Visibility_14: The Search (R)Evolution

763 views

Published on

Searchmetrics CTO Marcus Tober presents The Search (R)Evolution. His keynote presentation at Visibility_14 in Chicago, IL on July 17, 2014 looked at "Ranking Factors after Hummingbird, the Panda Update and the Future of Search".

Published in: Education
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
763
On SlideShare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
17
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide
  • There are severeal examples for the misintpretation of correlation as causation.
    There is the phenomenon, that the amount of storks living in certain rural areas correlates with a higher birth rate. Would you suggest that a higher number of storks causes the birth of more babies?
    In summer, people like to go to the beach and get tanned. Also, they like to eat ice cream more often when the sun is shining. Would you appreciate the conclusion that ice cream causes sunburn?
    We at Searchmetrics are aware of the fact that correlation is not causation. But our analyses based on a huge amout of data allow us to look at similarities and differences of rankings and URLs and our expertise lets us interpret the data.
  • What are the most important „factors“ for good rankings? Again: 10.000 Informational Search Keywords. Top 30

    Particulary after Hummingbird, it‘s about relevant Content – not the best optimized Content.
  • Low correlation, but in fact, nearly every URL in the Top 30 has a Description -> seems obligatory!
  • Consistently high average values. 2014 nearly 100% - even encreased from last year. Seems a „Must-Have“.
    Correlation even decreased slightly, but only because even more URLs in Top 30 have this Feature and therefore, the differences become even smaller.
  • Average count of internal links. Again: Brand Factor. Interesting to see, that better ranking URLs have slightly more internal Links. That means: Housekeeping and Optimization of internal Link Structure is important
  • Average count of internal links. Again: Brand Factor. Interesting to see, that better ranking URLs have slightly more internal Links. That means: Housekeeping and Optimization of internal Link Structure is important
  • Page load Time is a very important performance factor. Google is pushing it for your and obvivously many websites got faster. Influences User Signals (CTR, BounceRate, etc) and User Happiness! – Correlation increased significantly!
  • Average count of internal links. Again: Brand Factor. Interesting to see, that better ranking URLs have slightly more internal Links. That means: Housekeeping and Optimization of internal Link Structure is important
  • Do you like this kind of remote remote control? I don‘t.

    Less bells and whistles – more focus
  • Content alone is not enough. The technical environment as well as Site Architecture have to be at a top level to achieve good rankings. But: Content is still the most important thing.

    But it must be unique content with a benefit for the user.
  • Brand Factor - „Keyword in Descritpion“ seems a rather rampant feature (weit verbreitet)
  • Brand Factor - „Keyword in Descritpion“ seems a rather rampant feature (weit verbreitet)

    Last year: higher average of pages had KW in Description. Brand Factor was not that strong a year ago
  • On average increased per Pos. about 12.5% (in comparison to 2013). Very interresting
    There seems to be the trend to have more and longer content. Is it a disadvantage to be short?
  • Seems as if this would‘nt really work. As you know, Tom Cruise is a rather short guy. But it‘s about providing holistic Content. And this content is likely to be a little longer – not as short as Tom Cruise. Just did not fit for Katie Holmes.
  • Word Count increased about 20% on average (compared to 2013)
  • Hummingbird was the biggest update of the Google algorithm in the past 10 years. But nobody really recognized. Why? Because Google implemented methods in the algorithm to understand the meaning of the query and therefore they must understand the meaning of the content.
  • Analysis of (Proof and) Relevant Keywords in the Top Results with Searchmetrics Content Optimization.
    Keywords with high Relevancy for the Query (Focus Keyword) = Relevant Keywords
    Also: Keywords being used by most of the Top-URLs for that Query = Proof Keywords (of course: correlation is lower)
  • Better ranking URLs contain more relevant Terms around the Fokus-keyword -> Holistic Documents rank better!

    Top results contain significantly more „relevant“ terms than lower-ranked results
  • Top-Rankings containn significantly more „Proof“-Terms
  • The distribution of Keywords and further relevant Terms compared to the absolute number of terms, helps Google‘s Algorithms to understand the topic and intention od texts. Googles Word2Vec Project is open source.
  • Natural Language

    Keywords become more Complex (Questions, hole sentences)
  • Ca. 30 - 40% smaller
  • Texts are shorter
  • Differences often regarding Keywords out of Movie-, Music- and/or Games-Sector. Examples (based on URL-comparison):

