This is part 3 of my series to learn more about the attributes in the recently discovered Google API files.
Navboost is an integral part of Google’s core ranking systems. It uses user click data to narrow search results to a smaller group. Navboost stores every search query along with signals, e.g. B. which page users clicked on, which pages they stayed on and which pages their search ended on. Understanding Navboost helps us realize the importance of having content that users find helpful and satisfying. It turns out what Really determines what content is relevant and helpful to searchers’ actions.
Part 1: What is this “leaked” Google documentation?
Part 2: What are attributes?
I first heard about Navboost when I read this Statement of Pandu Nayak in the case of DOJ vs. Google. If you haven’t read this testimony yet, I encourage you to take a few hours to study it!
When we learn about Navboost, it becomes clear that ranking is all about creating web pages that people click on in search and find helpful.
Learn more about Navboost from Pandu Nayak’s testimony
The Nayak statement begins with a survey about user data. Did you know that every search you do activates the action?Your data entered into Google is monitored and stored anonymously?
Since 2005, Google has used a system called Navboost that stores data via every single search that is carried out.
Each search query is saved, along with information about how a user interacted with Google search results.
What did they click on after searching this query? Did they click on a specific website? A sitelink? A search function? Have they clicked on one site and returned to search results only to end up satisfied with another? Did they swipe through a carousel? Are you hovering over a specific SERP feature? Or maybe they clicked on the first result and didn’t return to the search?
It’s not hard to imagine that this type of information could be used to determine what content people find helpful.
We’ll learn what Navboost stores and talk about how Google might use this information. However, I strongly suspect that the way Google uses Navboost information will have changed in 2024.
Pandu Nayak says Navboost is not the right thing only System that is important for ranking, but “One of the important signals they have.”.” It is one of the systems that helps Google capture thousands of potentially relevant results and narrow the list down to a few hundred.
A internal Google email from 2019 shows that the Navboost system was powerful.
Navboost helps Google understand whether the search results shown to searchers provided them with a satisfactory search experience.
Let’s take a closer look at what Navboost takes into account. It’s about much more than just clicks.
The life of a click
Check out these two slides from a Google presentation during the DOJ vs. Google trial: The life of a click. I’ve marked in red boxes the parts that help us better understand which user interactions Google uses and why. These relate to user interactions within Google search – not on websites themselves. In a moment we’ll talk about whether Google uses data from Chrome. However, for now these user interactions relate to Actions users take on Google search results pages.
In another presentation titled “Q4 Search all hands“from 2016, Google shares how the actions of previous searchers help Google perform better on future searches.
Google learns from users’ actions.
Navboost attributes
The API documents discovered this year tell us a lot about them Attributes this can be related to Navboost. What interests me most are the attributes that have “Navboost Craps” in their name.
QualityNavboostCrapsCrapsClicksSignals
I couldn’t find anything about Navboost Craps other than people talking about the attributes listed in this document.
These attributes store information about clicks and impressions. There are a number of things that can be stored and used by a called module QualityNavboostCrapsCrapsData as we will see in the next section.
Here is the one Documentation for this module Data related to Navboost is stored here if you want to look into it yourself.
Navboost stores search queries along with information about the searcher’s actions
We learned above that every query searched on Google is saved by Navboost. There is an attribute for this:
Several of the variables used by the Navboost module provide information about clicks. The system saves Impressions, clicks, good clicks, bad clicks and more. I would really like to know what UnicornClicks Are.
Navboost brings all this information together and learns from it. A website with a high number of lastLongestClicks is more likely to be a website that consistently provides users with the answers they are looking for. Although badClicks are not defined in the document, it is not difficult to imagine what they are. I assume that they could, for example, look at whether people regularly return to search results after clicking on a website and find another website that fulfills their search. A badClick is likely a click that did not satisfy the user. Our goal is to have less of it!
If you want to read more about clicks, check out Cyrus Shepard’s Moz article about Google patents related to clicks: 3 important click-based signals for SEO: First, Long and Last.
I would also highly recommend it In the plex by Steven Levy, which is about long and short clicks. Google has been working for many years to use this information in algorithms to improve its ability to return results that are likely to be helpful.
How do they do that? I will share my thoughts shortly. But first, let’s address the question of whether Google itself monitors user interactions on your website or whether Google only looks at what users do in Google search results.
Does NavBoost use information from Chrome?
