Tweet
Tweet

Slide 1

Slide 1 text

Gianluca Fiorelli International & Strategic SEO Consultant X –LinkedIn IloveSEO.net BRAND, SEO & GENERATIVE SEARCH

Slide 2

Slide 2 text

Our path BRAND, SEO & GENERATIVE SEARCH 3 1 2 What How Why #WMF2024

Slide 3

Slide 3 text

WHAT 1 of 3

Slide 4

Slide 4 text

No content

Slide 5

Slide 5 text

No content

Slide 6

Slide 6 text

No content

Slide 7

Slide 7 text

No content

Slide 8

Slide 8 text

No content

Slide 9

Slide 9 text

No content

Slide 10

Slide 10 text

No content

Slide 11

Slide 11 text

No content

Slide 12

Slide 12 text

No content

Slide 13

Slide 13 text

No content

Slide 14

Slide 14 text

No content

Slide 15

Slide 15 text

No content

Slide 16

Slide 16 text

And all this happened after months of beta testing

Slide 17

Slide 17 text

No content

Slide 18

Slide 18 text

No content

Slide 19

Slide 19 text

No content

Slide 20

Slide 20 text

No content

Slide 21

Slide 21 text

No content

Slide 22

Slide 22 text

No content

Slide 23

Slide 23 text

And it seems that Google is serious with correcting the issues and regaining the trust of the users (and media).

Slide 24

Slide 24 text

Even at the cost of making AIO almost invisible for a time.

Slide 25

Slide 25 text

“If you're seeing them far less, you're correct, they're pretty much gone in a few sectors!” (Aleyda Solís)

Slide 26

Slide 26 text

No content

Slide 27

Slide 27 text

No content

Slide 28

Slide 28 text

No content

Slide 29

Slide 29 text

SGE AIO

Slide 30

Slide 30 text

What types of AI Overviews we have seen so far?

Slide 31

Slide 31 text

Check her post!

Slide 32

Slide 32 text

Duplicative AI Overviews

Slide 33

Slide 33 text

No content

Slide 34

Slide 34 text

No content

Slide 35

Slide 35 text

Summarizing/Complementary AI Overviews

Slide 36

Slide 36 text

Investopedia Chase Bank NerdWallet (2) Forbes Bankrate

Slide 37

Slide 37 text

Chasebank (#1) NerdWallet 1 (#2) Forbes (#3) Investopedia (#4) Bankrate (#7) NerdWallet 2 (#9)

Slide 38

Slide 38 text

Accelerators AI Overviews

Slide 39

Slide 39 text

Practically the same behavior of the Featured Snippets… Also note how AIO used the caption of the photo for generating its snapshot.

Slide 40

Slide 40 text

No content

Slide 41

Slide 41 text

AI Overview designed as a sort of Pillar/Landing page that…

Slide 42

Slide 42 text

… sends you to a new AIO pages Merchant/PDPs…

Slide 43

Slide 43 text

… offering so a completely alternative search journey experience than the classic organic PLP search results list.

Slide 44

Slide 44 text

It is the ”ecommerce” query space – also for branded searches – where AI Overview is experimenting the most.

Slide 45

Slide 45 text

Also, because Ads are going to come quite soon on AI Overviews… probably when Google will have fixed the most serious hallucination issues it experienced at its roll-out.

Slide 46

Slide 46 text

In the meanwhile, Ads tend to be shown above the AI Overview feature.

Slide 47

Slide 47 text

HOW 2 of 3

Slide 48

Slide 48 text

How does Google create the AI Overviews?

Slide 49

Slide 49 text

And what can we do to appear in AIO and how can we track and measure our presence?

Slide 50

Slide 50 text

A first answer is offered to us by Google itself in its post announcing the update and corrections they are rolling out after the media backlash followed to Google I/O.

Slide 51

Slide 51 text

AI Overviews […] are not simply generating an output based on training data. Liz Reid VP – Head of Search Google

Slide 52

Slide 52 text

While AI Overviews are powered by a customized language model, the model is integrated with our core web ranking systems and designed to carry out traditional “search” tasks, like identifying relevant, high-quality results from our index. That’s why AI Overviews don’t just provide text output but include relevant links so people can explore further. Liz Reid VP – Head of Search Google

Slide 53

Slide 53 text

LLMS

Slide 54

Slide 54 text

RAG

Slide 55

Slide 55 text

https://patents.google.com/patent/US11769017B1/en

Slide 56

Slide 56 text

The process

Slide 57

Slide 57 text

Query submission

Slide 58

Slide 58 text

Analyzing the query using different LLMs for understanding context and intent.

Slide 59

Slide 59 text

Content retrieval (text, images, videos…)

Slide 60

Slide 60 text

Processing additional content from related searches and information others found helpful under similar query circumstances.

