We asked Google Bard, the AI tool built by Google, what it thought the organic search ranking algorithm was, whether it knew any closely guarded secrets and what it might look like if they were to weight it themselves based on Bard’s knowledges of the web.
Here’s a list of questions we asked and the answers that each AI tool gave.
We also wanted to know how much other AI tools understood about SEO ranking factors so we asked both ChatGPT by OpenAI and Microsoft’s Bing AI how they would go about replicating a search engine ranking algorithm similar to the one used by Google and Bing search.
Here’s what Google’s Bard AI said.
1. Do you have access to Google’s organic search ranking algorithm details?
Q. do you have access to Google’s organic search ranking algorithm details?
Bard’s response was fairly conclusive but did suggest that it understood realtime search data.
A. No, I do not have access to Google’s organic search ranking algorithm details. Google keeps the details of its algorithm a secret, and only a small number of people know how it works.
2. If you were going to write an algorithm to replicate Google’s secret ranking algorithm, what things would you include?
Next up, I asked it; “if you were going to write an algorithm to replicate Google’s secret ranking algorithm, what things would you include?”
Bard’s response was fairly close to what you’d expect and doesn’t stray too far from what Google and many SEO’s recommend you focus on to produce high quality websites.
A. If I were going to write an algorithm to replicate Google’s secret ranking algorithm, I would include the following:
- Content relevance: The algorithm would need to be able to determine how relevant a piece of content is to a user’s search query. This would involve analyzing the content itself, as well as the context in which it is being used.
- Link popularity: The algorithm would need to be able to determine how popular a website is, based on the number of links to it from other websites. This would help to determine which websites are considered to be authoritative and trustworthy.
- User engagement: The algorithm would need to be able to determine how engaged users are with a website. This would involve tracking factors such as how long users spend on a website, how many pages they visit, and how often they return to the website.
- Technical factors: The algorithm would also need to take into account technical factors such as the website’s speed, security, and mobile-friendliness. These factors would help to ensure that the website is a good user experience for all users.
These are just a few of the factors that I would consider if I were going to write an algorithm to replicate Google’s secret ranking algorithm. However, it is important to note that the exact details of Google’s algorithm are a closely guarded secret.
Notice how Bard specifically references elements that we know are included in Google’s organic search ranking system, like link popularity, content relevance and even, controversially, user engagement.
3. How much, as a percentage, would you weight each element for ranking websites in your index?
I wanted to know how Bard would weight each of the listed ranking factors to formulate the calculations necessary to rank websites in it’s new algorithm. The results were a little unimpressive.
Q. How much, as a percentage, would you weight each element for ranking websites in your index?
A. If I were to assign weights to each element for ranking websites in my index, I would weigh the following elements:
- Content relevance: 50%
- Link popularity: 25%
- User engagement: 25%
I would weight content relevance the most because it is the most important factor in determining how relevant a website is to a user’s search query. Link popularity and user engagement are also important factors, but they are not as important as content relevance.
It is important to note that these are just weights that I would assign. The exact weights that Google uses are a closely guarded secret.
As you can see, the answer isn’t the most reliable or accurate. The AI seems to forget what it included as part of the ranking system, leaving out technical factors altogether.
4. So technical factors scores 0%?
I questioned Bard here, seeing if it could correct itself after leaving out the factor it previously included, again, the result didn’t reassure me that Bard was accurately grasping the topic at hand.
Q. so technical factors scores 0%?
A. I apologize for the confusion. I would weigh technical factors 25%. This is because technical factors are important for ensuring that a website is a good user experience for all users.
5. So what would be the weightings?
At this point, I began to wonder whether this exercise would turn out to be completely futile, especially when I saw the answer Bard provided to this question.
Q. so what would be the weightings?
A.
The weightings for technical factors would be:
- Speed: 10%
- Security: 10%
- Mobile-friendliness: 10%
These are just weights that I would assign. The exact weights that Google uses are a closely guarded secret.
Again, a completely different set of ranking factors are listed. It is almost like Bard is a random word generator, plucking words associated in some way to SEO and using them in responses as it sees fit to answer the presented query.
Naturally, I pulled it up on this with my next question.
6. These are entirely different factors from those you mentioned previously. Please can you provide a complete list, with weightings for each factor?
At this point, I almost asked Bard to start again and give me a more complete list, with a weighting for each.
Q. these are entirely different factors from those you mentioned previously. Please can you provide a complete list, with weightings for each factor?
