including a different agency, cannabis and Canada. Those Perplexity observations are therefore environment-limited rather than perfectly clean-context observations. DankRank itself was not named or primed and still did not surface.
Google Search Console on Day 0
The captured Search Results evidence showed 0 clicks, 0 impressions and no query or page rows in the last-24-hours view.
The submitted sitemap had been processed successfully and reported 74 discovered URLs. Sitemap discovery is not proof of indexation. The Page Indexing report was still processing, so the Day 0 indexed and not-indexed totals were not yet available. We do not turn unavailable indexation data into a zero.
GA4 on Day 0
GA4 recorded 9 sessions: 4 Unassigned, 3 Direct and 2 Referral. It recorded 0 engaged sessions and 0 key events.
The captured evidence did not identify the referral source, so those 2 Referral sessions are not classified as AI referral traffic.
Crawl and Discovery Configuration
Production verification showed 74 of 74 governed routes returning 200 and 5 of 5 utility routes returning 200. All 74 governed routes had self-referential canonicals and index, follow directives, and the sitemap contained the same 74 governed URLs.
Googlebot, GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and bingbot all received 200 responses. llms.txt and Markdown-for-Bots also returned 200.
That evidence establishes crawlability and discovery configuration at launch. It does not establish Google indexation, an AI citation, or that any AI system had already discovered and understood the site.
How to Read the Zero
The Day 0 result is a launch baseline. It is not a failure, it does not establish why DankRank was absent, and it is not a prediction of what happens next.
The same governed measurements will be repeated at the scheduled checkpoints so later movement can be compared with the starting state without rewriting the baseline after the fact.
02
What We Mean by an AI Visibility Score
An AI visibility score can help us compare the same prompt set over time. That’s the extent of the claim.
It isn’t a universal measure of whether a brand is winning in AI search. Change the prompts, engines, timing or scoring rules and the number can change with them.
When we publish a score, we’ll also show how it was calculated and the evidence behind it. That includes the citations, mentions, recommendations, sources and competitor appearances we observed.
The number is a shortcut for comparison. Anyone reading the page should still be able to inspect what produced it.
03
What the Launch Baseline Will Record
AI Citations
We’ll record every instance where a DankRank page appears as a cited source for one of the prompts in the test set. The record will include the engine, prompt, cited URL and observation date.
A brand mention without a linked or identifiable source won’t be counted as a citation.
Brand Mentions and Recommendations
Being named in an answer isn’t the same as being recommended.
We’ll track both, but keep them separate. Recommendations will only count when the prompt gives the AI system a meaningful reason to suggest an agency, service or source.
Source Visibility
We also want to see which DankRank pages appear among the sources displayed with an answer, even when DankRank isn’t the final recommendation.
That may show us which types of pages appear first and whether they continue showing up in later observations.
Organic Visibility
The organic baseline will include indexed pages, non-brand impressions, ranking queries and the Search Console data available after launch.
A new domain won’t produce a clean trend immediately. We’ll keep the early noise in the record and wait until there is enough data before drawing conclusions from it.
AI Referral Traffic
When analytics identifies a visit from ChatGPT, Gemini, Perplexity, Copilot, Claude or another AI platform, we’ll record it separately from citation and mention data.
Visibility doesn’t guarantee a click, and one referral doesn’t prove that the site has broad AI visibility. The measurements answer different questions.
Competitor Context
DankRank’s numbers mean more when we can see what happened around them.
We’ll use the same defined prompt set to observe a relevant group of competitors. That gives us a better chance of distinguishing a change specific to DankRank from a broader shift in how the results were generated.
Baseline
- AI Citations
- Brand Mentions and Recommendations
- Source Visibility
- Organic Visibility
- AI Referral Traffic
- Competitor Context
04
The Prompt Set Stays Documented
Ask a different question and you may get a different answer. Any AI visibility result is therefore tied to the prompts used to produce it.
We’ll publish the prompt categories, search intent, engines and testing schedule, then keep a stable core set for comparison. We won’t swap out an inconvenient prompt or quietly change the test because the result wasn’t what we hoped to see.
The set can grow as new buyer questions become relevant. When we add something, it will be dated and identified as a new prompt rather than folded into the earlier results.
05
We Expect the Early Data to Be Messy
A new domain doesn’t have enough history to produce a clean story at launch. Search engines still need to crawl and process the site, referral traffic may be too small to interpret and the same AI prompt can produce different answers from one test to the next.
We’ll record that instability rather than edit around it. Some measurements may settle quickly. Others may need several observation periods before they tell us anything useful.
This is why the starting point matters. If we begin changing pages before documenting it, we lose the ability to separate an actual improvement from normal early movement.
06
What This Baseline Does Not Measure
This page documents DankRank’s starting point. It isn’t an industry benchmark, and we won’t use one observed change to claim that a particular optimization caused it.
It also isn’t the audit we prepare for prospective clients. If you want to understand how your own business currently appears across Google and AI answers, start with the free Visibility Baseline.
07
How This Record Will Be Updated
The original launch baseline will remain on the page. Later observations will be added with their dates so nobody has to guess what changed or when.
If we revise the methodology, we’ll explain the change rather than compare two different tests as though they were identical. Missing data won’t be estimated, and a measured zero will remain zero.
Without that history, the baseline would lose most of its value.
08
See the Rest of the Evidence
This baseline documents where DankRank starts, but it is only one part of the public record.
The build record explains how the site was made and what changed along the way.
Experiment 001 follows what happens when a new domain tries to earn brand citations in ChatGPT.
Our Research work asks broader questions using more prompts, brands and search environments than a single DankRank test can support.