How Google can personalize search results?

5/5 - (1 vote)

The personalization of search results is one of the last steps in the ranking process and is applied after relevance scoring and quality classification.

As part of the November 2024 core update, there were speculations from the SEO community that the update was a personalization update. I myself was one of the first SEOs to share this assumption:

But what does this mean and to what extent can Google personalize search results? Based on official Google info, Google patents, the data from the API Leak and the info from the DOJ trial this article provides thoughts and ideas on this question.The Patent and Paper Analyzer, SEO Thought Leader Research Assistant and Google Antitrust trial & API Leak Analyzer supported me in this research.

Contents

1 How personalization at Google works?
2 Which metrics Google is using for personalization?
3 Is Google using search history in user profile or browser history for personalization?
4 What are the possible consequences for SEO of a higher degree of personalization?

How personalization at Google works?

According to official documentation personalization at Google is primarily limited to factors such as the user’s location, language, time, platform used, and the distribution to different data centers. Other factors like click-through rate, past searches, and other user signals have neither been confirmed nor denied by Google. However, personalization is about a user-related individual ranking and not about a general scoring of certain content.

For search queries that have a reference to entities, the search history, i.e., past search queries made with reference to the respective entity, plays a role. For example, if you search for an entity with an ambiguous name like “apple” the search results you get will be influenced by your past searches related to “apple”

Google emphasizes that search results are only minimally personalized based on interests and search history. The most common form of personalization takes place based on the searcher’s location and set language. The results displayed in the news carousel are never personalized.

You can update or change the personalization settings or preferences in your Google account at any time. You can also find content preferences like SafeSearch in settings, which help you make a choice about whether search results include graphic content that may be shocking for some users.

From the DOJ Antitrust trial documents the following conclusions can be drawn:

Location-Based Personalization:

Google personalizes results based on location for both desktop and mobile searches
As confirmed by Google’s representative: “Location matters on desktop also. If I come and search for weather on desktop, I hope I would get weather in Washington, D.C. as against, you know, in Chicago.”

User Data Storage:

Google maintains user search data for extended periods
The data is anonymized and de-identified
When asked about how long they keep data about searches (like “Taylor Swift”), the Google representative indicated they may keep anonymized data indefinitely, though they weren’t the expert on log storage specifics

Machine Learning and User Data:

Google uses user data for training their models
They might train models on specific subsets of users (e.g., U.S. users for U.S.-specific models)
For global models, they look at all users’ data
The volume of data increases over time – a month of data today contains more clicks and queries than the same period three years ago

It is important to distinguish between user data that is used for general relevance and quality assessment and user data that is used for user-specific personalization. Ranking systems based on user data, such as Navboost or Deeprank, are used for general evaluation, not for personalization.

Google’s personalization method, as described in the patent “Personalizing search results” (Patent ID: US8977630B1), involves re-ranking search results based on user preferences. This is achieved by modifying the global ranking algorithm to reflect individual or group biases. The aim is to enhance search result relevance by integrating user-specific preferences into the ranking process.

The patent explains that existing search engines use global ranking algorithms to rank documents based on search queries, but these algorithms do not account for individual user preferences, which can lead to less relevant search results for different users.

The patent claims involve identifying a first set of documents associated with a user, with each document being assigned a weight reflecting the user’s interest.

You can find more details about this patent here.

Which metrics Google is using for personalization?

Google uses a variety of metrics for personalization, as indicated in the patent summaries and research papers in the database and official Google info. Here are some key metrics:

User Interaction Metrics: These include metrics such as click-through rates, dwell time, and engagement with content.
User Satisfaction Metrics: Metrics based on user satisfaction provide a more comprehensive evaluation of search engine performance than click-based metrics. For example, satisfaction signals can be incorporated into algorithms to prioritize results that provide value directly on the SERP.
Past Interaction Measures: The patent “Personalizing search results” (Patent ID: US8977630B1) also mentions the use of data from the Past Interaction Measures System to refine the scores of documents. This suggests that past user interactions with search results could be a factor in personalization.
Location: Google uses the user’s location to provide more relevant search results. For example, if a user in New York searches for “restaurants,” Google will show restaurants in New York.
Language: Google uses the user’s language settings to provide search results in the user’s preferred language.
Time: The time of the search can also influence the search results. For example, if a user searches for “news” early in the morning, Google might show news articles published that morning.
Platform Used: The type of device used for the search (mobile, desktop, tablet) can also influence the search results. Google aims to provide a good user experience on all devices, so the search results might be different depending on the device used.
Distribution to Different Data Centers: Google uses different data centers around the world to provide faster and more reliable search results. The data center used for a particular search can influence the search results.
Search History: Google uses the user’s past search queries to provide more relevant search results. For example, if a user frequently searches for information about dogs, Google might show more dog-related search results in the future.
Click-Through Rate: Google might use the click-through rate on search results to determine how relevant those results are to the user. If a user frequently clicks on a particular type of search result, Google might show more of that type of result in the future.

Remember, these metrics are used to enhance the relevance of search results by integrating user-specific preferences into the ranking process.

Is Google using search history in user profile or browser history for personalization?

There is no indication from Google or other documents as to whether browser data is used for personalization or only the search history from Google user profiles.

Info from the DOJ trial and API leak:

Chrome/Browser History:

Despite previous public denials by Google representatives (like John Mueller and Matt Cutts stating “We don’t use anything from Chrome for Ranking”), the leak reveals:

Chrome data is indeed used for ranking purposes
Chrome data is used in sitelinks generation
A leaked internal presentation from 2016 confirmed Chrome data was being integrated into search

User Profile Data:

The leak shows Google uses:

Cookie history
Logged-in Chrome data
Pattern detection through “unsquashed” vs “squashed” clicks (for spam detection)
User engagement metrics including clicks and impressions

What are the possible consequences for SEO of a higher degree of personalization?

The consequences for ranking monitoring in the context of increased personalization can be significant:

Variability in Rankings: With increased personalization, rankings can vary greatly from user to user based on their search history, location, language, and other factors. This means that the traditional concept of a ‘universal ranking’ becomes less meaningful.
Need for Personalized Tracking: SEO professionals may need to consider tracking rankings in a more personalized way, taking into account different locations, languages, devices, and other factors that can influence personalization.
Focus on Average Position: Instead of focusing on specific ranking positions, it might be more useful to look at the average position of a page for a particular query over time. This can give a more accurate picture of how well a page is performing in search results.
Importance of Other Metrics: With the variability in rankings, other metrics like organic traffic, click-through rate, and conversions become even more important to measure the success of SEO efforts.
User-Centric Approach: SEO professionals need to focus more on understanding their target audience and their search behavior. This includes understanding the types of queries they use, their needs and intentions, and how these might influence the personalized search results they see.
User Engagement Optimization:
- Increased importance of engagement metrics like CTR and dwell time
- Need to minimize “pogo-sticking” (users quickly returning to search results)
- Focus on creating content that keeps users engaged longer
- Importance of monitoring and analyzing click data more rigorously
Brand Building Becomes More Important:
- Focus on building brand recognition and trust
- Encourage branded searches
- Build strong social media communities
- Invest in traditional PR and authoritative media mentions
Greater Focus on User Intent: Understanding and aligning content with potential user intent becomes crucial. This means SEO professionals should focus on identifying and using key phrases that are likely to be predictive of user intent and related queries.

About the Author
Latest Posts

About Olaf Kopp

Olaf Kopp is Co-Founder, Chief Business Development Officer (CBDO) and Head of SEO & Content at Aufgesang GmbH. He is an internationally recognized industry expert in semantic SEO, E-E-A-T, LLMO, AI- and modern search engine technology, content marketing and customer journey management. As an author, Olaf Kopp writes for national and international magazines such as Search Engine Land, t3n, Website Boosting, Hubspot, Sistrix, Oncrawl, Searchmetrics, Upload … . In 2022 he was Top contributor for Search Engine Land. His blog is one of the most famous online marketing blogs in Germany. In addition, Olaf Kopp is a speaker for SEO and content marketing SMX, SERP Conf., CMCx, OMT, OMX, Campixx...