SearchLove London 2016 | Larry Kim | Hacking RankBrain and Other Machine Learning Algorithms: 5 SEO Weapons to Survive SEO Judgement Day’
- 1. HACKING RANKBRAIN & MACHINE LEARNING-ENABLED ALGOS: The 5 SEO Strategies You’ll Need to Survive SEO Judgment Day
- 2. AN AVERAGE DAY IN THE NOT TOO FAR OFF FUTURE…
- 3. BOOM!
- 4. SEO JUDGEMENT DAY: Google Core Search Algorithms All Based on Machine Learning
- 5. Algos for Spam Detection, Image Search, Video Search, Query Interpretation, Core Ranking Algo (etc.) All Machine-Learning Enabled
- 8. JOHN CONNOR, Leader of the SEO Rebellion in 2020.
- 11. GOOGLE SEARCH TODAY: HEURISTIC ALGORITHMS Not Entirely Machine Learning Based (Yet)Google Algorithm Change History
- 12. • Machine Learning / AI- Powered • Checks if the search results met user expectations • Make adjustments to query & rankings based on the outcome Google Search In The Future: MACHINE-LEARNING BASED
- 13. WHERE DOES GOOGLE USE MACHINE LEARNING?
- 14. Ran
- 15. Google Cloud Vision API Google Cloud Speech API Google Cloud Translate API
- 17. Certain Search Verticals Like Porn?
- 18. Google AdWords Entirely Based on Machine Learning for Ad Targeting
- 19. FACEBOOK NEWS FEED & AD TARGETING
- 20. TWITTER ADS USE MACHINE LEARNING
- 21. Machine Learning Systems Pretty Much EVERYWHERE
- 22. Core Algo Changing? Validation Signals? WHO CARES! GET IN!
- 24. Why Didn’t They Just Send Rand?
- 26. Three Random Facts ABOUT ME
- 27. 1. Expert Drone Pilot
- 28. 2. STUDIED ELECTRICAL ENGINEERING FOR 5 YEARS
- 29. 3. Built Suite of Long Tail Keyword Research Tools
- 31. Google: “RankBrain Used on Long Tail Queries” ... “Query Interpretation ... That Changes Rank”
- 32. Last Year: RankBrain as It’s Own Thing? *Note: Not Official Google Diagram. The CORE Algorit hm! Panda Pengui n RankBra in (New!) The CORE Algorit hm! Panda Pengui n
- 33. The CORE Algorit hm! The Future: RankBrain as a Thing That Works Across Multiple Things *Note: Not Official Google Diagram. Panda Pengui n RankBrain (New!)
- 34. How Does RankBrain REALLY Work??
- 35. Long Tail Query Entered RankBrain Makes a Guess and Re-Writes The Query into Something it Recognizes (Query Interpretation) Did The Result Satisfy The User? Great. Next Time I See Queries Like This, I’ll Put This Page at the Top. Larry’s Greatly Simplified RankBrain Diagram* *Note: Not Official Google Diagram. System Doesn’t Update in Real-Time (yet). Rats. Next Time I See a Query Like This, I’ll Try Some Other Page
- 37. Organic Search Ranking What’s A Good Click Through Rate for Organic Search? (Spoiler Alert: It’s in Flux)OrganicSearchClickThroughRate
- 38. Organic Search Ranking What’s A Good Click Through Rate for Organic Search? (Spoiler Alert: It’s in Flux)OrganicSearchClickThroughRate CTRs Trending Higher in Top Positions CTRs Trending Lower in Lower Positions
- 39. Organic Search Ranking What’s A Good Click Through Rate for Organic Search? (Spoiler Alert: It’s in Flux)OrganicSearchClickThroughRate CTRs Trending Higher in Top Positions CTRs Trending Lower in Lower Positions
- 41. THE CTR vs RANKING CURVE is BENDING! WHY? Theory Impossible Because Elimination of Right Side Rail Ads? That Happened in February
- 42. THE CTR vs RANKING CURVE is BENDING! WHY? Theory Impossible Because Elimination of Right Side Rail Ads? That Happened in February Increase in Featured Snippets? That would shift the curve outwards, not bend curve.
- 43. THE CTR vs RANKING CURVE is BENDING! WHY? Theory Impossible Because Elimination of Right Side Rail Ads? That Happened in February Increase in Featured Snippets? That would shift the curve outwards, not bend curve. Change in Mobile? 78% Desktop / 22% Mobile (+/- 0.5%)
- 44. What Came First: The CTR or the Ranking??
- 45. +/- 3% Increase in Post Engagement +/- 1 Spot in Organic Position RankBrain Rewards & Penalties?
- 46. Outputs are The New Inputs!