    Indiana jones 5 (only 40% identical)
    Simcity 4 mac (only 42.5% identical)
    Packers game (only 40% identical)
    mumford & sons - "i will wait“ (only 23.53% identical)
  • Every word of a query, and more important, there sematic meaning, play a role
  • In the Movie: „Food Replicator“
    Machine reads your mind, figures out what you want to eat right know and delivers the desired dish.
  • Google also tries to read the user‘s mind in interpreting the intention behind a query -> then delivering a knowledge graph as the „desired dish“ above / nerxt to the organic SERP.
  • Example. „Italian restaurant san francisco“ / „italian lunch san francisco“ / „pizza nearby“ (local search in SF) / „eat pizza in san francisco“ / „san francisco pizza inn“ … etc.
  • Example. „Italian restaurant san francisco“ / „italian lunch san francisco“ / „pizza nearby“ (local search in SF) / „eat pizza in san francisco“ / „san francisco pizza inn“ … etc.
    BUT: of course, the SERP is influenced by factors such as search history, user specific cookies, location, logged in/ logged out etc.
  • So, what is good, what is bad? What is too much? Searchmetrics tells you.
  • Achtung: Dass, die Kurvenverläufe von URLs und Domains sich teilweise stark unterscheiden, liegt auch daran, dass die Anzahl jeweils abweicht. URLs: ca. 3.5 Mio
    Domains: ca. 500.000
  • informational keyword
  • More Focussed. -> Relevance …. Precise. Fast.

    Diversity decreases – more Competition !
  • So, what is good, what is bad? What is too much? Searchmetrics tells you.
  • Examples for SERPs for Keyword Pairs -> Before Hummingbird

    Universal Search Results removed

    (Bang = Pony)
  • It‘s about relevant, related topics and terms. The Content is the „Face“ of a page being the most important thing, the centre. It‘s immensely important, to provide unique and high-class content being perfectly tailored for your users. That means, the content must not be always the same. It‘s rather the opposite. Because you have to be distinguishable from your cometition. Content has to be holistic and satisfy the user‘s intention.
  • Number of Backlinks as such has increased! (Although correlation declined slightly)
  • Absolute Number (average values) actually decreased! Penguin?
  • Immensly increased (compared to 2013). Better rankign URLs have more links from pages with a high SEO Visibility
  • Backlinks of the last 30 Days to Homepage of respective URL
    NEW Definition of the Brand-Factor in 2014: 1. Small (niche) Brand / 2. Big Brand / 3. Wikipedia (universal Brand)
    -> URLs on second half of SERP 1 (Pos 4/5 to 8/9) seem to have more fresh Backlinks (directed to Homepage) -> Big Brands
    Obviously, (Niche) Brands get significantly fewer new Links to their Homepage, compared to pages ranking behind Pos 4. In contrast, these (Brand)pages have significantly more and better links
  • Share of „Brandlinks“. Z.B. spiegel.de / wsj.com

    URLs with more Brandlinks rank better.
  • Z.B. Spiegel Online / Wall Street Journal (Share of the most frequent Anchor Text, being not the exact Domain Name with TLD compared to all Anchor Texts of the URL)
  • Brands have more links, better links, more brandlinks – and they „need“ fewer fresh and/or news links.
    Your product must be good!
  • So, what is good, what is bad? What is too much? Searchmetrcis tells you.
  • Correlations calculated on Top 15 (not Top 30)
  • Low correlation. Top rankings have some 20 to 40 seconds more
  • Quiet stable average values over all positions. Pos 1-5 perform slightly better.
  • Very high correlation. Actually the highest, we have ever measured.
  • So, it‘s actually obvios. What must be the goal of SEO for Content?
  • Increase traffic for existing content (through content updates, more relevant content)

    New content performs better because more relevant.
  • Visibility_14: The Search (R)Evolution

