ChatGPT Is Not Replacing Google—It’s Expanding Search [Study]

As ChatGPT and different generative AI devices proceed to accelerate successful popularity, 1 mobility looms ample for marketers: Is ChatGPT replacing Google Search? 

To research this question, we analyzed really group interacted pinch Google Search aft their first ChatGPT interaction. 

Our findings revealed an absorbing pattern: ChatGPT take isn’t lowering Google usage—it’s expanding the measurement group search.

Two Hypotheses: Substitution vs. Expansion

The study was planned to corroborate 1 of 2 theories that could explicate what’s happening now:

• Substitution Hypothesis: ChatGPT take pulls users distant from Google Search. People “substitute” ChatGPT for really they antecedently utilized Google. 
• Expansion Hypothesis: ChatGPT take does not trim Google usage. In different words, it simply expands wide information-seeking behavior.

Methodology 

We analyzed 260 cardinal rows of clickstream data betwixt January 2024 and June 2025 to study the Google Search behaviour of usersbefore and aft their first relationship pinch ChatGPT. 

Note: This clickstream information group is based connected an anonymized sample of unsocial users (panelists) who opted successful to stock their browsing behavior. This method lets america understand online trends astatine scale—while keeping individual privateness afloat protected.

To understand if (and really much) Google Search activity changed aft group started utilizing ChatGPT, we monitored the number of hunt sessions from each personification 3 months earlier and aft their first ChatGPT session.

• Study group: U.S. users who began utilizing ChatGPT successful Q1 2025, pinch nary anterior ChatGPT activity successful 2024.
• Control group: Users who ne'er utilized ChatGPT during 2024–2025. This was our baseline for knowing really Google usage mightiness alteration people complete clip (without ChatGPT adoption).
• Platform & location: Desktop devices successful the U.S. only, allowing for accordant convention tracking.
• What was measured: Google Search sessions 90 days pre- and post-ChatGPT adoption.

This attack helped america measure whether ChatGPT take leads to a alteration successful accepted hunt behavior, aliases whether it corresponds pinch unchangeable aliases moreover increasing hunt activity.

Total Search Behavior Expands pinch ChatGPT Usage

Based connected the data, ChatGPT take did not trim Google Search usage. 

This supports the description hypotheses, meaning group did not substitute their emblematic Googling pinch ChatGPT-ing.

In fact, location was moreover a flimsy summation successful mean Google Search usage aft ChatGPT adoption.

Google Search Holds Steady 

The graphs beneath exemplify each cohort from the study: those who adopted ChatGPT successful January, February, aliases March of 2025.

The reddish spike indicates the time of the first ChatGPT session, and the statement indicates regular sessions afterwards.

The bluish lines correspond sessions connected Google (before and aft the day of ChatGPT adoption). 

January 2025 Cohort

February 2025 Cohort

March 2025 Cohort

This shows no statistically important change successful regular Google Search sessions erstwhile group commencement utilizing ChatGPT for the first time. 

This shape of dependable Google Search usage besides held for:

• New ChatGPT users (Q1 2025)
• Long-term ChatGPT users (from January 2024, tracked complete 500+ days)
• Users who ne'er utilized ChatGPT (control group)

Even contempt sustained ChatGPT usage aft adoption, Google Search usage remained constant. 

January 2024 Cohort

These conclusions were besides existent for a longer cohort: those who started utilizing ChatGPT successful January 2024.

For this cohort, we tracked their behaviour for 500 days and saw a akin shape of sustained Google usage, on pinch unchangeable ChatGPT usage aft adoption.

And to corroborate that location was nary earthy variety successful Google Search behaviour complete the aforesaid clip period, we looked astatine the power group (people who had nary ChatGPT sessions successful 2024 aliases 2025).

This confirms that Google Search behaviour besides remained unchangeable for those who ne'er utilized ChatGPT. 

What This Means for Marketing Leaders

It’s still early, but this study supports the Expansion Hypothesis: ChatGPT is not replacing accepted search. But it’s expanding the measurement group activity information.

For CMOs, the implications are applicable and immediate:

• Google is not getting overtaken by ChatGPT. You tin put successful some accepted SEO and GEO / AI SEO without fearfulness of 1 cannibalizing the other.
• Customer journeys are becoming multi-modal. Users judge some ChatGPT and Google, apt switching betwixt them depending connected intent and context.
• Generative AI adds caller existent estate. It’s not replacing aged hunt tools—it’s expanding online exploration and creating much ways to scope group pinch information.

So, What Should CMOs Do Next?

1. Keep investing successful SEO. Google Search remains a pillar of integer behavior, contempt ChatGPT and AI Overviews.
2. Add AI SEO to your contented strategy. Marketing efforts for ChatGPT and different AI engines should complement, not substitute, your accepted SEO efforts.
3. Monitor transmission operation quarterly. Benchmark AI take successful your circumstantial market. Different audiences adopt ChatGPT and AI engines astatine different rates. It whitethorn beryllium much important successful your industry. 

The early isn’t zero-sum. Search isn’t dying, and ChatGPT isn’t replacing Google. 

They’re coexisting and giving marketers much devices to link pinch customers.