Slide 61

Slide 61 text

Generation of the summary that’s supposed to be directly helpful to the user, who inputted the query.

Slide 62

Slide 62 text

Dynamic and Context-Aware Responses: The AI adapts its responses based on the specific context of each query submission.

Slide 63

Slide 63 text

Google learns from how users interact with the search results. If certain responses are consistently overlooked in favor of others, or if user clicks suggest a preference, the AI modifies future responses to align more closely with user behavior.

Slide 64

Slide 64 text

The AI continually updates its understanding based on new information and interactions.

Slide 65

Slide 65 text

Which LLM?

Slide 66

Slide 66 text

“…the system selects, from multiple candidate generative models: none, one, or multiple to utilize in generating response(s) to render responsive to the query.”

Slide 67

Slide 67 text

Query type, using classifiers. Factors for choosing the LLM(s).

Slide 68

Slide 68 text

Analysis of the SERP characteristics (top results and content types). Factors for choosing the LLM(s).

Slide 69

Slide 69 text

Processing of the query content using ML classifiers. Factors for choosing the LLM(s).

Slide 70

Slide 70 text

Application of rules based on presence or absence of words or voice characteristics. Factors for choosing the LLM(s).

Slide 71

Slide 71 text

Considerations about the quality and trustworthiness of search result documents. Factors for choosing the LLM(s).

Slide 72

Slide 72 text

Individuation of the presence of certain words or content types in the query. Factors for choosing the LLM(s).

Slide 73

Slide 73 text

Contextualization and personalization. Factors for choosing the LLM(s).

Slide 74

Slide 74 text

Resource optimization for minimizing computational costs. Factors for choosing the LLM(s).

Slide 75

Slide 75 text

What sources?

Slide 76

Slide 76 text

Query-dependent. Factors for choosing the sources.

Slide 77

Slide 77 text

Query-dependent. Factors for choosing the sources. Location.

Slide 78

Slide 78 text

Query-dependent. Factors for choosing the sources. Location. Language.

Slide 79

Slide 79 text

Query-independent. Factors for choosing the sources.

Slide 80

Slide 80 text

Query-independent. Factors for choosing the sources. Popularity.

Slide 81

Slide 81 text

Query-independent. Factors for choosing the sources. Popularity. Trustworthiness.

Slide 82

Slide 82 text

Query-independent. Factors for choosing the sources. Popularity. Trustworthiness. Content diversity.

Slide 83

Slide 83 text

Query-independent. Factors for choosing the sources. Popularity. Trustworthiness. Content diversity. Relevance over rank.

Slide 84

Slide 84 text

What does it mean for us?

Slide 85

Slide 85 text

What does it mean for us? Factors for choosing the sources. Popularity... aka being the source of reference for the quality and usefulness of the resources we create, so to earn natural links and mentions.

Slide 86

Slide 86 text

What does it mean for us? Factors for choosing the sources. Popularity... aka not limiting ourselves in creating “buyer personas” but also “audience personas”. Who (persons & websites) are the true influencers of our audience? Create content these influencers want to use and amplify. Involve them in your Digital PR campaigns.

Slide 87

Slide 87 text

What does it mean for us? Factors for choosing the sources. Trustworthiness... aka “listen to your users”, interact with them onsite and offsite, and embrace a product-led SEO strategy.

Slide 88

Slide 88 text

What does it mean for us? Factors for choosing the sources. Content diversity... aka stop considering only “text” as content: long videos, short videos, images, audio/podcasts. And text is not only blog posts and landing pages. Embrace the concept of repurposing.

Slide 89

Slide 89 text

What does it mean for us? Factors for choosing the sources. Relevance over rank... aka detect the search intents and their seasonalities and create content accordingly.

Slide 90

Slide 90 text

What does it mean for us? Factors for choosing the sources. Relevance over rank... also means semantic relevance: do not stress writers about “SEO Copywriting” but urge them to write naturally over a base of expert-led documentation. Natural language naturally leads to semantic correctness.

Slide 91

Slide 91 text

Factors for choosing the sources. User-dependent. Search History.

Slide 92

Slide 92 text

Factors for choosing the sources. User-dependent. Search History. ”Click story” of the user (on SERP).

Slide 93

Slide 93 text

Factors for choosing the sources. User-dependent. Search History. Dwell time (on site). ”Click story” of the user (on SERP).

Slide 94

Slide 94 text

Factors for choosing the sources. User-dependent. Search History. Scrolling behavior (on SERP). ”Click story” of the user (on SERP). Dwell time (on site).

