ChatGPT access is changing search demand, not just referral paths
Research covered by SEJ compares ChatGPT and Google clickstream behavior. Wider ChatGPT Search access was associated with fewer traditional searches, while ChatGPT sent outbound traffic far less often than Google.
Editor’s Note by Dageno: Dageno AI’s editorial team structurally distilled Search Engine Journal’s article “ChatGPT Access Tied To 9% Drop In Traditional Search” to help GEO teams understand referral-rate differences, query displacement, affected categories and measurement limits.
Understand the article in 60 seconds
- The research analyzed U.S. desktop browsing data and compared outbound clicks from ChatGPT sessions with referrals from Google searches.
- ChatGPT directed users to outside websites in 5.2% of sessions, compared with 31.1% for Google searches.
- ChatGPT’s outbound clicks skewed toward reference, academic, developer, tool and SaaS destinations.
- Broader ChatGPT Search access was linked to a 9.4% drop in weekly traditional search queries.
- The largest reductions appeared in informational searches, while transactional and recreational searches changed less.
The search shift has two separate effects
ChatGPT sends fewer outbound clicks than Google, and broader ChatGPT Search access is associated with fewer traditional searches.
For GEO teams, those are different risks: lower referral capture and lower SERP demand before a user ever reaches Google.
ChatGPT referrals go to a different web
The paper found that ChatGPT clickouts favor reference, knowledge, academic, developer, tool and SaaS sites more than the destinations usually reached from Google search.
Brands should identify which source types AI systems prefer in their category and build evidence where those systems actually look.
Informational demand is most exposed
The strongest reductions appeared around informational content, while transactional and recreational searches were comparatively stable.
Educational and category-definition assets need AI visibility because they shape demand before buyers reach commercial pages.
Dageno view: The practical question is not only what changed, but which measurement, source, technical, trust or paid-media workflow should change because of it.
Referral rate is not the same as influence
A lower clickout rate does not mean the session has no business impact. A user can receive a recommendation, comparison or answer without visiting the cited site.
Dageno would track brand mentions, answer inclusion, cited source type, post-answer branded searches and assisted conversions.
Access timing matters for measurement
OpenAI expanded ChatGPT Search access in stages, creating useful windows for comparing behavior change.
AI product rollouts can shift search demand even when the brand’s site has not changed. Measurement should include platform-access events.
Facts to convert into AI evidence
These source facts should become reusable evidence blocks: short, specific, current and easy for AI systems to retrieve without guessing.
ChatGPT sent outbound traffic in 5.2% of sessions versus 31.1% for Google searches.
AI answers can influence without sending equivalent clicks.
Wider ChatGPT Search access was associated with a 9.4% weekly drop in traditional searches.
SERP demand can shrink before website metrics explain why.
Informational searches saw the clearest declines.
Educational assets need GEO visibility.
ChatGPT referrals favored specialized and knowledge-heavy sites.
Source quality matters more than broad reach.
How teams should respond across the AI visibility workflow
The response is cross-functional because AI visibility now depends on visual answer surfaces, social trust, crawl economics, paid search automation and no-click measurement.
| Owner | Execution focus | Signal to watch |
|---|---|---|
| SEO / GEO | Translate answer-surface changes into crawlable evidence, source strategy and measurement confidence. | AI citations, branded query lift, indexed evidence pages and answer-surface share of voice. |
| Content | Prioritize original experience, unique facts and audience-fit assets over scaled commodity production. | Information gain, recrawl frequency, engagement, backlinks and decay after launch. |
| Paid media | Align AI Search eligibility, landing-page controls and future-conversion metrics with business validation. | AI Max coverage, URL exclusions, QFC variance and CRM revenue match. |
| Trust & Safety | Treat synthetic content, spam detection and provenance labels as visibility infrastructure. | AI labels, account-level spam risk, C2PA status and high-risk topic review. |
| Analytics | Report influence beyond sessions while keeping conversions and revenue as the outcome layer. | No-click exposure, mentions, branded search, direct traffic lift and assisted conversion paths. |
Signals to monitor next
Measurement should show whether the brand is being found, cited, trusted, crawled, matched and recommended as AI search and content systems change.
- AI referral rate by platform compared with Google organic referral baselines.
- Traditional search-query changes around AI product access or feature rollouts.
- Brand mentions and citations in ChatGPT answers for informational category prompts.
- Source-type mix for AI referrals: reference, academic, developer, SaaS, news and ad-supported media.
- Post-answer branded search, direct visits and assisted conversions where clickouts are lower.
Turn AI visibility signals into execution.
Dageno AI is a complete execution platform from data monitoring to strategy, content generation and result attribution. Use it to monitor AI answer surfaces, find strategy gaps, generate content actions and connect visibility shifts to business impact.
Alexa for Shopping