Google Display The Most Accurate Results Based on Query Intent
Google will try to understand the intent of a query in order to display the most accurate page in search results. For example, if they view the search query to be more transactional they will show a product page instead of an informational page. John recommends reviewing real queries that users are making and ensuring the intent is accurate for the page that is being displayed.
BERT is Used To Determine Which Pages Are Most Relevant for Different Queries
BERT is being used by Google to help them to understand both user queries and the content on websites, in order to serve the most relevant results to users. John reiterated that, besides ensuring you are creating content that is easily understandable for users, there is nothing specific that needs to be done to websites to make them work better for BERT. Google are constantly expanding and testing their understanding of queries in order to apply it for different languages.
Numerous Algorithms Used to Determine Which Search Results Features to Display
Google has a number of algorithms that are used to determine which of the search result feature blocks (e.g. videos, images and news results) to show above others, based on what is deemed relevant for individual queries. These algorithms are being used to understand what is being searched for specifically with the query.
Unnatural Links Can Hurt a Site Regardless of Algorithm Updates
If you believe your site is suffering from poor link-building schemes, John recommends focusing on removing these unnatural links, regardless of any algorithmic updates. This can be done in a number of ways, including using the disavow file or removing links from the source site.
BERT is Currently Just Used For English Content
Google currently just use BERT for English content, this is because most Natural Language Processing models are very language-dependent. It is therefore important to understand how things work in one language first, before generalising them for other languages.
BERT Will Not Cause Sites to Drop in Rankings or Lose Visibility
BERT is used by Google in order to better understand queries and text. This won’t cause it to see a site as less relevant, but rather helps to identify if the website is relevant to the query being asked. This may affect traffic from long tail queries, but will help to send more relevant users to the page.
Google Algorithms are Constantly Evaluating the Quality of Content on Websites
There are different algorithms running all the time to collect signals about a website’s content in order to gain a notion of the quality and relevance for individual queries. A core algorithm update will not lead to a website being completely re-evaluated but will update the scores used.
Social Media Likes Are Not Used by Google as a Ranking Signal
Google doesn’t use social signals, such as Facebook likes, as a ranking signal for websites. This is because it can be hard for them to keep up with the signals as the pages change so frequently.
Google’s Algorithms Should be Able to Detect & Prioritize Original Content From Near Duplicate Versions
Google’s algorithms will ideally be able to detect spun content which has been rewritten from another source and see the original content as more valuable.