- 47. A Simple CTR Optimization Example
- 48. Behold The Power of CTR Optimization WHAT? Page Title CTR Rank Guerrilla Marketing: 20+ Examples & Strategies to Stand Out 1% 8 20+ Jaw-Dropping Guerrilla Marketing Examples 4.19% 5
- 50. Key Takeaway: ML (in General) Makes Stuff Increasingly Winner-Take-All. Larry’s RankBrain Theory: High CTR Long Tail Rankings Get *Additional* Bonus Ranking Boost
- 51. If True, What Should SEOs Do To Prepare For RankBrain and Other ML- Enabled Algos?
- 52. They’re Not Going To Tell You Everything…
- 53. He’s Delusional. Do Nothing.
- 54. We Need To Crush More Auditions.
- 55. New ML-Algos Tougher Than Previous Ones…
- 56. Five SEO Weapons to Survive SEO Judgement Day
- 57. LARRY’S #1 SEO WEAPON The CTR Donkey Detector!
- 59. DOWNLOAD THIS QUERY DATA
- 60. What’s a Good Click Through Rate?
- 61. CTR Study (2014) by. Philip Petrescu for Moz.
- 63. Use Your Own Data To Plot The Average CTR for YOUR Industry! Unicorns Donkeys CTR vs. Average Position
- 65. Unicorns are Up To 6x Better Than Donkeys. Avg. CTR Donkey CTRs (Bottom 10%) = 3x Worse Than Avg. Unicorn CTRs (Top 10%) = +2x Better Avg.
- 66. DONKEYS DETECTED. How to Convert to Unicorn??
- 67. Larry’s #2 SEO Weapon: DONKEY TO UNICORN CONVERTER
- 68. These SEO Title Tags All Suck:
- 70. Keyword Headlines Using Dynamic Keyword Insertion Do OK, Not Great Relative Frequency Average CTR Top 15% Top 5% Top 1% Relative CTR Ads without DKI Ads with DKI
- 71. Keyword-y Headlines Less Likely To Be UnicornsRelative Frequency Top 5% Ads without DKI Ads with DKI Top 1% Relative CTR
- 73. Leverage Emotional Triggers I Know Now Why You Cry…
- 74. The Top 9 Emotions That Make People Click Like Crazy: Step 1: Pick One of These Emotional Triggers • Laughter • Amusement • Curiosity • Awe • Anger • Fear • Joy • Empathy • Sadness
- 75. Step 2. Write Headline Copy From Perspective of One of These Personas The Bearer of Bad News The Hero/Villai n The Comedian The Feel Good Friend
- 76. 3. Use This Title Template Commonly Employed By ‘Viral’ Articles Format List post, quiz, Infographic Emotional Hook Emotional word or superlative Content Type Images, quotes, pictures, facts Topic Love, cats, dogs, fitness, health, Donald Trump 5 AWESOME SEO STRATEGIES TO DEFEAT RANKBRAIN
- 77. Unicorn Headlines = Top 10% Odds of Finding Unicorn = 1:10 So Test 10 Different Headlines
- 78. Quiz: How Many Headlines Do You See Here?
- 79. It’s the Same Headline with Different Capitalization, Punctuation Are We Learning Yet??
- 80. You Need 10 *Different* Headlines! How to Spot the Unicorn?
- 81. Use AdWords to Test Headlines
- 82. Bid to Position (eg: Top Spot)
- 83. OK To Target Cheaper Markets Average Cost Per Click Around The World Greater than US Average CPC Average US CPC 1-20% Less than Average US CP 21-40% Less than Average US CPC41-60% Less than Average US C 61-80% Less than Average US CPC>81% Less than Average US CPC Limited Data
- 84. Use Broad Match Keyword Match Type
- 85. • Identify Donkeys (Bottom 10%) • Write 10 Different Headlines • Test 10 Ads on AdWords • Replace Old Headline With Winner Larry’s Donkey to Unicorn Converter: Summary
- 86. The Same Emotions That Make People to Share Things Cause High CTR.
- 87. elative Search CTR vs. Facebook Post Engagement Rat
- 88. Social Signals Were Suspected Ranking Factors A Long Time Ago!
- 89. Meaning: CTR As Ranking Factor Has Been In Place For a Long Time!
- 90. "In addition, click data (the website links on which a user actually clicks) is important for evaluating the quality of the search results page.” Udi Manber, Google's former chief of search quality, Testimony to US Federal Trade Commission, Obtained by The Wall St. Journal, August 8, 2012
- 92. 3. The Grand Unified Unicorn Quaternary Theory: They All Have High Engagement in Common SEO Unicorn Paid/Organic Social Unicorn CRO Unicorn PPC Unicorn
- 93. 4. Bonus (New) Featured Snippets!
