Reimagining SEO in the AI Era [SoCi Reimagine 2024 Talk]

Slide 1

Slide 1 text

Reimagining SEO in the AI Era Keynote Speaker Michael King, CEO of iPullRank

Slide 2

Slide 2 text

2 2

Slide 3

Slide 3 text

3 3 Download this deck: https://speakerdeck.com/ipullrank

Slide 4

Slide 4 text

4 Salutations! I’m Mike King (@iPullRank)

Slide 5

Slide 5 text

Slide 6

Slide 6 text

6 People are Searching Differently?

Slide 7

Slide 7 text

7 7 “Search It Up” is the new “Google it”

Slide 8

Slide 8 text

8 8

Slide 9

Slide 9 text

9 9 What does it mean when a brand ceases to be a verb?

Slide 10

Slide 10 text

10 There’s A Lot of Discussion of ChatGPT Replacing Google The same is true for Tiktok, Perplexity, Bing’s CoPilot, and [insert new genAI search tool here].

Slide 11

Slide 11 text

11 11 40% of People Leaving ChatGPT Go to Google My assumption is that many of these people are fact-checking. That is a bad behavior to establish for nearly half of your users. This is also an indication that people are deeply aware of the issues related to hallucinations. In other words, people don’t trust the product.

Slide 12

Slide 12 text

12 12 Yes, TikTok is a (Nascent) Search Engine 41% of Tiktok users perform searches, but the search volume around a series of broad and meaningful queries is not there to make it more than a small supplement to Google Search.

Slide 13

Slide 13 text

13 13 21% of People Going to Tiktok Come from Google 24.5% of People Leaving Tiktok Go to Google

Slide 14

Slide 14 text

14 14 The Last Time the Major Search Engine Died it Looked Like This

Slide 15

Slide 15 text

15 15 Google Search Still Dwarfs Everything, but More People are Using More Channels

Slide 16

Slide 16 text

16 16 It means there is fragmentation in how information needs are being met.

Slide 17

Slide 17 text

17 17 I wish OpenAI the best with this. It will be very diﬃcult to supplant Google as the search engine of record.

Slide 18

Slide 18 text

18 18 Users Have Very High Expectations of a Search Box

Slide 19

Slide 19 text

19 19 …but search outside of major search engines is still really bad

Slide 20

Slide 20 text

20 20 Google is still the main event, but we are going back into a world where we need to optimize for multiple search engines across a series of channels.

Slide 21

Slide 21 text

21 GenAI Comes from Search Tech

Slide 22

Slide 22 text

22 22 Real quick. Let’s talk about how search engines work.

Slide 23

Slide 23 text

23 23 Search Engines Work based on the Vector Space Model Documents and queries are plotted in multidimensional vector space. The closer a document vector is to a query vector, the more relevant it is.

Slide 24

Slide 24 text

24 24 TF-IDF Vectors The vectors in the vector space model were built from TF-IDF. These were simplistic based on the Bag-of-Words model and they did not do much to encapsulate meaning.

Slide 25

Slide 25 text

25 25 Relevance is a Function of Cosine Similarity When we talk about relevance, it’s the question of similar is determined by how similar the vectors are between documents and queries. This is a quantitative measure, not the qualitative idea of how we typically think of relevance.

Slide 26

Slide 26 text

26 26 The lexical model counts the presence and distribution of words. Whereas the semantic model captures meaning. This was the huge quantum leap behind Google’s Hummingbird update and most SEO software has been behind for over a decade. Google Shifted from Lexical to Semantic a Decade Ago

Slide 27

Slide 27 text

27 Word2Vec Gave Us Embeddings Word2Vec was an innovation led by Tomas Mikolov and Jeff Dean that yielded an improvement in natural language understanding by using neural networks to compute word vectors. These were better at capturing meaning. Many follow-on innovations like Sentence2Vec and Doc2Vec would follow.

Slide 28

Slide 28 text

28 28 Tomas Mikolov Led the Word2Vec Research Tomas is a Czech computer scientist behind many of these natural language understanding innovations.

