Lily Ray, Sr. SEO Director and Head of Organic Research at Amsive, shares the latest from Google at PubCon 2023 in Austin, TX.
With Google’s new Helpful Content Update, search marketers are forced to reconsider the value of every single piece of content on their site. Looking at content through the lens of Google Panda, E-A-T and Needs Met from the Google Rater Guidelines, and the new Helpful Content Update, learn how to analyze current content on a scale of helpfulness, how to make the tough decisions on content to remove, and how to read between the lines of all Google’s content related documentation to create the types of content Google wants to see ranking at the top of the search results.
7. BERT Crisis information Deduplication Exact match domain
Freshness Helpful content
Link analysis &
PageRank
Local news
MUM Neural matching Original content
Removal-based
demotion
Page experience Passage ranking Rankbrain Reliable information
Site diversity Spam detection
Google’s Active Ranking Systems
14. What the classifier looks for…
“In search of content with little
value, low-added value, or that is
not particularly helpful”
speak
actually
What
means:
SEO spam
15. If the majority of
site content is
“unhelpful”…
…even for “helpful” content.
16. The (un)helpful
content classifier:
o Is automated
o Uses machine learning
o Runs continuously
o Is not a manual action
o Can affect a site at any time
o Can be “lifted” after unhelpful
content is removed
29. Which turned into uncovering a network of 17 niche
sites all affected by the Helpful Content Update
0.
0.75
1.5
2.25
3.
May 22 June 2022 July 2022 Aug 2022 Sept 2022 Oct 2022 Nov 2022 Dec 2022 Jan 2023
SEO visibility
30. What are the 17 sites doing?
o Thousands of articles with mediocre content
o Trying to cover every topic within the niche, but
lacking true E-E-A-T
o Most articles start with “what is”
o Venturing into topics outside of its wheelhouse
(art or music site writing about health)
o Excessive interlinking between sites in network
o Stock photos only
o Aggressive ads and affiliate links
o Lying about readership statistics &
mentions in major brands (no evidence)
o Fake E-E-A-T signals e.g. “Fact Checked”
labels and (possible) fake authors
47. AI Content Experiment
• Published brand new niche site
• Scraped relevant People Also Ask
questions & used GPT3 API to answer
across ~10k URLs
• Added a few links from other sites
53. What did Google say? TL;DR:
o It is original and high-quality
o It demonstrates E-E-A-T
o It satisfies the Helpful Content System
o It doesn’t propagate misinformation
o It isn’t created with the primary purpose of
manipulating search results
AI content can be fine if:
54. Google also
recommends listing
author names for content
when the reader might
ask “who wrote this?”
However, there are 3 important caveats:
Google recommends
disclosing to the reader
when AI was used in
the content creation
process.
Google recommends
against listing AI as
the “author” of the
content.
55.
56. 43% of authors do not want to disclose AI content alongside their name
58. Let’s not forget these Google patents…
Agent Rank Author Vectors
Speaker
Identification
Identifying
Authoritative
Results
Website
Representation
Vectors
59. Takeaways
o Helpful content requires effort
o Taking shortcuts can catch up to you
o Instead of focusing on scale, focus on
content quality and E-E-A-T
o Stay in your lane
o Invest in your authors’ online reputation &
build out their Knowledge Graphs
65. Let’s be real. There are way more than 3 intents in SEO.
66. More accurate list of actual search intents:
o Buying the product
o Reading product reviews
o Finding local businesses & activities
o Seeing recent news
o Fact checking
o Defining the word or phrase
o Tutorials & troubleshooting
o Finding stock photos
o Recipes
o Medical symptoms
o Researching reputation
o Finding government/authoritative resources
o Seeing pictures
o Booking a trip
o Watching a video
o Hearing a song
o Translating a phrase
o Spell check
o How to
o Lyrics
o Adult videos/images