Voices of Search Podcast: SEO Under Siege – 2025’s Collapse of Google Traffic

Host: Jordan Cooney
Guest: Malta Lanvier, CMO & CPO at Peak AI
Date: March 9, 2026

Overview

This episode dives into the rapidly evolving landscape of search and SEO, particularly focusing on the impact of AI-powered search engines and the anticipated decline in Google-driven traffic by 2025 and beyond. Jordan Cooney hosts Malta Lanvier, a veteran of both agency and in-house SEO (notably at Searchmetrics and Idealo), who now leads at Peak AI. Together, they explore actionable strategies for marketers and SEO professionals to adapt to the seismic shifts in online discovery, AI-generated content, and measurement challenges in a clickless, volume-less world.

Key Discussion Points & Insights

Malta’s Journey & the AI Epiphany (01:41–04:13)

• Malta’s Background: From teen web-builder to agency co-founder, Searchmetrics, Idealo (“largest e-commerce price comparison in the world”), and now Peak AI.
• South Korea Vacation: Malta challenged himself to use ChatGPT instead of Google for daily queries; everything worked fine ("zero food poisoning, zero allergic reactions… Everything was fine." – C, 03:29).
• Takeaway Epiphany: “This change that is happening with AI search is going to come faster and it's going to be more severe than I thought.” (C, 03:55)
• Industry Crossroads: Malta chose to actively participate in change instead of being a “victim of this very fundamental change” in content discovery and consumption.

Skillsets for the AI Search Era (05:40–08:53)

• SEOs Are Equipped: “Skill wise SEOs are probably the best equipped people in the world to work on AI search.” (C, 06:19)
• Fundamental Change:
  • Traditional focus was on clicks, measurable search volume, and keyword attribution – all transparent with Google.
  • “With AI search…there's no search volume…results are personalized…there is no click…” (C, 07:27)
  • AI results are “much, much more extreme” in terms of unpredictability and lack of direct click-throughs or clear attribution.

Lessons from Price Comparison & Legal Battles with Google (10:33–13:50)

• Google Shopping & Antitrust: Idealo’s legal battle with Google delayed but couldn’t stop the giant’s dominance.
• AI Will Reduce Clicks: “There's no reason to fight it. I think that was the biggest skill thing I learned.” (C, 12:51)
• Evolving Mindset: Shift from optimizing keywords to focusing on goals, strategies, and embracing the reality of fundamental ecosystem changes.

The AI Discovery Landscape – Who’s Who and What’s Changing (15:27–20:12)

• Current Leaderboard: Google leads (AI Overviews), followed by OpenAI/ChatGPT, Gemini, then Perplexity, Grok, and Anthropic; regional players elsewhere.
• How Prompts Differ: With LLMs, user input is longer, more natural; results are generated (not listed), often with retrieved contextual grounding from external sources.
• SEO’s New Frontier: “If people want to see success quickly, then… trying to influence this grounding process is the one thing that actually works.” (C, 19:36)

Ads in AI: What Do They Mean For Marketers? (22:09–24:46)

• Limited Tracking Initially: “Almost nothing…there will be little to no performance tracking or metrics or insights.” (C, 22:14)
• Validation of Commercial Intent: Ads will provide directional validation that AI platforms contain commercial users and topics.
• Organic remains key: “If you're willing to spend hundreds of thousands…you should also invest time on organic visibility.” (C, 23:32)

Grounding – The Undiscussed Backbone of AI Search (24:46–28:03)

• What is Grounding?
  • LLMs generate “fan out queries” from user prompts, querying search indices (Google for Google, Bing, Brave, etc. for others).
  • They select a subset of documents, then passages, to inform the final AI-generated answer.
  • Quote: “Some documents are selected based on some criteria…and all of that is then used together with my original prompt to generate my answer. That is the grounding process essentially.” (C, 27:53)

Implications for SEOs – Expanding the Game Board (31:03–34:13)

• SEO Still the Base Layer: Good content, indexation, and ranking remain foundational.
• Where the Pipeline Breaks:
  • Crawling/indexing
  • Ranking for fan out queries
  • Selection as candidate source
  • Cited (and in a positive light)
• Beyond Your Website: Influence important third-party sources (e.g., Reddit, YouTube, LinkedIn, news media); be mindful and ethical in engagement.

Consistency & Training Data: Controlling The Narrative (35:57–37:57)

• Key for Brands: Consistency is critical, especially for smaller brands and personal entities.
  • “Be consistent everywhere where you describe yourself…that can be enough that overnight the LLMs understand…” (C, 36:48)
• Check & Optimize Prompts: Evaluate how models describe your brand and ensure positivity and accuracy on key third-party sources.