Slide 29

Slide 29 text

29 He was accompanied by the Chuck Norris of Computer Science Jeff Dean Jeff Dean has been a part of nearly every major innovation that has powered Google Search.

Slide 30

Slide 30 text

30 30 This Allows for Mathematical Operations Comparisons of content and keywords become linear algebraic operations.

Slide 31

Slide 31 text

31 31 Words are Converted to Multi-dimensional Coordinates in Vector Space

Slide 32

Slide 32 text

32 32 We Went from Sparse Representations to Dense Representations

Slide 33

Slide 33 text

33 33 Word2Vec Gave Us Hummingbird

Slide 34

Slide 34 text

34 34 8 Google Employees Are Responsible for Generative AI https://www.wired.com/story/eight-google-employees-invented-modern-ai-transformers-paper/

Slide 35

Slide 35 text

35 35 The Transformer The transformer is a deep learning model used in natural language processing (NLP) that relies on self-attention mechanisms to process sequences of data simultaneously, improving eﬃciency and understanding in tasks like translation and text generation. Its architecture enables it to capture complex relationships within the text, making it a foundational model for many state-of-the-art NLP applications.

Slide 36

Slide 36 text

36 36 The Transformer Gave us BERT

Slide 37

Slide 37 text

37 37 Word2Vec Captured Relationship, but Not Context – BERT Captures Context

Slide 38

Slide 38 text

38 38 BERT Yields Embeddings with Higher Dimensionality and Information Capture

Slide 39

Slide 39 text

39 39 Leveraging generative AI is a combination of content strategy, your unique creative angles, and deep understanding of the technical nuances of a channel.

Slide 40

Slide 40 text

40 40 SEOs are uniquely positioned for this moment.

Slide 41

Slide 41 text

41 41 This is our opportunity to get up from the kids table. This is our opportunity to get up from the kids table

Slide 42

Slide 42 text

42 The threat of Google’s AI Overviews

Slide 43

Slide 43 text

43 43 Queries are Longer and the Featured Snippet is Bigger 1. The query is more natural language and no longer Orwellian Newspeak. It can be much longer than the 32 words that is has been historically in order 2. The Featured Snippet has become the “AI snapshot” which takes 3 results and builds a summary. 3. Users can also ask follow up questions in conversational mode.

Slide 44

Slide 44 text

44 44 The Search Demand Curve will Shift With the change in the level of natural language query that Google can support, we’re going to see a lot less head terms and a lot more long tail term.

Slide 45

Slide 45 text

45 45 The CTR Model Will Change With the search results being pushed down by the AI snapshot experience, what is considered #1 will change. We should also expect that any organic result will be clicked less and the standard organic will drop dramatically. However, this will likely yield query displacement.

Slide 46

Slide 46 text

46 Rank Tracking Will Be More Complex As an industry, we’ll need to decide what is considered the #1 result. Based on this screenshot positions 1-3 are now the citations for the AI snapshot and #4 is below it. However, the AI snapshot loads on the client side, so rank tracking tools will need to change their approach.

Slide 47

Slide 47 text

47 47 Context Windows Will Yield More Personalized Results AIO maintains the context window of the previous search in the journey as the user goes through predeﬁned follow questions. This will need to drive the composition of pages to ensure they remain in the consideration set for subsequent results.

Slide 48

Slide 48 text

48 AI Overviews are Susceptible to Spam and Lower Quality Sites AI Overviews operate outside of core Organic Search so they don’t always have the same spam ﬁlters.

Slide 49

Slide 49 text

49 49 Luckily Users Love it So Much

Slide 50

Slide 50 text

50 50 Ranking in AI Overviews is more about relevance than the other signals.

Slide 51

Slide 51 text

51 What is Retrieval Augmented Generation (RAG)?

Slide 52

Slide 52 text

52 52 Combining a Search Engine with a Language Model is called “Retrieval Augmented Generation” Neeva (RIP), Bing, and now Google’s Search Generative Experience all use pull documents based on search queries and feed them to a language model to generate a response. This concept was developed by the Facebook AI Research (FAIR) team.

