Exploring the relationship between traditional SERPs and Gen AI search

Transcript

2. Hi, I’m @raygrieselhuber Founder & CEO of DemandSphere 18+ years experience in data systems, engineering, and enterprise SEO Speakerdeck.com/raygrieselhuber @raygrieselhuber
3. Shift to multipolarity? Traditional Search Monopoly Gen AI Search Landscape Speakerdeck.com/raygrieselhuber @raygrieselhuber
4. Where is the audience today vs. the future? 2004-2023 Effective Monopoly 2023-? Gen AI emergence One or multiple winners? Speakerdeck.com/raygrieselhuber @raygrieselhuber
5. Rapidly emerging behavior patterns in Gen AI search https:/ /blog.adobe.com/en/publish/2025/03/17/adobe-analytics-traffic-to-us-retail-websites-from-generative-ai-sources-jumps-1200-percent Speakerdeck.com/raygrieselhuber @raygrieselhuber
6. Traffic from Gen AI search is still a rounding error… https:/ /gs.statcounter.com/search-engine-market-share/all/united-states-of-america Gen AI search traffic < .03% Speakerdeck.com/raygrieselhuber @raygrieselhuber
7. But how are younger generations searching? Speakerdeck.com/raygrieselhuber @raygrieselhuber
8. Also, is traffic still our main KPI? Speakerdeck.com/raygrieselhuber @raygrieselhuber
9. Zero Click World: the shift from traffic to influence & SoV https:/ /sparktoro.com/blog/2024-zero-click-search-study-for-every-1000-us-google-searches-only-374-clicks-go-to-the-open-web-in-the-eu-its-360/ Speakerdeck.com/raygrieselhuber @raygrieselhuber
10. Questions abound… Speakerdeck.com/raygrieselhuber @raygrieselhuber
11. SoV Measurement & Optimization needs to be in context “10 blue links” Modern SERPs ChatGPT Perplexity Speakerdeck.com/raygrieselhuber @raygrieselhuber
12. The 10 blue links are back! Speakerdeck.com/raygrieselhuber @raygrieselhuber
13. Modern SERP Features Speakerdeck.com/raygrieselhuber @raygrieselhuber
14. The missing detail is factors that impact CTR Avg. Position (Max traffic potential) Impressions (depends on Avg. Pos.) Clicks CTR (clicks / impressions) What happened? Derived metric What happened? Derived metric WHY?! Speakerdeck.com/raygrieselhuber @raygrieselhuber
15. Pixels = Attention but do they matter in Gen AI? Speakerdeck.com/raygrieselhuber @raygrieselhuber
16. “SERP Features” in Gen AI Speakerdeck.com/raygrieselhuber @raygrieselhuber
17. “SERP Features” will continue to emerge in Gen AI too Visual Rank Sentiment Layout Shift Scroll Depth Pixel Depth CTR Modeling Ad Copy Suggested Keywords # of Elements Ad Location Locations Reviews Merchant IDs Business Titles Ranking URLs Title, Meta, etc. Custom Extraction SERP Screenshots Pixel Height SERP Features Search Intent Keyword Clusters Topic Modeling Search Volume Ad Presence Ad Performance Co-Occurrence Competitor Performance Competitor Discovery Share of Voice Visual Share of Voice URL Screenshots NLP Analysis Video Discussions & Forums Social Knowledge Graph SERP Feature Interiors FAQ Flight Details Hotel Details Review details PLA Text Ads People Also Ask Refine this search Organic Commerce Related Products Shops News News Details Price & Currency Google vs. Bing Buying Guide Howto Job Details
18. It feels a lot like the early 2000s • It’s somewhat of an achievement to simply have data (this is changing quickly) Speakerdeck.com/raygrieselhuber @raygrieselhuber
19. It feels a lot like the early 2000s • It’s somewhat of an achievement to simply have data (this is changing quickly) • There is no equivalent of GSC for Gen AI search engines (will there ever be?) Speakerdeck.com/raygrieselhuber @raygrieselhuber
20. It feels a lot like the early 2000s • It’s somewhat of an achievement to simply have data (this is changing quickly) • There is no equivalent of GSC for Gen AI search engines (will there ever be?) • The narrative around measurement must evolve toward clarity Speakerdeck.com/raygrieselhuber @raygrieselhuber
21. Let’s look at Gen AI Search Speakerdeck.com/raygrieselhuber @raygrieselhuber
22. The dirty “secret”: AI search depends on traditional organic indexes Speakerdeck.com/raygrieselhuber @raygrieselhuber
23. Two reasons Speakerdeck.com/raygrieselhuber @raygrieselhuber
24. The training data is almost always stale Speakerdeck.com/raygrieselhuber @raygrieselhuber
25. They also hallucinate Speakerdeck.com/raygrieselhuber @raygrieselhuber
26. Live Retrieval (RAG) is the key Search Index Foundational Model Speakerdeck.com/raygrieselhuber @raygrieselhuber
27. Solving the stale content problem with Retrieval Augmented Generation (RAG) and Google / Bing LLMs Search Engines Answer engine uses combines LLM answers with organic index from Google or Bing (RAG) User receives higher quality and more relevant answer
