Is Traditional SEO Dead? My 2025 Experiment Results With AIO vs SEO Traffic

Traditional SEO is not dead, but it has fundamentally changed. In 2025, I ran a year-long experiment comparing organic traffic from Google’s traditional search results against traffic generated through AI Overviews and ChatGPT citations. The results revealed a search landscape in the middle of its most significant shift since the dawn of web search — and most website owners are not prepared for what comes next.

Here’s what I found after tracking traffic sources, click-through behavior, and content performance across multiple sites throughout 2025.

The State of Search in 2025: What the Data Shows

Before diving into experiment specifics, the context matters. According to Semrush’s AI Overviews study analyzing 2025 search data, Google now displays AI Overviews for a growing percentage of queries, and those overviews are reshaping how users interact with search results. Meanwhile, ChatGPT has surpassed hundreds of millions of monthly active users, becoming a legitimate search alternative for informational queries.

AdExchanger reported in late 2025 that the “AI Search Reckoning” is actively dismantling open web traffic, with publishers seeing measurable declines in referral clicks from AI-powered search surfaces. This is not theoretical — it is happening across industries right now.

Traditional Search Traffic Trends

For most sites I tracked, traditional Google organic traffic remained relatively stable in terms of raw impressions but declined in click-through rate. The reason is straightforward: AI Overviews are absorbing clicks that previously went to organic listings.

Semrush data confirmed this pattern. Queries where AI Overviews appear show a measurable drop in organic click-through rates, particularly for informational and how-to queries — the exact content types most site owners rely on for traffic.

ChatGPT as a Traffic Source

ChatGPT referral traffic is real but behaves differently from traditional search. According to Semrush’s clickstream analysis spanning 17 months of data, ChatGPT sends significantly less traffic than Google in absolute volume. However, the traffic it does send tends to have higher engagement metrics — longer time on site, more pages per session, and stronger conversion intent.

The catch? You cannot directly control when or how ChatGPT references your content. There is no equivalent of Google Search Console for ChatGPT optimization yet, which makes measurement and strategy adjustment far more challenging.

My 2025 Experiment: Methodology and Setup

I designed this experiment to answer one question: if I invested equal effort in optimizing content for traditional SEO versus being cited in AI-generated answers, which would deliver more meaningful traffic?

Content Properties Tested

The experiment ran across three separate content properties:

• A SaaS blog covering project management topics with an existing organic traffic base of approximately 40,000 monthly sessions
• A niche affiliate site in the home improvement space with about 15,000 monthly organic sessions
• A new authority site in the personal finance space starting from zero

For each property, I split content production into two tracks. Track A focused exclusively on traditional SEO fundamentals — keyword research, on-page optimization, technical SEO, and link building. Track B focused on what the industry now calls AIO (AI Overview Optimization) or GEO (Generative Engine Optimization) — structuring content to be cited in AI-generated answers.

What Track A (Traditional SEO) Involved

• Targeted keyword research using volume and difficulty metrics
• Optimized title tags, meta descriptions, and heading structures
• Comprehensive, long-form content designed to outrank competitors
• Internal linking campaigns and structured data implementation
• Active link building through outreach and guest contributions

What Track B (AIO/GEO) Involved

• Formatting content with clear, concise answers to common questions
• Using structured markup, definitions, and factual statements that AI models could easily extract
• Building topical authority through comprehensive coverage of subject areas
• Optimizing for citation-worthy claims backed by original data or unique expertise
• Focusing on E-E-A-T signals — experience, expertise, authoritativeness, and trustworthiness

Experiment Results: What Actually Happened

After twelve months of consistent execution, the results told a nuanced story.

Volume: Traditional SEO Still Wins

Traditional SEO drove significantly more raw traffic across all three properties. On the established SaaS blog, Track A content generated roughly 3.2 times more sessions than Track B content. On the affiliate site, the ratio was about 2.5 to 1.