Slide 53

Slide 53 text

53 53 Google’s Initial Version of this is called Retrieval-Augmented Language Model Pre-Training (REALM) from 2021 REALM identiﬁes full documents, ﬁnds the most relevant passages in each, and returns the single most relevant one for information extraction.

Slide 54

Slide 54 text

54 54 DeepMind followed up with Retrieval-Enhanced Transformer (RETRO) DeepMind's RETRO (Retrieval-Enhanced Transformer) is a language model that combines a large text database with a transformer architecture to improve performance and reduce the number of parameters required. RETRO is able to achieve comparable performance to state-of-the-art language models such as GPT-3 and Jurassic-1, while using 25x fewer parameters.

Slide 55

Slide 55 text

55 Google’s Later Innovation Retroﬁt Attribution using Research and Revision (RARR) RARR does not generate text from scratch. Instead, it retrieves a set of candidate passages from a corpus and then reranks them to select the best passage for the given task.

Slide 56

Slide 56 text

56 56 AIO is built from REALM/RETRO/RARR + PaLM 2 and MUM MUM is the Multitask Uniﬁed Model that Google announced in 2021 as way to do retrieval augmented generation. PaLM 2 is their latest (released) state of the art large language model. The functionality from REALM, RETRO, and RARR is also rolled into this.

Slide 57

Slide 57 text

57 57 Sounds cool, but how does it work?

Slide 58

Slide 58 text

58 58 Documents are Broken into Chunks and the Most Relevant Chunks are Fed to the Language Model to Generate a Response

Slide 59

Slide 59 text

59 How to Appear in LLMs

Slide 60

Slide 60 text

60 60 Blocking LLMs is a Mistake. Appearing in these places will be recognized as brand awareness opportunities very soon.

Slide 61

Slide 61 text

Embrace Structured Data There are three models gaining popularity: 1. KG-enhanced LLMs - Language Model uses KG during pre-training and inference 2. LLM-augmented KGs - LLMs do reasoning and completion on KG data 3. Synergized LLMs + KGs - Multilayer system using both at the same time https://arxiv.org/pdf/2306.08302.pdf Source: Unifying Large Language Models and Knowledge Graphs: A Roadmap

Slide 62

Slide 62 text

62 62 What is Mitigation for AIO? 1. Manage expectations on the impact 2. Understand the keywords under threat 3. Re-prioritize your focus to keywords that are not under threat 4. Optimize the passages for the keywords you want to save

Slide 63

Slide 63 text

63 63

Slide 64

Slide 64 text

64 64 We Can Also Show You Per Keyword How You Show Up

Slide 65

Slide 65 text

65 65 It’s all about the Fraggles. (Fragment + Handle)

Slide 66

Slide 66 text

66 66

Slide 67

Slide 67 text

67 67 The Fraggles Show What AIO Used for the AI Snapshot

Slide 68

Slide 68 text

68 68 Scroll to Text

Slide 69

Slide 69 text

69 69 Fraggles Relevance Relevance against the chunks to keyword: Relevance against AI Snapshot:

Slide 70

Slide 70 text

Check out MarketBrew’s Free Tool to Help

Slide 71

Slide 71 text

No content

Slide 72

Slide 72 text

No content

Slide 73

Slide 73 text

No content

Slide 74

Slide 74 text

74 The GEO team shared their ChatGPT prompts The GEO team also shared the ChatGPT prompts that help them improve their visibility. You can augment them and put them to work right away. https://github.com/GEO-optim/GEO/blo b/main/src/geo_functions.py

Slide 75

Slide 75 text

Check out @GarrettSussman’s post on how to optimize for AI Overviews: https://ipullrank.com/optimize-content-for-sge

Slide 76

Slide 76 text

76 So, It Turns Out Google is All About the Clicks

Slide 77

Slide 77 text

77 77 Google’s Algorithms Inner Workings Have Been Put on Full Display Lately Through a combination of what’s come out of Google’s DOJ antitrust trial and the Google API documentation leak, we have a much clearer picture of how Google actually functions.