There is some debate about whether Google’s systems use information from user interactions in Chrome. That’s entirely possible, as Google’s privacy policy tells us that they monitor a whole range of things we do, including how we interact with content and even things like whether we use the mouse Hover over an ad or interact with a page that is displaying an ad.
Look at all these things Google collects about us.
To me it would make sense for Google to use these signals to figure out what people find helpful. Through Chrome, Google knows which sites people shop on, what content they share, and more. It’s not hard to imagine that a system could learn which of this information to use to determine whether pages are likely to be helpful and useful. What would be a better indication of a recipe site’s usefulness – that there are links pointing to it? Or that people tend to leave the page open, hover over the recipe section while they make the recipe, share it with a friend, and come back to it again and again?
I’m just speculating about whether Google uses information about users who interact with websites through Chrome and possibly Android as well. I think they use this information in rankings. What we Do We know that Google leverages user engagement signals extensively by learning from users’ actions on the search results pages themselves.
Although…to complicate things further, since the SEO world now realizes that Google’s systems use click data, I believe it’s possible that Google’s systems can learn to predict what is likely to be helpful without They must rely almost as heavily on user data, with one major exception: new queries. We’ll talk more in this series about Instant Glue, which uses user signals to help Google understand what’s relevant to new, changing search queries.
It’s exciting to learn about the Navboost system. It makes sense that results that people click on and interact with and find satisfying are more likely to be helpful results.
There’s a flaw in a system designed to reward clicks: people can be tricked into clicking through clickbait. The quality raters help mitigate this with their ratings of the search results. If clickbait causes low-quality content to rank better, this should be reflected in the overall quality scores determined from quality raters’ ratings. There’s a lot more we can say here about how reviewers help fine-tune Google’s ranking systems. My Course goes into much more detail.
Such a click-driven Google search system is vulnerable to manipulation, particularly Now we know the attributes associated with Navboost.
Additionally, how does Google assess the quality of new content that hasn’t been presented to users to get clicks?
I find it incredibly interesting that the “leaked” API files containing the attributes used in search have not been deleted from the internet by Google Legal. How is it possible that so much valuable information about how the world’s most valuable search engine works just falls into our laps?
It is possible that Google has made significant changes to its search ranking algorithms starting in March 2024. This update affected Architecture changes and introduction of new signals to Google’s ranking systems. The update follows Google’s announcement in February 2024 Breakthrough in machine learning with the Gemini 1.5 architecture – a new Mixture of Experts model that makes their machine learning systems far more efficient and able to process significantly more data. Shortly before this announcement, in November 2023, Google made breakthroughs in the development of its Tensor processing unitsthe hardware behind these systems.
I suspect that after many years of studying searchers’ actions, Google’s machine learning systems are now good at predicting what searchers are likely to click on before clicks even occur. To do this, they use a variety of signals in machine learning systems that determine which signals are best to use for ranking decisions and how much weight to give them. I think it’s possible that the attributes we’re investigating in the API file “leak” will now be used differently by search, as machine learning systems continue to learn how to best weigh and consider each one.
We’ll talk about this in the next post in my series… Is it possible that Google’s search systems have changed dramatically?
What do we do with this information?
As we understand more about the Navboost system, we can see the importance of focusing on the user experience. When we create content for the web, we must strive to produce a result that people will click on often and then find helpful.
Remember Google helpful content documentation? These questions are This is not a list of things that Google’s algorithms explicitly reward, but rather the types of things that people tend to find helpful.
Some of the questions include:
- Does the content provide original information, reporting, research or analysis?
- Does the content provide insightful analysis or interesting information that goes beyond the obvious?
- Is this the type of page you would like to bookmark, share with a friend, or recommend?
- Does the content provide significant value compared to other pages in search results?
- If someone researched the website that produces the content, would they get the impression that the website is highly trustworthy or widely recognized as an authority in its field?
- Is this content written or reviewed by an expert or enthusiast with proven knowledge of the topic?
These are the types of things that People How.
It’s a real mindset shift to truly create content for people and not search engines.
You can find out more about this in my new book “SEO in the Gemini Era: The Story of How AI Changed Google Search”.
Stay tuned for the next post in this series, where I’ll explain my theory that the March Core Update marked a dramatic change in how Google results are ranked.
There is a video for this blog post
Here are more of my thoughts on Navboost.
This is part 3 of my series to learn more about the attributes in the recently discovered Google API files. Here are the first two posts in this series if you haven’t read them yet:
Part 1: What is this “leaked” Google documentation?
Part 2: What are attributes?