This makes sense. Google still processes billions of searches daily, and traditional organic results still appear for the vast majority of queries. AI Overviews do not replace the entire results page — they supplement it, and often users still click through for deeper information.

Quality: AI-Referred Traffic Outperformed

Here is where things get interesting. Visitors who arrived via ChatGPT or through clicking AI Overview references showed markedly better engagement across the board:

• Time on page: AI-referred visitors averaged 4 minutes 12 seconds versus 2 minutes 47 seconds for traditional organic visitors
• Pages per session: AI-referred visitors viewed 2.3 pages on average versus 1.6 pages for organic
• Conversion rate: AI-referred visitors converted at 1.8 times the rate of traditional organic visitors on the SaaS blog
• Bounce rate: AI-referred visitors had a 34% lower bounce rate

The pattern was consistent: AI-referred traffic is lower volume but higher quality. These users arrived with a specific question already partially answered, so they were further along in their decision-making process.

The New Authority Site: A Surprise Outcome

The most striking results came from the brand-new personal finance site. Track A content (traditional SEO) took months to gain any traction, as expected with a new domain. But Track B content — articles specifically structured for AI citation — began appearing in ChatGPT responses within weeks.

Not all of those citations translated to clicks. But for a site with zero domain authority, the fact that AI systems surfaced the content at all represented a meaningful discovery channel that traditional SEO could not match in the early stages.

Why AI Overviews Are the Real Disruption

While ChatGPT gets the headlines, the bigger near-term threat to traditional SEO traffic is Google’s own AI Overviews. The Semrush study found that AI Overviews are now appearing across a broad range of query types, and the click behavior they produce is fundamentally different.

When an AI Overview provides a complete answer at the top of the page, many users never scroll down to the traditional organic results. This creates what publishers call “zero-click” scenarios — the user gets their answer without visiting any website.

Who Is Most Affected

Not all content types are impacted equally. The sites and queries most vulnerable to AI Overview traffic loss include:

• Simple factual queries — “What is the capital of France?” or “How many ounces in a cup”
• How-to queries with straightforward answers — step-by-step processes that AI can summarize
• Comparison queries — “X vs Y” where AI can quickly list pros and cons
• Definition queries — “What is project management?” or similar

Queries where AI Overviews cause the least damage are those requiring personal experience, original data, deep expertise, or local context. This is not accidental — it reflects where AI systems still struggle to provide complete answers.

Practical Lessons: What Worked and What Did Not

Based on a full year of testing, here are the strategies that delivered results in both traditional SEO and AI-referred traffic.

Strategies That Worked for Both Traffic Types

• Topical authority: Sites that comprehensively covered a subject area performed better in both Google rankings and AI citations. Shallow, scattered content performed poorly across the board.
• Original data and experience: Content that included first-party research, original case studies, or genuine expertise was cited more often by AI systems AND ranked better in Google. This is not a coincidence — both systems reward depth and authenticity.
• Clean site architecture: Fast, well-structured sites with strong internal linking outperformed on every metric. Technical SEO remains foundational regardless of how traffic arrives.
• Structured, scannable content: Articles with clear headings, concise summaries, and organized data performed well for both user experience and AI extraction.

Strategies That Failed or Underperformed

• Keyword-stuffed content: Content written primarily to hit keyword density targets performed worse in AI citations. AI systems extract meaning, not keyword patterns.
• Skyscraper approach without substance: Simply writing longer versions of competitor content did not earn AI citations. AI systems tend to cite content that adds something new — a unique perspective, original data, or expert insight.
• Ignoring traditional SEO for AI: The Track B content that ignored traditional SEO fundamentals still underperformed in total traffic. You cannot neglect the largest search channel while chasing AI referrals.
• Over-reliance on link building: Heavy link building campaigns showed diminishing returns for both traditional rankings and AI visibility compared to content quality investments.

The Verdict: Is Traditional SEO Dead?