Slide 78

Slide 78 text

78 78 I was the First to Publish on the Google Leak

Slide 79

Slide 79 text

79 79 We Now Have a Much Stronger Understanding of the Architecture https://searchengineland.com/how-google-search-ranking-works-445141

Slide 80

Slide 80 text

80 80 These speciﬁcs help us zero in on what really matters.

Slide 81

Slide 81 text

81 The Primary Takeaway is the Value of User Behavior in Organic Search Google’s Navboost system keeps track of user click behavior and uses that to inform what should rank in various contexts.

Slide 82

Slide 82 text

82 82 Google Has Denied this Many Times

Slide 83

Slide 83 text

83 83 Many times…

Slide 84

Slide 84 text

84 84 But it’s Been Conﬁrmed in Pandu Nayak’s Testimony

Slide 85

Slide 85 text

85 85 And in Google’s documentation for their Cloud Search services

Slide 86

Slide 86 text

86 The Leaked Docs Have Several Features Measuring Clicks “Long clicks” are those clicks that resulted in a successful user session meaning that the user did not come back to Google or otherwise demonstrated that they found what they were looking for.

Slide 87

Slide 87 text

87 87

Slide 88

Slide 88 text

88 88 That’s Also Why Google Was So Mad when @DejanSEO Did This…

Slide 89

Slide 89 text

89 89 User Click Data is What Makes Google More Powerful Than Any Other Search Engine The court opinion in the DoJ Antitrust trial, Google’s leaked documents, and Google’s own internal documentation all support the fact that click behavior is what makes Google perform the way that it does.

Slide 90

Slide 90 text

90 90 13 Months of Google Data = 17 Years of Bing Data

Slide 91

Slide 91 text

91 Modern SEO Needs UX Baked-in Google has expectations of performance for every position in the SERP. The user behavior signals collected reinforce what should rank and demote what doesn’t perform just like a social media channel. The best way to scale this is by generating highly-relevant content with a strong user experience.

Slide 92

Slide 92 text

92 Local Insights from the Google Leak

Slide 93

Slide 93 text

93 93 The leak indicates that Google is trying to localize every page and every query.

Slide 94

Slide 94 text

94 94 Although this yields signals that impact rankings positively, this is not what you should be doing for long term success.

Slide 95

Slide 95 text

95 User Click Behavior Can Brand a Page with a Location Again, continued performance in Google is all about the clicks. The clickRadius50Percent indicates that Google will assign a location to a document based on the location where half its clicks come from. Action: For location based content, make it clear as early as the metadata that this page is relevant to speciﬁc locations and leverage rel-canonical tags effectively.

Slide 96

Slide 96 text

96 96 Google Attempts to Localize All Documents Google looks to semantic and explicit signals within a document to associate a location with it. Action: Clarify which location a page is associated with in language on the page.

Slide 97

Slide 97 text

97 Google Uses Many Semantic Signals to Determine Location Google looks to semantic and explicit signals within a document to associate a location with it. Action: Clarify which location a page is associated with in language on the page.

Slide 98

Slide 98 text

98 98 Your Primary Category Is Given a Ton of Value The primary category in your GMB holds a lot of weight for what you rank for. If you have locations going rogue, show them this slide and tell them they are shooting themselves in the foot. Action: Make sure your primary category is set to the one that matters most to the business.

Slide 99

Slide 99 text

99 99 Signals from Location Pages are Very Valuable As Well People generally neglect the location landing pages as secondary. The LocalWWWInfo module in the leak suggests that the topics and relationships of these pages have impact on how your location ranks. Action: Spend the time to make these pages robust and original.

Slide 100

Slide 100 text

10 0 10 0 Locations Are Associated with “Conceptual Centers” Google identiﬁes the centroid for a given concept in a given area. There’s not much to do here if the business location is already set, but if the business is considering moving, you may want to weigh in with this data to give your locations a better opportunity.

