SEO in 2013, 2023, and 2033 - The More Things Change...

Slides from my talk at Friends of Search 2023 where I looked back at the past 10 years of search and SEO, and make some predictions about where the industry is heading in the next decade.

Transcript

1. SEO in 2013, 2023, and 2033:
   The more things change…
   Barry Adams
   March 2023
   

   View Slide

2. View Slide

3. View Slide

4. View Slide

5. View Slide

6. SEO in 2013
   • Post-Panda SEO;
   ➢ Thin & low quality content wasn’t rewarded
   ➢ ‘Great content’ started to make sense
   • Penguin in full swing;
   ➢ Targeting link schemes & unnatural links
   ➢ Penalises keyword stuffing & over-optimisation
   • Hummingbird update confirmed in Sept 2013;
   ➢ Full query text considered for rankings
   ➢ From ‘keywords’ to ‘topics’
   • Knowledge Graph in its infancy;
   ➢ Early stages with limited SERP impact
   

   View Slide

7. SEO in 2013
   • Post-Panda SEO;
   ➢ Thin & low quality content wasn’t rewarded
   ➢ ‘Great content’ started to make sense
   • Penguin in full swing;
   ➢ Targeting link schemes & unnatural links
   ➢ Penalises keyword stuffing & over-optimisation
   • Hummingbird update confirmed in Sept 2013;
   ➢ Full query text considered for rankings
   ➢ From ‘keywords’ to ‘topics’
   • Knowledge Graph in its infancy;
   ➢ Early stages with limited SERP impact
   Business as usual
   (more or less)
   

   View Slide

8. 2015:
   RankBrain
   

   View Slide

9. RankBrain
   • Enables relevant SERPs for unfamiliar queries
   • First time Google uses Machine Learning in search;
   ➢ Reluctantly
   • It opened the door for what came in 2018…
   

   View Slide

10. Amit Singhal
    

    View Slide

11. John Giannandrea
    Amit Singhal
    

    View Slide

12. 2018:
    The first ‘core algorithm update’?
    

    View Slide

13. 2018–now: The era of ‘Core Algorithm Updates’
    • Machine Learning is increasingly driving SERP rankings;
    ➢ Less reliance on manually crafted ranking factors
    ➢ BERT & MUM are ML systems working on SERPs
    • Every core update is an improved ML model;
    ➢ Constant training on feedback provided by human quality raters
    • ML can do what manual ranking algorithms can’t;
    ➢ Deal with the insane chaos that is the World Wide Web
    

    View Slide

14. But what about?
    ✓ Mobile-first Indexing
    ✓ HTTPS
    ✓ Core Web Vitals
    

    View Slide

15. But what about?
    ✓ Mobile-first Indexing
    ✓ HTTPS
    ✓ Core Web Vitals
    Social Engineering Updates
    Intended to change WWW behaviours
    Mostly irrelevant to the bigger picture
    

    View Slide

16. Google 2013-2023: Three Key Trends
    1. Machine Learning & AI;
    ➢ Moving away from manual factors
    

    View Slide

17. View Slide

18. View Slide

19. View Slide

20. View Slide

21. Ads
    Local pack
    People also ask
    Organic results
    Related searches
    Knowledge graph
    Related entities
    

    View Slide

22. SERP Features
    

    View Slide

23. Google 2013-2023: Three Key Trends
    1. Machine Learning & AI;
    ➢ Moving away from manual factors
    2. SERP Features;
    ➢ Increased diversity in ranking types
    

    View Slide

24. SERP Features & Knowledge Graph
    

    View Slide

25. Knowledge Graph
    Arnold
    Schwarzenegger
    Bodybuilding
    Terminator films
    Governor of
    California
    Maria Shriver
    Ronnie
    Coleman
    

    View Slide

26. Knowledge Graph
    

    View Slide

27. Knowledge Graph
    • Topical Authority
    • Authorship
    • E-E-A-T
    Expressions of the
    Knowledge Graph
    

    View Slide

28. Knowledge Graph
    • Underpins most Google SERP features
    • Enables Google to understand entities & relationships
    • Empowered by Structured Data
    • Critical aspect of Google search quality
    

    View Slide

29. Google 2013-2023: Three Key Trends
    1. Machine Learning & AI;
    ➢ Moving away from manual factors
    2. SERP Features;
    ➢ Increased diversity in ranking types
    3. Knowledge Graph;
    ➢ Not keywords, but entities & relationships
    

    View Slide

30. Generative AI
    

    View Slide

31. View Slide

32. ChatGPT / Bard = LLM
    • Large Language Models
    • Basically predictive text on steroids
    

    View Slide

33. LLM’s key problem
    

    View Slide

34. LLM’s key problem
    Industrial Grade
    Bullshit Generators
    

    View Slide

35. The solution..?
    

    View Slide

36. The solution
    

    View Slide

37. Knowledge Graph
    The solution
    

    View Slide

38. Search in 2033..?
    1. Machine Learning & AI;
    ➢ Black box systems
    2. SERP Features;
    ➢ Powered by AI and KG
    3. Knowledge Graph;
    ➢ Entities & relationships, accuracy
    

    View Slide

39. View Slide

40. View Slide

41. 2033?
    

    View Slide

42. Insane* Predictions
    • Conversational search;
    ➢ Not queries, but conversations
    • Predictive search;
    ➢ Your device will give you the answer before you ask
    • Headless websites;
    ➢ No front-end, just data for search engines
    *Probably, maybe
    

    View Slide

43. Thank You
    barry@polemicdigital.com
    @badams
    /in/barryadams/
    

    View Slide