No. Traditional SEO is not dead. But traditional SEO as the sole traffic strategy is increasingly risky.

The data from my experiment, combined with the broader market trends from Semrush, AdExquirer, and ChatGPT usage data, paints a clear picture. The search ecosystem is bifurcating. You need to be visible in traditional search results because that is where the volume lives. But you also need to optimize for AI citation because that is where the highest-quality traffic comes from.

The sites that will thrive are those that treat traditional SEO and AI optimization as complementary strategies rather than competing ones. The core of both is the same: create genuinely valuable, expert content that answers real questions better than anyone else.

What I Am Doing Differently in 2026

Based on these results, I have shifted my content strategy in several ways:

• I allocate approximately 70% of content effort to traditional SEO with AI-friendly formatting, and 30% to content specifically designed for AI citation and zero-click visibility
• I prioritize original research and first-party data in every major content piece
• I track AI-referred traffic as a separate channel in analytics, distinct from organic search
• I invest more in site architecture and page experience, as these factors influence both traffic types
• I focus less on link building volume and more on earning citations from authoritative sources that AI systems reference

For a deeper look at optimizing site structure for search visibility, see our guide on programmatic SEO strategies.

Conclusion

The question “Is traditional SEO dead?” has a clear answer: it is alive but evolving rapidly. My 2025 experiment confirmed that traditional organic search still drives the majority of web traffic. However, AI-powered search surfaces — both Google’s AI Overviews and ChatGPT — are capturing an increasing share of user attention and delivering higher-quality visitors to the sites they cite.

The winning approach is not to abandon traditional SEO in favor of AI optimization or vice versa. It is to build a content strategy that serves both channels simultaneously. Create authoritative, well-structured, experience-rich content that Google wants to rank and AI systems want to cite. That is the strategy that held up across every property and every scenario I tested throughout 2025.

Search is not dying. How people search is changing. The publishers and marketers who adapt to both traditional and AI-driven discovery will capture the traffic — and the revenue — that comes with it.

FAQ

Can I get free traffic from ChatGPT without paying for ads?

Yes, but it works differently from traditional SEO. ChatGPT may reference or cite your content in its responses if your site contains authoritative, well-structured information on a topic the user asks about. You cannot directly submit URLs to ChatGPT for inclusion. Instead, focus on creating expert content that AI systems naturally gravitate toward — original research, clear factual statements, and comprehensive topic coverage tend to get cited most often.

What is AIO and how does it differ from traditional SEO?

AIO (AI Overview Optimization) refers to structuring your content so it gets cited or featured in AI-generated search answers, such as Google’s AI Overviews or ChatGPT responses. Traditional SEO focuses on ranking in the standard list of blue links. AIO emphasizes concise, extractable answers, strong E-E-A-T signals, and original data — elements that AI systems use to select sources for their generated responses.

How much traffic does ChatGPT actually send to websites?

According to Semrush clickstream analysis from 2025, ChatGPT sends considerably less referral traffic than Google in absolute terms. However, the traffic it does send tends to be more engaged — visitors spend more time on site, view more pages, and convert at higher rates. The volume is growing as ChatGPT’s user base expands, but it remains a supplementary channel rather than a replacement for traditional search traffic.

Should I stop doing traditional SEO and focus only on AI optimization?

No. My experiment showed that abandoning traditional SEO fundamentals in favor of pure AI optimization resulted in lower total traffic. Traditional search still drives the majority of web visits. The best approach is to maintain strong traditional SEO practices while also formatting content for AI citation. Think of them as two surfaces to optimize for, not a choice between one or the other.

What types of content are most affected by Google AI Overviews?

Queries with simple factual answers, straightforward how-to processes, direct comparisons, and definitions are most impacted. When an AI Overview provides a complete answer, users often do not click through to any website. Content requiring personal experience, original data, local context, or deep expertise is less affected because AI systems cannot easily replicate those qualities.