Slide 101

Slide 101 text

10 1 10 1 You’re Not Imagining it Chains are Treated Differently Google treats bigger brands separately and applies what it understands across the masterbrand to all locations. If you’re not a brand, this does not apply to you, but it’s effectively a multiplier effect that the bigger brands beneﬁt from.

Slide 102

Slide 102 text

10 2 10 2 NAP is Still Critical The basics of making sure your name, address, and phone number are consistent across all citations is still a very valuable thing to do.

Slide 103

Slide 103 text

10 3 10 3 For Local Search the Leak Data Reinforces Best Practices Get your on-page SEO house in order, but really focus on your GMB, reviews, and citations.

Slide 104

Slide 104 text

10 4 Less is More, More or Less It’s time to cut out the content madness

Slide 105

Slide 105 text

105 10 5 You Don’t Need Link Volume, You Need Link Quality Indexing Tier Impacts Link Value A metric called sourceType that shows a loose relationship between the where a page is indexed and how valuable it is. For quick background, Google’s index is stratified into tiers where the most important, regularly updated, and accessed content is stored in flash memory. Less important content is stored on solid state drives, and irregularly updated content is stored on standard hard drives. The higher the tier, the more valuable the link. Pages that are considered “fresh” are also considered high quality. Suffice it to say, you want your links to come from pages that either fresh or are otherwise featured in the top tier.

Slide 106

Slide 106 text

10 6 10 6 No one is actually improving content when they do the Skyscraper Technique

Slide 107

Slide 107 text

10 7 Indexing is Also Harder It’s not being talked about as much, but indexing has gotten a lot harder since the Helpful Content update. You’ll see a lot more pages in the “Discovered - currently not indexed” and “Crawled - currently not indexed” than you did previously because the bar is higher for what Google deems worth capturing from the web.

Slide 108

Slide 108 text

10 8 10 8 Google Wants to Crawl Even Less Gary Illyes has indicated that he wants to have Google crawl less. Search quality certainly cannot suffer, so crawlin has to get increasingly intelligent.

Slide 109

Slide 109 text

10 9 10 9 I Believe This is a Function of Information Gain Conceptually, as it relates to search engines, Information Gain is the measure of how much unique information a given document adds to the ranking set of documents. In other words, what are you talking about that your competitors are not?

Slide 110

Slide 110 text

11 0 110 In conclusion: “More content” is no longer inherently the most effective approach because there’s no guarantee of traﬃc from Google. Google’s sophistication won’t allow it.

Slide 111

Slide 111 text

11 1 11 1 The only content you should be making

Slide 112

Slide 112 text

11 2 11 2 I’m Leaving Y’all with Three Actions Today 1. How to Prune Your Content 2. How to Use LLMs to Generate Valuable Content 3. AI Tools to Use for SEO

Slide 113

Slide 113 text

11 3 The Content Pruning Process

Slide 114

Slide 114 text

11 4 11 4 Pruning and Optimization Work Quite Well Together

Slide 115

Slide 115 text

11 5 Aleyda Has a Process Aleyda’s workﬂow is a great place to work through whether your content should be pruned or not. https://www.aleydasolis.com/en/crawli ng-mondays/how-to-prune-your-website- content-in-an-seo-process-crawlingmon days-16th-episode/

Slide 116

Slide 116 text

11 6 11 6 We like automate to get to a Keep. Revise. Kill. (Review.)

Slide 117

Slide 117 text

11 7 11 7 Content Decay The web is a rapidly changing organism. Google always wants the most relevant content, with the best user experience, and most authority. Unless you stay on top of these measures, you will see traﬃc fall off over time. Measuring this content decay is as simple comparing page performance period over period in analytics or GSC. Just knowing content has decayed is not enough to be strategic.

Slide 118

Slide 118 text

11 8 11 8 It’s not enough to know that the page has lost traﬃc.

Slide 119

Slide 119 text

11 9 11 9

Slide 120

Slide 120 text

12 0 12 0 The Content Potential Rating (CPR).

