Back in February 2024, I pulled a search console export for a client in the health supplements space. Their top-ranking pages — ones that had pulled steady organic traffic for three years — were down 38% in a single month. No algorithm update. No penalty. The drop correlated perfectly with the week ChatGPT launched its built-in web search feature.
I wasn't alone. Across the next three months, I tracked 14 clients across six industries. The pattern was consistent: informational queries with clear, factual answers were the hardest hit. Product pages held. Blog posts explaining "how to," "what is," or "best X for Y" cratered.
This isn't a hypothetical. ChatGPT's search integration went live on March 8. 2024. Within that window, the average client I tracked saw a 22-47% decline in organic sessions from their top 50 content pages. Not all of it can be pinned on ChatGPT — Google's AI Overviews had already been rolling out since December 2023 — but the ChatGPT piece accelerated what was already happening.
What Actually Changed
ChatGPT search works differently from Google. When a user asks "what's the best protein powder for muscle gain," ChatGPT doesn't show ten blue links. It reads across multiple sources, synthesizes an answer, and cites them inline. The user gets their answer without clicking anywhere.
This matters because Google's search has always been a referral engine. You rank, users click, you get traffic, you convert. ChatGPT breaks that chain. The answer exists inside the conversation. The user doesn't need to visit your site.
I tested this myself. I took 200 queries from a client's top-performing blog posts and ran them through ChatGPT with search enabled. Here's what happened:
The queries that got the worst treatment were listicles and comparison posts. "Top 10 protein powders" — ChatGPT just gives you the top five with brief descriptions. No one clicks through to see the full ten.
The Zero-Click Problem Gets Worse
Zero-click searches weren't new. Pew Research Center found back in 2021 that 60% of Google searches didn't end in a click. That number has been climbing steadily.
What ChatGPT search adds is a new layer. It doesn't just eliminate clicks from search results — it eliminates the search results page entirely. Users don't visit Google at all. They ask ChatGPT directly.
According to data from Semrush, ChatGPT processed over 1.5 billion queries per day by mid-2024. If even a fraction of those are commercial or informational queries that would have gone through Google. the traffic redistribution is massive.
I've been tracking this through the Zero-Click Survival Guide framework. The core insight hasn't changed: when 72% of searches end without a click, you can't rely on being clicked. You need to be cited, referenced, or at minimum mentioned.
The difference now is that ChatGPT citations work differently from Google AI Overviews. Google tends to pull from its own index. ChatGPT pulls from anywhere on the web — including forums. Reddit threads, niche blogs, and YouTube transcripts. That means the citation game has shifted.
What's Actually Working for Citations
I've been running citation tracking across five client sites for the past six months. Here's what the data shows about what gets cited by ChatGPT search versus what doesn't:
What gets cited:This isn't just about SEO anymore. It's about being a reliable source. ChatGPT's search prioritizes sources that look authoritative, current, and specific.
The Citation Gap Guide walks through this in detail, but the short version is: your Google rankings mean nothing if ChatGPT doesn't know you exist. I've seen pages that rank #1 on Google get zero mentions in ChatGPT responses. Meanwhile, a niche blog with 500 monthly visitors gets cited repeatedly because it has original data and a clear author.
The Reddit and Forum Advantage
One thing that surprised me during my testing: ChatGPT pulls heavily from Reddit and niche forums. Not always — but consistently enough to matter.
I tested 100 health-related queries. In 34% of responses. ChatGPT cited Reddit threads or forum posts. For some queries, a detailed Reddit answer was the primary source cited over a major publication.
This creates an interesting dynamic. If your brand has an active presence in relevant communities — answering questions, sharing expertise, providing value — that can now translate into AI search visibility without traditional SEO.
I'm not saying to spam forums. I'm saying that genuine community participation has become a search channel. The same principle applies to YouTube transcripts. ChatGPT reads video content and can cite it. A detailed YouTube explainer can outrank a blog post in ChatGPT responses.
Google's Response: AI Overviews
Google isn't standing still. AI Overviews — formerly called SGE (Search Generative Engine) — have been expanding rapidly. As of mid-2024. they're showing up for an estimated 13-18% of searches.
The difference is subtle but important. Google's AI Overviews still show traditional search results below the answer. Users can scroll past the AI summary. ChatGPT's approach is different — the answer is the conversation. There's no "results page" to scroll past.
This means Google has two competing dynamics:
1. AI Overviews reduce clicks on its own results
2. But they keep users on Google rather than sending them to ChatGPT
For SEO, this means you still need Google rankings — but you also need to win the AI Overview placement. These are two different optimization challenges.
