7 documents related to Google Search ranking in evidence in antitrust trial
The U.S. Justice Department has released a number of new pieces of evidence, including internal Google presentations, documents and emails related to ranking in Google Search. Seven of these are discussed below, which specifically address elements of Google Search ranking.
1. life of the click (user interaction)
The first presentation, “The Life of a Click,” is a heavily censored document by Google’s Eric Lehman. Although it lacks full context, what is found here is of interest to SEO professionals.
The three pillars of Google ranking
In a presentation titled “Three Pillars of Ranking,” Google highlights three key areas:
- Content: What the document says about itself.
- Anchors: what the network says about the document.
- User interactions: What users say about the document.
It is worth noting that Google uses the term “clicks” as a substitute for “user interactions” in some places. User interactions include clicks, result-focused attention, scrolling and entering a new query. This approach is reminiscent of Mike Grehan’s E-A-T theory (page content, hyperlink analysis and usage data), which is important for SEO.
Google user interaction signals
The next slide, “User Interaction Signals,” illustrates the relationship between queries, user interactions and search results. Google lists specific interactions such as reads, clicks, scrolls and mouse hovers. It’s worth noting that while Lehman testified that Google uses clicks for ranking, individual clicks are not a sufficient signal for ranking. Google has publicly stated that it uses click data for training, experimentation and personalization of results.
2. ranking
The second presentation was about ranking in Google Search and included seven slides prepared by Eric Lehman.
Google false understanding
In the first slide, Google admits it doesn’t understand the documents and “fakes” that understanding. Instead, the company monitors how people react to the documents and relies on their reactions.
How Google search really works
The next slide explains that the real source of Google’s “magic” is the two-way conversation with users. Google points out that each user query and response from Google create chunks of knowledge that add up, teaching the algorithms.
3. ranking in research
The third presentation was on research and quality of Google search results. It included information on 18 aspects of search quality, including relevancy, page quality, popularity and others.
Disadvantages of Google live traffic
One of the slides discussed the drawbacks of evaluating live traffic, and Google points out that there is little connection between observed user behavior and the quality of search results.
Google search result ranking
Another slide presented a different understanding of Google’s search result ranking. The document also provided valuable information on attempts to manipulate search results and the need to exercise discretion in matters of algorithm performance.
4. Google is magic
The fourth presentation focused on how Google search works.
Google bie this is how search works
Google admits that the traditional search concept of directly understanding documents is insufficient. Instead, the company relies on users’ reactions to search results.
How Google search really works
Google explains that the key is a two-way conversation with users to gather insights from users and improve search results.
5 Login and Ranking
The fifth presentation focused on the role of logs in ranking and search.
Google Magic Dialogue
Google compares the search process to a potluck, where each person brings one dish to share. Similarly, search is driven by the vast mass of knowledge provided by users.
Google extracts valuable information from user data
The presentation describes how Google is trying to convert user behavior into value judgments of search results. It’s a difficult process that relies on statistical correlations.
6. mobile vs. desktop ranking
The sixth presentation addressed the differences between search engine ranking on mobile devices and desktop computers.
Comparison of metrics
Google analyzed various metrics, such as click-through rate, number of manual revisions, number of queries per task, query length and more. The results of this analysis suggest that it is necessary to take into account the different intentions of users on different devices.
7 Key points for Sundar to present
The last presentation did not contain surprising information, but one interesting point was the discussion of the impact of the BERT algorithm on ranking.
BERT algorithm
According to the paper, early experiments applying the BERT algorithm to various areas in search suggest significant improvements in understanding queries, documents and user intent. The BERT algorithm is seen as a revolutionary change in natural language understanding technologies. In conclusion, these presentations reveal a lot of valuable information about how Google Search ranking works and explain why end-user data plays such an important role in the ranking process. It is worth remembering that Google’s algorithms are constantly evolving and adapting to the changing needs of users, so SEO must be dynamic and flexible to stay on top of search results.