Slide 121

Slide 121 text

12 1 12 1 Content Potential Score

Slide 122

Slide 122 text

12 2 12 2 Interpreting the Content Potential Rating 80 - 100: High Priority for Optimization 60 - 79: Moderate Priority for Optimization 40 - 59: Selective Optimization 20 - 39: Low Priority for Optimization 0 - 19: Minimal Benefit from Optimization If you want quick and dirty, you can prune everything below a 40 that is not driving significant traffic.

Slide 123

Slide 123 text

12 3 12 3 Combining CPR with pages that lost traﬃc helps you understand if it’s worth it to optimize.

Slide 124

Slide 124 text

12 4 12 4 Step 1. Pull the Rankings Data from Semrush Organic Research > Positions > Export

Slide 125

Slide 125 text

12 5 12 5 Step 2: Pull the Decaying Content from GSC Google Search Console is a great source to spot Content Decay by comparing the last three months year over year. Filter for those pages where the Click Difference is negative (smaller than 0) then export.

Slide 126

Slide 126 text

12 6 12 6 Step 3: Drop them in the Spreadsheet and Press the Magic Button

Slide 127

Slide 127 text

12 7 The Output is List of URLs Prioritized by Action Each URL is marked as Keep, Revise, Kill or Review based on the keyword opportunities available and the effort required to capitalize on them. Sorting the URLs marked as “Revise” by Aggregated SV and CPR will give you the best opportunities ﬁrst.

Slide 128

Slide 128 text

12 8 12 8 Get your copy of the Content Pruning Workbook : https://ipullrank.com/cpr-sheet

Slide 129

Slide 129 text

12 9 How to Kill Content Content may be valuable for channels outside of Organic Search. So, killing it is about changing Google’s experience of your website to improve its relevance and reinforce its topical clusters. The best approach is to noindex the pages themselves, nofollow the links pointing to them, and submit an XML sitemap of all the pages that have changed. This will yield the quickest recrawling and reconsideration of the content.

Slide 130

Slide 130 text

13 0 13 0 How to Revise Content Review content across the topic cluster Use co-occurring keywords and entities in your content Add unique perspectives that can’t be found on other ranking pages Answer common questions Answer the People Also Ask Questions Restructure your content using headings relevant to the above Add relevant Structured markup Expand on previous explanations Add authorship Update the dates Make sure the needs of your audiences are accounted for Add to an XML sitemap of only updated pages

Slide 131

Slide 131 text

13 1 How to Review Content The sheet marks content that has a low content potential rating and a minimum of 500 in monthly search volume as “Review” because they may be long tail opportunities that are valuable to the business. You should take a look at the content you have for that landing page and determine if you think the effort is worthwhile.

Slide 132

Slide 132 text

13 2 Using AI for your SEO

Slide 133

Slide 133 text

13 3 With AI, I’m giving y’all legos. What you build is up to you, but I’m going to show things to consider.

Slide 134

Slide 134 text

13 4 Setting Up LLMs Locally You don’t need ChatGPT anymore

Slide 135

Slide 135 text

13 5 13 5

Slide 136

Slide 136 text

13 6 LLaMa 3.1 is SOTA Facebook’s open source model is outperforming the best closed-source models on a variety of different evaluation metrics.

Slide 137

Slide 137 text

13 7 13 7

Slide 138

Slide 138 text

13 8 13 8

Slide 139

Slide 139 text

13 9 13 9

Slide 140

Slide 140 text

14 0 14 0

Slide 141

Slide 141 text

14 1 14 1

Slide 142

Slide 142 text

14 2 142 You can now unlock state of the art generative AI use cases from your laptop for free.

Slide 143

Slide 143 text

14 3 Make Sure You Hook It Up To Your GPU On a Windows machine you’ll need to go to the NVIDIA Control Panel and add the Ollama server application under Manage 3D Settings.