Slide 95

Slide 95 text

Factors for choosing the sources. User-dependent. Search History. Result expansion/Highlighting (on SERP). ”Click story” of the user (on SERP). Dwell time (on site). Scrolling behavior (on SERP).

Slide 96

Slide 96 text

User-dependent. Factors for choosing the sources. Search History. Interaction with Search Features (on SERP). ”Click story” of the user (on SERP). Dwell time (on site). Scrolling behavior (on SERP). Result expansion/Highlighting (on SERP).

Slide 97

Slide 97 text

Related queries. Factors for choosing the sources.

Slide 98

Slide 98 text

Recent queries performed by the user. Factors for choosing the sources.

Slide 99

Slide 99 text

Implied queries. Factors for choosing the sources.

Slide 100

Slide 100 text

What does it mean for us?

Slide 101

Slide 101 text

Factors for choosing the sources. What does it mean for us?

Slide 102

Slide 102 text

Factors for choosing the sources. What does it mean for us? Analyze the Search Journey inside Google and the Messy Middle.

Slide 103

Slide 103 text

Factors for choosing the sources. What does it mean for us? Create/Update the architecture of your website based on that analysis & own the space.

Slide 104

Slide 104 text

Factors for choosing the sources. What does it mean for us? Stop thinking in “One query One click” aka do not fall in the melancholy of the 0-click SERPs. SEO is about Search Journeys and Search Sessions, not single SERPs.

Slide 105

Slide 105 text

E-E-A-T? Experience and Expertise lead to relevance. Relevance leads to Trustworthiness. Trustworthiness leads to Authority. Owning and demonstrating E- E-A-T mean being a “brand”.

Slide 106

Slide 106 text

E-E-A-T? A Brand is a named entity that Google considers an important node in the Knowledge Graph, hence firmly connected to other nodes (i.e.: topics, products) and that have distinctive attributes. All this translates to visibility for the brand when people search about topics/products and related attributes.

Slide 107

Slide 107 text

Does AIO search in real time?

Slide 108

Slide 108 text

NO!

Slide 109

Slide 109 text

It uses the pre-existing training data and the Google’s index.

Slide 110

Slide 110 text

Can we track and measure traffic from AIO?

Slide 111

Slide 111 text

YES and NO

Slide 112

Slide 112 text

Can we track and measure traffic from AIO? Nothing on Google Search Console.

Slide 113

Slide 113 text

Can we track and measure traffic from AIO? Tools like Semrush started tracking AIO.

Slide 114

Slide 114 text

Can we track and measure traffic from AIO? Brightedge is very focused on AIO and uses an approach to tracking it different than other suites.

Slide 115

Slide 115 text

Can we track and measure traffic from AIO? Extension for Chrome started to be created. This below is probably the most useful.

Slide 116

Slide 116 text

Can we track and measure traffic from AIO? Extension for Chrome started to be created. This below is probably the most useful.

Slide 117

Slide 117 text

Can we track and measure traffic from AIO? Ziptie.dev is maybe the most interesting too right now.

Slide 118

Slide 118 text

Can we track and measure traffic from AIO? Ziptie.dev is maybe the most interesting too right now.

Slide 119

Slide 119 text

Can we track and measure traffic from AIO? Ziptie.dev is maybe the most interesting too right now.

Slide 120

Slide 120 text

Can we track and measure traffic from AIO? Ziptie.dev is maybe the most interesting too right now.

Slide 121

Slide 121 text

WHY 3 of 3

Slide 122

Slide 122 text

We have found that with AI Overviews, people use Search more... Liz Reid VP – Head of Search Google

Slide 123

Slide 123 text

Our data are telling us that people click a lot on the sources linked in CoPilot ... Fabrice Canel Principal Program Manager BING

Slide 124

Slide 124 text

Can we track and measure traffic from AIO? Experiments like this one below seem confirming that AIO generates traffic.

Slide 125

Slide 125 text

Google is plagiarizing me! Not refers me! It is stealing my traffic! Many bloggers. How to ”forbid” AIO to show us?

Slide 126

Slide 126 text

How to ”forbid” AIO to show us? https://developers.google.com/search/docs/appearance/ai-overviews https://developers.google.com/search/docs/appearance/snippet#nosnippet

Slide 127

Slide 127 text

What do I recommend to my clients?

Slide 128

Slide 128 text

What do I recommend to my clients?

Slide 129

Slide 129 text

Summary of what we have seen 3 1 2 AIO won’t disappear Don’t be silly Optimize! #WMF2024

Slide 130

Slide 130 text

VOTE FOR THE INTERVENTION ON “IBRIDA”

Slide 131

Slide 131 text

Gianluca Fiorelli International & Strategic SEO Consultant X –LinkedIn IloveSEO.net BRAND, SEO & GENERATIVE SEARCH