    1. 1. The Search (R)Evolution Ranking Factors after Hummingbird, the Panda Update and the Future of Search 2014 MARCUS TOBER 07/17/2014 Chicago
    2. 2. Marcus Tober
    3. 3. Once upon a time… SEO
    4. 4. SEO Censored
    5. 5. 2014 SEO
    6. 6. Differences of URLs from #1 to #30?
    7. 7. Correlation ≠ Causation
    8. 8. 2014 Edition
    9. 9. IRON MAN RANKING
    10. 10. Techniques Content
    11. 11. RANKING FACTORS 2014
    12. 12. RANKING FACTORS 2014 vs 2013 Content/ Technics Social
    13. 13. OnPage TECHNIQUES
    14. 14. TECHNICAL ELEMENTS RANKING FACTORS 2014
    15. 15. vs 2013 TECHNICAL ELEMENTS RANKING FACTORS 2014
    16. 16. 35 www.searchmetrics.com 7/23/2014 ~98%𝜌2014 = 0.05 Description existing
    17. 17. 36 www.searchmetrics.com 7/23/2014 ~95% 𝜌2013 = 0.06 𝜌2014 = 0.05 Description existing vs 2013 ~98%
    18. 18. 37 www.searchmetrics.com 7/23/2014 Sitespeed 𝜌2014 = 0.14
    19. 19. 38 www.searchmetrics.com 7/23/2014 Sitespeed vs 2013 𝜌2014 = 0.14
    20. 20. 40 www.searchmetrics.com 7/23/2014 Brand Factor 𝜌2014 = 0.16 Number of internal Links
    21. 21. 41 www.searchmetrics.com 7/23/2014 Number of internal Links
    22. 22. Less is More!
    23. 23. What‘s even more important?
    24. 24. What‘s inside! CONTENT
    25. 25. CONTENT ELEMENTS RANKING FACTORS 2014
    26. 26. vs 2013 CONTENT ELEMENTS RANKING FACTORS 2014
    27. 27. 50 www.searchmetrics.com 7/23/2014 Keyword in Description ~7.500𝜌2014 = 0.13 Brand Factor ~55%𝜌2014 = 0.01
    28. 28. 51 www.searchmetrics.com 7/23/2014 𝜌2013 = 0.02 Keyword in Description vs 2013 ~55%𝜌2014 = 0.01 ~62%
    29. 29. 52 www.searchmetrics.com 7/23/2014 Brand Factor ~7.500𝜌2014 = 0.13 Text Length in Characters
    30. 30. 53 www.searchmetrics.com 7/23/2014 ~6.700𝜌2013 = 0.11 Text Length in Characters vs 2013 ~7.500𝜌2014 = 0.13
    31. 31. 55 www.searchmetrics.com 7/23/2014 Brand Factor ~650𝜌2014 = 0.13 Word Count
    32. 32. 56 www.searchmetrics.com 7/23/2014 ~540𝜌2013 = 0.11 vs 2013 Word Count ~650𝜌2014 = 0.13
    33. 33. „precise and fast“
    34. 34. GOOGLE HUMMINGBIRD UPDATE Contextual Search Conversational Search Semantic Search 7/23/201459 www.searchmetrics.com
    35. 35. Contextual Search1.
    36. 36. 65 www.searchmetrics.com 7/23/2014 Content Complexity and Quality – NEW 2014 Pretty high correlations! CONTENT ELEMENTS RANKING FACTORS 2014
    37. 37. 66 www.searchmetrics.com 7/23/2014 Increasing CONTENT ELEMENTS RANKING FACTORS 2014 vs 2013Content Complexity and Quality
    38. 38. 67 www.searchmetrics.com 7/23/2014 𝜌2014 = 0.34 Relevant Keywords/ Terms
    39. 39. 𝜌2014 = 0.20 68 www.searchmetrics.com 7/23/2014 Proof Keywords/ Terms
    40. 40. 69 www.searchmetrics.com 7/23/2014 „You shall know a keyword by the company it keeps” (Firth, 1957) What’s behind? RANKING FACTORS 2014
    41. 41. Source: Google 70 www.searchmetrics.com 7/23/2014 What’s behind? RANKING FACTORS 2014
    42. 42. Proof KW Relevant KW Content Optimization
    43. 43. 72 www.searchmetrics.com 7/23/2014 Content Optimization
    44. 44. 73 www.searchmetrics.com 7/23/2014 Content Optimization
    45. 45. 75 www.searchmetrics.com 7/23/2014 Aggregators (few/ no own Content) • Couponing • Press Portals • Celebrity News Pages • Software Portals • Forums • Etc… Google Panda Update 4.0
    46. 46. There are no tips and nothing the user expected on the page. Why should it rank?
    47. 47. Conversational Search2.
    48. 48. Page 81 Hummingbird – Conversational Search 7/23/2014 ® Searchmetrics Inc. 2014 │ Android Chrome Glass
    49. 49. MOBILE vs DESKTOP MOBILE
    50. 50. 83 www.searchmetrics.com 7/23/2014 MOBILE vs DESKTOP Sitespeed
    51. 51. 84 www.searchmetrics.com 7/23/2014 MOBILE vs DESKTOP Filesize