Slide 144

Slide 144 text

14 4 How to Use LLMs for Content and SEO

Slide 145

Slide 145 text

14 5 14 5 The Three Laws of Generative AI content 1. Generative AI is not the end-all-be-all solution. It is not the replacement for a content strategy or your content team. 2. Generative AI for content creation should be a force multiplier to be utilized to improve workﬂow and augment strategy. 3. You should consider generative AI content for awareness efforts, but continue to leverage subject matter experts for lower funnel content.

Slide 146

Slide 146 text

14 6 14 6 Think back to 10 Minutes Ago - Retrieval Augmented Generation

Slide 147

Slide 147 text

14 7 14 7 It’s Not Diﬃcult to Build with Llama Index sitemap_url = "[SITEMAP URL]" sitemap = adv.sitemap_to_df(sitemap_url) urls_to_crawl = sitemap['loc'].tolist() ... # Make an index from your documents index = VectorStoreIndex.from_documents(documents) # Setup your index for citations query_engine = CitationQueryEngine.from_args( index, # indicate how many document chunks it should return similarity_top_k=5, # here we can control how granular citation sources are, the default is 512 citation_chunk_size=155, ) response = query_engine.query("YOUR PROMPT HERE")

Slide 148

Slide 148 text

14 8 148 Everyone can code now.

Slide 149

Slide 149 text

14 9

Slide 150

Slide 150 text

15 0 15 0

Slide 151

Slide 151 text

151 PAGE GENERATIVE AI PRODUCTIVITY USE CASES RAG opens up a series of generative AI use cases that work well for your situation. Briefing & Business Cases Content Analysis First-pass Brand Review First-pass Legal Review Content First Draft Keyword Insertion Structured Data Generation Link Identification & Insertion Generating Voiceovers Generating Images Generating Videos Writing Code

Slide 152

Slide 152 text

15 2 @BritneyMuller’s Guide to Using Colab Britney talked about how easy it is to use Colab with Python. Now it’s even easier to using LLMs. https://github.com/BritneyMuller/colab- notebooks?tab=readme-ov-ﬁle

Slide 153

Slide 153 text

15 3 Just describe what you want You can tell your language model what you want the code to do and it will handle the rest. If it doesn’t work, just describe what went wrong or paste the error and it will ﬁx it for you. In this example my prompt is: {write python code for colab that takes a csv file of keywords and using bertopic with the chatgpt to compute the natural language topics for each row.}

Slide 154

Slide 154 text

15 4 15 4 Colab + Gemini

Slide 155

Slide 155 text

15 5 15 5 Llama Index - RAG - https://www.llamaindex.ai/

Slide 156

Slide 156 text

15 6 15 6 LangChain - Build Agents - https://www.langchain.com/

Slide 157

Slide 157 text

15 7 15 7 LangFuse - Prompt Management & Observability tool - https://langfuse.com/

Slide 158

Slide 158 text

15 8 15 8 Literal AI

Slide 159

Slide 159 text

15 9 159 You don’t need to code for any of this.

Slide 160

Slide 160 text

16 0 16 0

Slide 161

Slide 161 text

16 1 16 1

Slide 162

Slide 162 text

16 2 16 2

Slide 163

Slide 163 text

16 3 163

Slide 164

Slide 164 text

16 4 164

Slide 165

Slide 165 text

16 5 165

Slide 166

Slide 166 text

16 6 16 6 Bubble - No Code Apps - https://bubble.io/

Slide 167

Slide 167 text

16 7 Integrate Promptitude with Zapier or Make

Slide 168

Slide 168 text

16 8 Prompts You Need To Write ChatGPT is very effective at doing the following SEO related tasks: Page Title writing Meta Description writing Keyword Insertion Link Insertion You should use your own prompts for these though so you don’t copy other people’s patterns.

Slide 169

Slide 169 text

16 9 Page Titles Feature Token Count features Hypothesis There’s no hard max page title length indicated in the attributes so we can test lengths longer than the 60-70 characters to determine impact.