Reverse-Engineering ChatGPT Shopping (39:08–43:57)

• ChatGPT is Scraping Google Shopping:
  • “They are scraping Google Shopping, they are using a third-party scraping provider…you get an 80%...overlap.” (C, 40:22)
  • Product feeds and identifiers provide traceable evidence.
• Implications: For now, maximize reach by sharing accurate product data; in the future, ChatGPT will likely own more product feeds and side-step legacy middle-men.

The Future: Shops as Fulfillment Centers (43:57–47:28)

• Marketplace Risks: AI agents/LLMs will drive users directly to product details, making online shops mere fulfillment centers.
• Prediction: “We will see a lot fewer online shops in the future…factory to LLM” cycles emerging, cutting out retailer middlemen.

Measurement: What to Track In an AI-Search Dominated World (48:14–50:21)

• Clicks & Search Volumes are Dead: Instead, focus on:
  1. Self-reported Attribution (“Where did you hear about us?”—sometimes 22% of leads attribute to ChatGPT, per case study. C, 48:34)
  2. Visibility on a Set of Representative Prompts
  3. Citation Share (How often is your brand/website cited in sourced answers?)
  4. Source Coverage (Are you mentioned positively in the most-used documents?)
• Actionable Wisdom: “If you get the citation share and the source coverage up, I guarantee you you will get up your visibility.” (C, 49:37)

Notable Quotes & Memorable Moments

• On the Rate of Change
  “This change that is happening with AI search is going to come faster and it's going to be more severe than I thought.”
  (Malta, 03:55)

• What’s Really Different Now?
  “There is no click…The whole mindset with the focus on clicks with easily attributable keywords and search volume that all goes away…”
  (Malta, 07:27)

• AI as Enemy or Opportunity
  “There's no reason to fight it. I think that was the biggest skill thing I learned… Let's not complain about this, let's embrace this…”
  (Malta, 12:51)

• How Grounding Works
  “If I write a prompt, the LLM…generates multiple fan out queries… The search engines return a list of results…some text passages from these documents are selected…to generate my answer. That is the grounding process essentially.”
  (Malta, 27:53)

• ChatGPT Shopping Secret
  “They are just scraping Google Shopping and that is what the ChatGPT shopping is based on.”
  (Malta, 41:13)

• On the Endgame for Online Shops
  “I predict that we will see a lot fewer online shops in the future. And I also believe we will see factory to LLM…”
  (Malta, 46:18)

• Top Measurement Priority
  “Ask your leads, ask your users, where did you hear about us?...I’ve seen one case where a click-based attribution said 0% of leads from ChatGPT; asking people revealed 22% from ChatGPT. Crazy difference.”
  (Malta, 48:34)

Important Timestamps

• [01:41] Malta’s origin story in SEO and AI search pivot
• [03:55] "Epiphany" about the speed/severity of AI search impact
• [07:27] The end of click-based attribution in AI search
• [13:50] Embracing rather than fighting inevitable industry changes
• [19:36] The case for focusing on influencing grounding processes
• [22:14] Paid ads in AI search: What (little) advertisers learn
• [27:53] In-depth explanation of grounding in LLM-driven search
• [36:48] Importance of brand and content consistency
• [40:22] Detailed reverse engineering of ChatGPT’s shopping mode
• [46:18] The “factory to LLM” prediction for online commerce
• [48:34] New measurement priorities: Self-reported attribution

Lightning Round – Actionable Takeaways (51:14–56:48)

• SEO Habits to Unlearn: Ditch click-based attribution; act like a brand marketer (C, 51:24)
• Auto-Generated Content at Scale: Works briefly, always penalized in the long-run (C, 51:49)
• Where to Invest: Content across the funnel and measurement—don’t obsess over perfect tracking; pursue directional signals (C, 53:11)
• Reddit, Wikipedia, Community Sites: Audit which are cited, build visibility in an ethical and brand-aligned manner (C, 54:23)
• Best Alternative Signal to Traffic: Still, self-reported attribution (C, 55:15)
• Prompt Selection: Use broad topics and user language; always analyze in aggregate, never obsess over one prompt or day (C, 55:35)

Key Takeaways

• SEO isn’t obsolete, but it’s morphing: Adapt to AI search, focus on visibility in grounding sources, and spread your influence—all senses of brand and authority matter more.
• Clicks and search volume are out; measurement must pivot: Self-reported attribution, prompt-based aggregate visibility, and citation/share coverage are the new KPIs.
• The future is “factory to LLM”: Merchants, especially those without differentiation, risk being reduced to logistics while AI agents own discovery and recommendations.
• Stay consistent and audit your data footprint: Inconsistencies confuse LLMs—harmonize your messaging across all digital presences.
• Embrace change, stay nimble, and think like a brand marketer.

For anyone navigating SEO in the age of generative AI, this episode is a roadmap for survival and growth. The traditional rules are crumbling, and those who adapt now will shape the next era of search.