- 94. Snippet Position CTR most expensive google adwords 1.1 0.6939 weight loss keywords 1 0.5595 insurance keywords 1 0.5408 most expensive adwords 1 0.5303 meta tags for seo 1.1 0.4768 how to increase ctr 1.5 0.4556 best call to action buttons 1.1 0.4503 meta tag for seo 1.4 0.4481 marketing buzz words 1 0.434 what is a good click through rate 1.1 0.4127 Theory: Unusually High CTR Appears to Play a Major Role in Snippet Selection. Beat Expected CTR by 50%
- 96. A Note of Caution
- 97. Don’t Use Spam Sites Like CrowdSearch.me
- 99. We’ve Got RankBrain By The Balls
- 100. They Have a Back-Up System…
- 101. We’ve Got Rankbrain By The Balls Seems to Have a Back-Up System… We’re Gonna Need Bigger Guns..
- 102. LARRY’S #3 SEO WEAPON Engagement Rate Unicorn/Donkey Detector
- 103. Bounce Rate vs. Organic Position Pass: Bounce Rate Low Enough Fail: Bounce Rate Too High
- 104. We Know Google Measures Dwell Time
- 105. We Know Google Measures Dwell Time
- 106. “Back” On Mobile Triggers “People Also Search For”
- 107. Time on Site vs. Organic Position PASS: Good Enough Time On Site Fail: Time on Site not Good Enough
- 108. Key Takeaway: “Task Completion Rates” Used To Validate CTR
- 109. Engagement Rate Donkey Detector
- 110. Search Conversion Rate Distribution Median: 2.35% Top 25%: 5.31% and Above Top 10%: 11.45% and Above SHAREOFALLACCOUNTS SEARCH CONVERSION RATE. (WordStream Customer Data – Not Official Google Data) What’s a Good Conversion Rate?
- 111. How to Turn Conversion Rate Donkeys into Unicorns?
- 112. Larry Kim, Founder WordStream, Inc. SearchLove London, October 17, 2016 #searchlove@larrykim
- 113. Unicorn CRO! 1. On-Page Elements: UX, Design, Images, Copy, Fonts, Emotional Triggers, Psychology, Etc. 2. Brand Awareness: People overwhelmingly favor the brands they're familiar with. 3.Growth Hack: Engineer an offer to be 10x more compelling and pull uses through the funnel 10x faster. Key Takeaway: Focus on All Three Areas! #searchlove@larrykim
- 115. Time Travel to Past: Before Users Do Searches
- 116. Larry’s #4 SEO Weapo FACEBOOK ADS and Google Display Network
- 117. Wait.. Advertising Impacts Organic CTR and Conversion Rates??
- 118. Brand Affinity DRAMATICALLY Impacts CTR (+2-3x!)
- 119. Brand Affinity DRAMATICALLY Impacts CONVERSION RATES (+2-3x!)
- 121. Warning: Facebook Ads Employ Machine Learning Tech: So You Can’t Promote Junk
- 122. Interests Targeting
- 124. Behavioral Targeting
- 125. Custom Affinity Audiences in Google Display Network
- 126. In Google Display Network Targeting, Set Keyword Setting to “Content”
- 127. Clone Existing Customer Lists
- 128. Time Travel to FUTURE: Ensure Subsequent Searches Go Your Way
- 129. I’LL BE BACK Larry’s #5 SEO Weapon: Remarketi ng
- 130. LovePop Cards:
- 133. Unicorn Video Ads!
- 136. Returning Visitors: 1. Existing Customers 2. 1-Day Remarketing Audience 3. 7, 30, 180-day Remarketing Audiences New Visitors: 1. Lookalike Audiences 2. Interest, Behavior & Demographic- Based Audiences. Note: Exclude the overlap (duh). arry’s Ad Targeting Stack (Powered By Tim Travel)
- 137. Delete Your “Bad Neighborhoods” Self Destruct Sequence Initiated
- 138. Mission Complet e
- 139. What Does it All Mean?
- 140. Old SEO Tactics Clearly Don’t Work Anymore
- 141. They Might Even Make Things Worse!
- 142. Don’t Just Wait for RankBrain & ML-Algos.
- 143. Better CTR & Conversion Rates -> More Clicks & Conversions. Improved CTR & CVR -> Better Rankings. Better Rankings -> Even more clicks and Virtuous Cycle of Unicorn Land:
- 144. Terrible CTR & Engagement Rates -> Less Clicks, Leads, Sales, etc. Crappy User Engagement -> Terrible Rankings. Terrible Rankings -> Even Fewer Clicks & Sales. Vs. Death Spiral of RankBrain and Other ML- Algos
- 145. How Marketing REALLY Works STEP 1: Promote Inspirational / Memorable Content about Your Brand to your Target Market STEP 2: People See The Ad, But Don’t Necessarily Take Action Right Away. (But Become Biased) STEP 3: Later when the Need Arises, People either: Do a Branded Search for Your Stuff Do Un-Branded Search but Biased Towards Clicking & Buying From You.
- 146. Be a Unicorn in a Sea of Donkeys
- 147. Thank You, Will Critchlow & SearchLove Team!