ChatGPT Visibility Drives Purchases: What the Data Means for GEO
A new behavioral study from Profound, Kevin Indig, and Eric Van Buskirk measured something that's been assumed but never clearly demonstrated: whether being cited by ChatGPT actually influences what people buy.
The answer is yes — and the correlation is stronger than most marketers expected.
The study observed 56 people comparing products inside ChatGPT across 221 real purchase tasks. Brands that participants chose appeared in ChatGPT's answers approximately twice as often as brands they passed over — 24% share of voice for chosen brands versus 11% for the ones not selected.[1] The correlation between AI visibility and purchase choice was 0.57.[2]
The secondary finding is equally significant: 92.8% of tasks ended without a click to the open web.[3] When ChatGPT recommended a brand, participants acted on that recommendation directly — they didn't leave to verify it on the open web first.
This data has direct implications for how to think about GEO investment and strategy.
What a 0.57 Correlation Actually Means
A 0.57 Pearson correlation between AI visibility and purchase choice is a strong signal. For context: a correlation of 0.3 is considered moderate in behavioral research; 0.5+ is strong. The finding isn't that visibility guarantees selection — it's that brands appearing more frequently in AI answers are substantially more likely to be the brand someone chooses.
The researchers controlled for the study's limitations (56 participants, 221 tasks, U.S.-skewed sample), but the directional signal is clear: if your brand isn't appearing when people ask ChatGPT about your category, you're missing the moment when purchase intent is highest.
The No-Click Implication Changes the SEO Math
Traditional SEO drove traffic: you ranked, people clicked, they arrived on your site. The 92.8% no-click finding means AI search is fundamentally different. People are making decisions inside ChatGPT without visiting your site. Your content needs to do its persuasive work within the AI answer itself — in the excerpt ChatGPT cites, the brand name it mentions, and the credibility signals that make it recommend you over a competitor.
This changes what GEO optimization is actually for. It's not primarily about driving traffic from AI search — it's about being the brand that gets named when your category is asked about.
That means:
-
Being cited matters more than driving clicks from citations
-
The content that gets extracted needs to be persuasive, not just informative — the 75-150 words ChatGPT quotes in a citation need to make your brand sound like the credible, recommended choice
-
Citation frequency across a category's prompts is the leading indicator of brand presence, not click-through rate
The Prompts That Drive Purchase Decisions
The Profound study focused specifically on product comparison prompts — queries where someone is actively evaluating options and asking ChatGPT to help them decide. These are the highest-intent prompts in any category, and they're where the visibility-to-choice correlation is strongest.
For GEO strategy, this identifies a content priority: product comparison and "best of" content in your category is where citation presence converts most directly to consideration and selection. If your category has prompts like "best [tool] for [use case]" or "which [product] is right for [situation]" — those are the prompts where appearing in ChatGPT's answer has the highest downstream impact.
The practical implication: don't only optimize for definitional and how-to content. Optimize for comparison and recommendation prompts in your category.
What This Means for Schema and Content Implementation
The behavioral data reinforces the technical GEO practices that improve citation probability:
FAQPage schema with comparison-specific Q&As. If the high-value prompts are comparison questions, your FAQPage schema should include those comparisons as Q&A pairs. "What's the best [tool] for [use case]?" should appear as an explicit FAQ with a clear, confident answer that names your product.
Answer-first format in the exact words ChatGPT will quote. If 92.8% of tasks end without a web click, the excerpt ChatGPT quotes is often the only thing about your brand the buyer reads. That excerpt needs to do the work. Write your key claim as if it's the only sentence about your product the prospect will see.
Entity consistency across your site. The Profound study shows that brands appearing more often for a category's prompts are chosen more often. Citation frequency is a function of how thoroughly your brand is associated with the relevant entity graph — the schema connections between your organization, your product category, and the specific use cases you serve.
Volume matters. A 24% vs 11% share of voice difference drives a 0.57 correlation. Share of voice in ChatGPT is a function of citation frequency across the prompts in your category. Each additional piece of answer-first content on a priority prompt cluster is a citation opportunity — and the data suggests those citations compound into selection.
Frequently Asked Questions
Does ChatGPT visibility affect purchase decisions? Yes — a behavioral study published in July 2026 by Profound, Kevin Indig, and Eric Van Buskirk[4] found a 0.57 correlation between ChatGPT AI visibility (share of voice in ChatGPT answers) and purchase selection across 221 real product comparison tasks. Brands chosen by participants appeared in ChatGPT's answers approximately twice as often as brands passed over (24% share of voice vs 11%). The study found that 92.8% of tasks ended without a click to the open web, meaning ChatGPT's recommendation often directly influenced the decision without further research.
What is ChatGPT share of voice for brands? ChatGPT share of voice measures how often your brand appears in ChatGPT's answers for a set of prompts within your product category, relative to all brands that appear. A brand with 24% share of voice appears in roughly one in four ChatGPT responses for that category's prompts. Higher share of voice correlates strongly with being chosen by buyers — the Profound study found a 0.57 correlation between share of voice and purchase selection.
How do I increase my brand's visibility in ChatGPT? The highest-leverage actions for increasing ChatGPT visibility are: (1) verify GPTBot can access your site with no robots.txt blocks; (2) add FAQPage schema in @graph format to your most important pages; (3) rewrite content to answer-first format, especially for comparison and recommendation prompts in your category; (4) publish content specifically targeting the "best [product] for [use case]" prompts in your category, which is where the visibility-to-purchase correlation is strongest; (5) track citation frequency across your priority prompt set to measure improvement.
Why do most ChatGPT users not click through to websites? The Profound/Indig/Van Buskirk study found that 92.8% of purchase comparison tasks inside ChatGPT ended without a click to the open web. This reflects how ChatGPT presents information: it synthesizes a recommendation with enough supporting detail that many users feel confident acting on the AI's recommendation directly, without verifying on the brand's website. This is fundamentally different from traditional search, where clicking through to the source is the expected behavior.
What types of content get cited most in ChatGPT product comparisons? Content that directly addresses comparison and recommendation prompts — answering "which product is best for [use case]" with a clear, confident response — performs best in the prompts where the visibility-to-purchase correlation is strongest. FAQPage schema with comparison-specific question-answer pairs helps structure this content for extraction. Answer-first format (leading with the direct recommendation rather than context) increases the probability of being cited and ensures the excerpt is persuasive on its own.