    52. 52. 85 www.searchmetrics.com 7/23/2014 ~7.500 ~6.000 MOBILE vs DESKTOP Text Length in Characters
    53. 53. 86 www.searchmetrics.com 7/23/2014 MOBILE vs DESKTOP What’s the difference? RANKING FACTORS 2014
    54. 54. Semantic Search3.
    55. 55. Topic (Complex Entity) Who is the richest person in the world? ? Verb Entitity 1 Entity 2 Complex Entitity (1+2) Single Keyword Complex Query with several Entities
    56. 56. Google “understands“ the question “Answer“ above organic results ? Source?
    57. 57. User (Query) SERP (Knowledge Graph)
    58. 58. Hummingbird Query A1 Similar (or even synonymous) Keyword Queries SERP A1 Query A2 SERP A2 Query A3 SERP A3 Query A4 SERP A4 Query A5 SERP A5 Query A6 SERP A6 Pre- 92 www.searchmetrics.com 7/23/2014
    59. 59. Query A1 Hummingbird Query A6 Query A5 Similar (or even synonymous) Keyword Queries Query A2 Query A4 Query A3 SERP A Post- 93 www.searchmetrics.com 7/23/2014
    60. 60. Analysis 1: SERP Diversity
    61. 61. Diversity of SERPs All Keywords 95 www.searchmetrics.com 7/23/2014 ~6.5%
    62. 62. Diversity of SERPs All Keywords 96 www.searchmetrics.com 7/23/2014 34 ~6% ~2.5% 1 2 3 What happened?
    63. 63. Week 33 2013 1 Week 45 2013 2 Week 03 2014 3 10 104
    64. 64. Diversity of SERPs Keywords with Question Words What, Where, Who, How, Why … 98 www.searchmetrics.com 7/23/2014 34 ~6% ~10%
    65. 65. Query SERP C SERP A SERP B Pre- Hummingbird SERP Y SERP J SERP I SERP H SERP G SERP F SERP L SERP K SERP M SERP N SERP O SERP P SERP Q SERP R SERP S SERP T SERP U SERP V SERP Z SERP X 100 www.searchmetrics.com 7/23/2014
    66. 66. SERP H SERP M SERP J SERP R SERP F Query SERP C SERP A SERP B Post- Hummingbird SERP Y SERP I SERP G SERP L SERP K SERP N SERP O SERP P SERP Q SERP S SERP T SERP U SERP V SERP Z SERP X 101 www.searchmetrics.com 7/23/2014
    67. 67. Analysis 2: Keyword Pairs
    68. 68. 4/10 July 2013
    69. 69. 9/10 March 2014
    70. 70. 2/10 July 2013
    71. 71. 7/10 March 2014
    72. 72. HOLISTIC BE
    73. 73. LINK PROFILE
    74. 74. 109 www.searchmetrics.com 7/23/2014 BACKLINK ELEMENTS RANKING FACTORS 2014
    75. 75. 110 www.searchmetrics.com 7/23/2014 BACKLINK ELEMENTS RANKING FACTORS 2014 vs 2013
    76. 76. 111 www.searchmetrics.com 7/23/2014 More Backlinks! 𝜌2014 = 0.31 Number of Backlinks
    77. 77. 112 www.searchmetrics.com 7/23/2014 Better Links! 𝜌2014 = 0.26 SEO Visibility/ Source URL
    78. 78. 113 www.searchmetrics.com 7/23/2014 NEW Features 2014 BACKLINK ELEMENTS
    79. 79. 114 www.searchmetrics.com 7/23/2014 Fewer fresh Backlinks! more 𝜌2014 = 0.20 Number of new Backlinks
    80. 80. 115 www.searchmetrics.com 7/23/2014 𝜌2014 = 0.16 Anchor = Domain Name
    81. 81. 116 www.searchmetrics.com 7/23/2014 𝜌2014 = 0.15 Anchor = Brand/ URL
    82. 82. Evolution of SEO
    83. 83. USER SIGNALS Analysis 3: User Traffic
    84. 84. 125 www.searchmetrics.com 7/23/2014 USER TRAFFIC ELEMENTS RANKING FACTORS 2014
    85. 85. 126 www.searchmetrics.com 7/23/2014 𝜌2014 = 0.09 Time on Site
    86. 86. 127 www.searchmetrics.com 7/23/2014 Ø 𝜌2014 = 0.04 Bounce Rate
    87. 87. 128 www.searchmetrics.com 7/23/2014 𝜌2014 = 0.67 Click-Through Rate
    88. 88. ? SEO What‘s the goal of SEO for Content?
    89. 89. Content → Ranking → Traffic 7/23/2014130 www.searchmetrics.com What‘s the goal of SEO for Content?
    90. 90. 131 www.searchmetrics.com 7/23/2014 1. Conserve Traffic – conventional SEO-Methods 2. New Traffic through fresh Content What‘s the best optimized website in the world? Content → Ranking → Traffic
    91. 91. S E O
    92. 92. Search Engine Optimization
    93. 93. Search Experience Optimization
    94. 94. Techniques Content
    95. 95. twitter.com/searchmetrics facebook.com/searchmetrics THANK YOU FOR YOUR TIME

    ×