Slide 170

Slide 170 text

17 0 Page Title Test Hypothesis: A page title that’s longer than the standard best practice will negative impact rankings for primary keyword target. Variables: Control Short page title Long page title Metrics: Ranking Increase

Slide 171

Slide 171 text

17 1 lastSigniﬁcantUpdate lastSigniﬁcantUpdate - The date of the last time Google encountered the page as materially updated. Feature Hypothesis Making substantial updates to pages regularly yields improved crawl activity and more opportunities to rank better.

Slide 172

Slide 172 text

17 2 17 2 TL;DR Generation for Signiﬁcant Updates

Slide 173

Slide 173 text

17 3 Test Structure The goal of this test is to determine how much content is considered a “signiﬁcant update” that yields crawl activity. Create control and variants pages testing the length of added content: We measure the impact on organic traﬃc in order to capture changes to rankings and/or changes to clickthrough rate.

Slide 174

Slide 174 text

17 4 Let Screaming Frog Do the Heavy Lifting Generate embeddings while you crawl using Screaming Frog SEO Spider. Take the ﬁle to Colab and do the following things: Keyword - Landing Page Relevance Scoring Keyword Mapping Link Building Target Identiﬁcation Redirect Mapping Internal Link Mapping https://ipullrank.com/vector-embedding s-is-all-you-need You can also work with your language model to combine crawl data with SERP data and do things like information gain calculations.

Slide 175

Slide 175 text

17 5 17 5 Brief Writing

Slide 176

Slide 176 text

17 6 17 6 Persona Modeling based on Desk Research Data

Slide 177

Slide 177 text

17 7 17 7 Persona Writing with SparkToro

Slide 178

Slide 178 text

17 8 17 8 Brand Voice and Tone Review

Slide 179

Slide 179 text

17 9 17 9 Schema Markup Generator https://chatgpt.com/g/g-MvH0WHO3e-schema-markup-generator-gpt

Slide 180

Slide 180 text

18 0 18 0 Taskade - AI All-in-One - https://www.taskade.com/

Slide 181

Slide 181 text

18 1 18 1 AIPower - All-in-One for WP - https://aipower.org/

Slide 182

Slide 182 text

18 2 18 2 Thunderbit - Build No Code AI Automation Tools - https://thunderbit.com/

Slide 183

Slide 183 text

18 3 18 3 Keyword Insights - Keyword Clustering Tool

Slide 184

Slide 184 text

18 4 18 4 Octoparse - Combine a scraper with Generative AI - https://www.octoparse.ai/

Slide 185

Slide 185 text

18 5 18 5 DejanSEO’s LinkBERT - https://www.linkbert.com https://dejanmarketing.com/tools/li nkbert/

Slide 186

Slide 186 text

18 6 18 6 InLinks - Automated Internal Links - https://inlinks.com/

Slide 187

Slide 187 text

18 7 18 7 Link Whisper - Internal Link automation - https://www.linkwhisper.com

Slide 188

Slide 188 text

18 8 18 8 Respona - AI-enabed Link Building - https://www.respona.com

Slide 189

Slide 189 text

18 9 18 9 SEOJuice - https://seojuice.io/

Slide 190

Slide 190 text

19 0 19 0 UX Sniff - A AI-enabled HotJar https://uxsniff.com/

Slide 191

Slide 191 text

19 1 Roll the Credits

Slide 192

Slide 192 text

19 2 19 2 What you should know and do to win Google is still the primary show in town Relevance is a quantitative measure GenAI works on the same math as search engines Focus on making your chunks for relevant to rank in GenAI Search Improve UX to drive more long clicks Improve web search factors and local search factors Focus on content your audience wants, prune what they don’t Use RAG to generate content with AI Embrace AI tools to improve your workﬂows and your ability to test

Slide 193

Slide 193 text

19 3

Slide 194

Slide 194 text

Thank You | Q&A [email protected] Award Winning, #GirlDad Featured by Get Your AIO Threat Report: https://ipullrank.com/AIO-report Play with Raggle: https://www.raggle.net Download the Slides: https://speakerdeck.com/ipullrank Mike King Chief Executive Oﬃcer @iPullRank

Slide 195

Slide 195 text

195