21 November 2024

I lost 30% search traffic to AI… What do I do?

2Frameworks

7Insights

Frameworks in this episode

Marketing6 steps

AI Search Visibility Flywheel

Publish authoritative content that earns mentions across AI-generated answers

Read →

Strategy6 steps

Search Disruption Response Plan

Protect current search returns while building demand beyond search and paid ads

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 2

Hot Take09:30

AI Source Links May Build Trust Without Sending Meaningful Traffic

The hosts argue that users will usually ask follow-up questions instead of opening citations and conducting their own research. Source links reassure users that the system did the work, but the resulting referral traffic is unlikely to replace the organic visits publishers lose.

Follow-up prompts can validate claims without leaving the assistant
Citations increase confidence in an AI-generated response
Users may click only one source across many searches
AI referrals of roughly 0.5% to 1% do not offset reported traffic losses

“I think the sources are there to make you feel confident in the answer”
Kieran · 12:30

“traffic from chat GPT and perplexity is like half a percent to a percent”
12:00

#citations#referral traffic#publishers#ai assistants

Hot Take22:00

AI Search Depends on Content While Weakening the Reason to Publish It

AI systems need fresh informational content, but answer engines can remove the visits and revenue that motivate smaller publishers to create it. Synthetic data and publisher deals are possible responses, yet the hosts see no clear economic model for the long tail of independent websites.

AI models depend on informational content from the public internet
Reduced organic traffic weakens advertising and affiliate incentives
Synthetic data may reduce dependence on human-created material
Publisher partnerships do not automatically solve incentives for small sites

“they're dependent upon informational content continuing to be published but the but the reasons the publish informational content are going to go down over time”
Kieran · 22:00

“for the little sites like the smaller sites that are really just doing it because are dependent upon cuz they make money from organic traffic…”
Kieran · 22:30

#publishers#synthetic data#content economics#ai training

Explainer· 3

Explainer01:30

Why Informational Search Traffic Is Falling by Up to 30%

The hosts report that search-heavy organizations are seeing the greatest declines on informational pages such as blog posts and resources. AI overviews and assistants answer straightforward questions directly, while transactional product and solution pages appear less affected.

Reported declines range from 10% to 30% for many search-focused organizations
Informational queries face more cannibalization than transactional queries
Google AI overviews reduce the need to click through to source pages
Expansion into additional languages could broaden the impact

“most people I talk to have seen somewhere between a 10% to 30% decline in their search traffic”
02:00

“all of these informational blog posts and informational content are starting to be cannibalized by the AI assistance”
Kieran · 02:30

#seo#search traffic#ai overviews#content

Explainer06:00

AI Search Turns Dozens of Sources Into One Tailored Answer

A detailed HubSpot competitive-analysis prompt illustrates how AI search handles questions that would be cumbersome in a conventional search engine. The assistant searches many pages, synthesizes their contents, and returns a formatted response instead of requiring the user to inspect links individually.

Conversational search supports long, multi-part research requests
One response can synthesize roughly 20 or more web pages
AI performs collection, comparison, formatting, and initial analysis
Underlying links increasingly function as a data repository rather than the main interface

“this is a very complex query with a very detailed and complex result”
07:30

“20 different web pages were used to give you this one search response”
08:00

#ai search#research#chatgpt#search experience

Explainer14:00

The Next SEO Surface Is the Prompt AI Suggests Before You Ask

AI search interfaces are beginning to suggest queries and supply prepared prompts. The hosts expect assistants eventually to formulate better questions on the user's behalf, making inclusion in recommended prompts a potential successor to conventional keyword ranking.

AI search products are adopting familiar search-engine interface patterns
Suggested queries expose common consumer interests
Prepared prompts can improve how users formulate requests
Brands may compete for inclusion in the answers to recommended prompts

“it will figure out how to ask itself the question better than you can ask it”
Kieran · 17:30

“it used to be like oh do I rank for this keyword it's now like I know that users of search GPT are going to…”
18:00

#prompts#seo#search interfaces#recommendations

Takeaway· 2

Takeaway18:30

SEO Is Becoming a Maintenance Game, Not a Dead Channel

Search disruption is expected to unfold gradually, leaving substantial traffic available even after meaningful declines. Mature websites may shift from maximizing net-new organic growth toward protecting existing returns and allocating resources according to profitability.

A 20% decline still leaves 80% of traffic available to compete for
Large sites may feel disruption before the average site
Search should continue while its returns justify the investment
Resource allocation should reflect declining marginal efficiency

“a lot of websites go into maintenance mode”
Kieran · 19:00

“you should keep doing it until whatever degree it's like profitable for you”
Kieran · 19:30

#seo#profitability#maintenance#channel strategy

Takeaway25:30

Use AI Search Yourself Before Rewriting Your Marketing Strategy

The closing advice is to make AI search part of one's own research behavior rather than relying solely on secondhand commentary. Direct experience reveals how prompts, sources, recommendations, and follow-up interactions are changing customer discovery.

Regularly use ChatGPT search, Perplexity, and similar tools
Compare AI workflows with conventional Google searches
Observe how brands and products appear in recommendations
Use firsthand behavior to anticipate effects on customers and the business

“force yourself to use AI search perplexity search GPT all these things that you understand that how this is all changing”
25:30

“you do have to experience it for yourself”
25:30

#experimentation#ai search#marketing skills#customer behavior