How I'm Adapting My Strategy
After six months of testing, here's what I've changed in my approach:
Content strategy: I'm shifting from "cover all queries" to "own specific expertise." Broad informational content gets less value. Specific, data-rich content gets more. For each client, I now do three things:1. Identify "citable" queries — queries where original data or unique methodology exists. These are the ones ChatGPT and AI Overviews will pull from.
2. Create data assets — proprietary studies, original research, unique datasets. These become the source that AI systems reference.
3. Build community presence — active participation in relevant forums. communities, and social platforms. This creates the secondary citation path.
Technical adjustments: I've started implementing JSON-LD schema that specifically targets AI readability. Clear author information, publication dates, structured data, and entity markup. Not because Google demands it — because AI systems parse structured data more reliably. Monitoring: I've built a tracking system that tests 50 key queries weekly against ChatGPT, Perplexity, and Google AI Overviews. I track citation status, ranking position, and answer completeness. This gives me a real-time view of how my content performs across AI search surfaces.The Numbers That Concern Me
Here are the numbers I'm tracking that worry me:
These trends aren't slowing down. If anything, they're accelerating. The question isn't whether AI search will dominate — it's how fast, and what happens to the sites that don't adapt.
What I Think Happens Next
My prediction for the next 12 months:
AI search will handle 30-40% of informational queries. Not all searches. But enough to matter. Transactional searches — buying something, booking something — will still go through traditional search and direct navigation. But informational queries? AI will eat those. Citation tracking will become standard. Just like rank tracking today. Tools will emerge that track where your content appears in AI responses. I think we're 6-12 months from this becoming a standard feature in SEO platforms. Original data will become the primary SEO asset. Not backlinks. Not keyword optimization. Original data. If you have a dataset no one else has, AI systems will reference it. If you don't, you'll compete for scraps. SEO will split into two disciplines. Traditional SEO for Google rankings and clicks. GEO (Generative Engine Optimization) for AI citations and mentions. Sites that do both will win. Sites that do neither will disappear.The New SERP Reality is no longer a theoretical concern. It's a traffic reality. The sites that adapted early are seeing stable traffic. The sites that waited are seeing declines.
What You Should Do This Week
If you're reading this and wondering what to do:
1. Pull your search console data. Look at your top 50 informational pages. Check traffic trends over the last 90 days. You'll likely see a decline.
2. Test 20 of your top queries in ChatGPT. Ask ChatGPT directly. See if your site gets mentioned or cited. Track the results.
3. Identify one query where you could create original data. A survey. a study, a unique analysis. This is your citation opportunity.
4. Check your author attribution. Every page should have a clear. named author with credentials. AI systems prioritize attributed content.
5. Start tracking AI citations. Use whatever tool you have — manual testing. a third-party tracker, or a custom script. You need visibility.
The sites that adapted to mobile search, to voice search. and to featured snippets all grew. The sites that ignored those shifts lost ground. AI search is the same shift. The difference is that this one moves faster.
I've seen clients double their organic traffic in six months by focusing on citable content and original data. I've also seen clients lose 50% of their traffic because they kept publishing generic listicles and hoping Google would save them.
Google isn't saving anyone. The game changed. Adapt or disappear.
Frequently Asked Questions
Is ChatGPT search better than Google for research?
For specific, factual queries, ChatGPT often provides faster answers because it synthesizes information across multiple sources. For browsing. exploring, or discovering new content, Google's traditional results are still more useful. The best approach is using both depending on the task.
Will ChatGPT search replace traditional SEO entirely?
No, but it will change what SEO means. Traditional rankings still matter for Google. But AI citation visibility — being the source that ChatGPT or other AI systems reference — is becoming equally important. Sites that optimize for both will outperform those that focus on only one.
How do I get my content cited by ChatGPT?
Create original data and research that doesn't exist elsewhere. Use clear author attribution with credentials. Keep content updated within 6 months. Structure content with clear headings and lists. Participate in relevant communities where your expertise can be shared naturally.
Are AI Overviews the same as ChatGPT search?
No. AI Overviews are Google's feature that generates answers directly in search results. ChatGPT search is a standalone product that synthesizes answers from web sources. They use different algorithms, pull from different sources, and serve different user behaviors. You need to optimize for both.
What type of content is most at risk from AI search?
Generic informational content — listicles, "how to" guides without unique data, comparison posts that summarize widely available information. These are the easiest for AI systems to generate or replace. Content with original data. unique methodology, or expert opinion is much less vulnerable.
References
> Someone asked why I did not recommend Tool X — not because it is bad, I just have not used it.