28. Google & Bing form the backbone of AI search Google AI Overviews Speakerdeck.com/raygrieselhuber @raygrieselhuber
29. ChatGPT is using BOTH Google and Bing indexes These citation sources are from Google. See next slide.
30. ChatGPT is using BOTH Google and Bing indexes Speakerdeck.com/raygrieselhuber @raygrieselhuber
31. AI Overviews are Google’s RAG in the SERP Speakerdeck.com/raygrieselhuber @raygrieselhuber
32. Data Study (more on this later) • 10,000 prompts • 17 industries • Google • Bing • ChatGPT • Perplexity
33. How we gathered data Speakerdeck.com/raygrieselhuber @raygrieselhuber
34. The key to measuring visibility is knowing what your audience is asking Speakerdeck.com/raygrieselhuber @raygrieselhuber
35. Prompt research is the new keyword research Speakerdeck.com/raygrieselhuber @raygrieselhuber
36. But where to get data for prompt research? Speakerdeck.com/raygrieselhuber @raygrieselhuber
37. It turns out, yet again, Google is the winner here Speakerdeck.com/raygrieselhuber @raygrieselhuber
38. Google’s “database of intentions” is also their “database of interactions” Speakerdeck.com/raygrieselhuber @raygrieselhuber
39. The database of user interactions • The Google antitrust case has proven that Google uses “user interactions” to influence its ranking systems • Google’s total dominance over the search market in the last 20 years gives them an unmatched data advantage • This data advantage reveals itself in SERP Features designed to keep more interactions on the SERPs • This data can be mined Speakerdeck.com/raygrieselhuber @raygrieselhuber
40. This means that Google has more data than anybody on what people will ask about People Also Ask (PAA) and similar features are based on Google’s extensive data about what people ask next for a given search. Side note: they also know what people are most likely to buy
41. Questions to AI search engines “look” like informational queries: • Longer query length ◦ “best winter tires 4Runner” vs. “What are the best winter tires for my 4Runner” Speakerdeck.com/raygrieselhuber @raygrieselhuber
42. Questions to AI search engines “look” like informational queries: • Longer query length ◦ “best winter tires 4Runner” vs. “What are the best winter tires for my 4Runner” • Lower measurable search volume per query Speakerdeck.com/raygrieselhuber @raygrieselhuber
43. Questions to AI search engines “look” like informational queries: • Longer query length ◦ “best winter tires 4Runner” vs. “What are the best winter tires for my 4Runner” • Lower measurable search volume per query • You can use SERP data to tie informational queries back to commercial intent Speakerdeck.com/raygrieselhuber @raygrieselhuber
44. This data can be used for prompt discovery Speakerdeck.com/raygrieselhuber @raygrieselhuber
45. Tie prompt targeting to desired search intent and volume Commercial best winter tires People Also Asked People Also Asked People Also Asked People Also Asked Which brand is the best for winter? Why are Blizzak tires so good? Which is better for snow awd or snow tires? Do snow tires really make a difference?
46. Start with desired search intent and search volume
47. Mine PAA questions for prompt opportunities
48. Mine PAA questions at scale for commercial intent Speakerdeck.com/raygrieselhuber @raygrieselhuber
49. Mine PAA questions at scale for commercial intent Speakerdeck.com/raygrieselhuber @raygrieselhuber
50. Clustering across questions occurs in many places Speakerdeck.com/raygrieselhuber @raygrieselhuber
51. Data Study • 10,000 prompts • 17 industries • Google • Bing • ChatGPT • Perplexity
52. (White paper in the works) Speakerdeck.com/raygrieselhuber @raygrieselhuber
53. Reliance on Google & Bing Speakerdeck.com/raygrieselhuber @raygrieselhuber 51.72% Google Index ChatGPT 4o 14.26% Bing Index ChatGPT 4o 34.02% Custom / Remix ChatGPT 4o 37.12% Google Index Perplexity 45.63% Custom / Remix Perplexity 17.25% Bing Index Perplexity
54. Evaluating SoV and other results Speakerdeck.com/raygrieselhuber @raygrieselhuber
55. Ranking & Appearance Speakerdeck.com/raygrieselhuber @raygrieselhuber 5.8 Google Avg. Rank ChatGPT 4o 5.6 Top Pos. Avg. Google Rank ChatGPT 4o 8 75% of results, Google Rank ChatGPT 4o 5.1 Google Avg. Rank Perplexity 7 75% of results, Google Rank Perplexity 6.78 Top Pos. Avg. Google Rank Perplexity
56. This points to some serious re-ranking Speakerdeck.com/raygrieselhuber @raygrieselhuber
57. Re-ranking of the search index results Search Index Foundational Model Speakerdeck.com/raygrieselhuber @raygrieselhuber Ranking System
58. You can just ask them how they work Speakerdeck.com/raygrieselhuber @raygrieselhuber
59. But they are not always forthcoming Speakerdeck.com/raygrieselhuber @raygrieselhuber
60. Takeaway #1: Do not abandon dominant traffic sources today to chase emerging trends Speakerdeck.com/raygrieselhuber @raygrieselhuber
61. Takeaway #2: Recognize this as a time of transition Speakerdeck.com/raygrieselhuber @raygrieselhuber
62. Takeaway #3: Keep multiple paradigms in mind at the same time Speakerdeck.com/raygrieselhuber @raygrieselhuber
63. Takeaway #4: Be quick to test and experiment emerging platforms and behaviors Speakerdeck.com/raygrieselhuber @raygrieselhuber
64. State of the industry in Gen AI search monitoring • Results in UI differ from API • Web UIs must be scraped for links • Direct model monitoring can be used via API for brand mentions • Very aggressive anti-bot / anti-scraping measures are in place
65. State of the industry in Gen AI search monitoring • Monitoring Gen AI is more expensive with fewer players • Achieve scale by combining Google + Bing with Gen AI monitoring • Focus Gen AI monitoring on priority queries • The future lies with platforms that can do both
66. Come visit our booth! Speakerdeck.com/raygrieselhuber @raygrieselhuber x.com/demandsphere linkedin.com/in/raygrieselhuber