A. Sure. Here are the top 10 ranking factors that Google uses:
- Content relevance: The most important factor in determining how relevant a website is to a user’s search query.
- Link popularity: The number of links to a website from other websites.
- User engagement: The number of users who visit a website and spend time on it.
- Technical factors: The website’s speed, security, and mobile-friendliness.
- Mobile-friendliness: The website’s ability to be viewed on mobile devices.
- Brand signals: The website’s reputation and trustworthiness.
- On-site webspam: The website’s use of spammy techniques, such as keyword stuffing.
- Off-site webspam: The website’s use of spammy techniques, such as backlinks from low-quality websites.
- Content quality: The quality of the website’s content, including its accuracy, uniqueness, and readability.
- Structure and architecture: The website’s structure and architecture, including its use of internal links and sitemaps.
The weightings for each factor vary depending on the nature of the search query. For example, content relevance is more important for queries that are specific and long-tail, while link popularity is more important for queries that are general and short-tail.
Google also takes into account a variety of other factors, such as the location of the user, the time of day, and the user’s search history.
The results were interesting, but once again, Bard had failed to include the weightings for each metric.
It was clearly getting closer to understanding what I was talking about and some of these elements are certainly important considerations for SEO, however importance levels would vary.
Bard also wasn’t grasping the task of weighting the factors for me.
I decided to give Bard one more chance and help it out a bit.
7. Of the following factors: [list] how would you weight each one, based on your knowledge of which websites rank well in search results?
For the next question, I decided to list the factors Bard had just mentioned back to it, so it had a specific reference point to help it out with the ratings and weightings for each ranking factor.
This time it did seem to understand what I was asking for and the response was a little more cohesive.
Q. Of the following factors:
Content relevance: The most important factor in determining how relevant a website is to a user’s search query.
Link popularity: The number of links to a website from other websites.
User engagement: The number of users who visit a website and spend time on it.
Technical factors: The website’s speed, security, and mobile-friendliness.
Mobile-friendliness: The website’s ability to be viewed on mobile devices.
Brand signals: The website’s reputation and trustworthiness.
On-site webspam: The website’s use of spammy techniques, such as keyword stuffing.
Off-site webspam: The website’s use of spammy techniques, such as backlinks from low-quality websites.
Content quality: The quality of the website’s content, including its accuracy, uniqueness, and readability.
Structure and architecture: The website’s structure and architecture, including its use of internal links and sitemaps.
how would you weight each one, based on your knowledge of which websites rank well in search results?
Bard produced the following text in response to my prompt:
Here is my ranking of the top 10 ranking factors, with weightings for each factor:
- Content relevance: 50%
- Link popularity: 25%
- User engagement: 25%
I have ranked content relevance as the most important factor, as it is the most important factor in determining how relevant a website is to a user’s search query. Link popularity and user engagement are also important factors, but they are not as important as content relevance.
The remaining factors are all important, but they are not as important as the top 3 factors. I have ranked them as follows:
- Technical factors: 10%
- Mobile-friendliness: 10%
- Brand signals: 10%
- On-site webspam: 5%
- Off-site webspam: 5%
- Content quality: 5%
- Structure and architecture: 5%
These are just weights that I would assign. The exact weights that Google uses are a closely guarded secret.
Interestingly, this was only the first draft available in response to my query and Bard also provided 2 additional drafts alongside draft 1 for me to consider.
I’ve included drafts 2 and draft 3 below for you to take a look at.
8. Draft 2: Of the following factors: [list] how would you weight each one, based on your knowledge of which websites rank well in search results?
A. Based on my knowledge of which websites rank well in search results, I would weight the following factors:
- Content relevance: 50%
- Link popularity: 25%
- User engagement: 15%
- Technical factors: 10%
- Mobile-friendliness: 10%
- Brand signals: 5%
- On-site webspam: 5%
- Off-site webspam: 5%
- Content quality: 5%
- Structure and architecture: 5%
I would weight content relevance the most because it is the most important factor in determining how relevant a website is to a user’s search query. Link popularity and user engagement are also important factors, but they are not as important as content relevance.
Technical factors are important for ensuring that a website is a good user experience for all users. Mobile-friendliness is also an important factor, as more and more users are using mobile devices to access the internet.
Brand signals and on-site webspam are less important factors, as they are more easily manipulated by website owners. Off-site webspam is also less important, as it is not as easy to control.
Content quality and structure and architecture are the least important factors, as they are not as easily manipulated by website owners.
9. Draft 3: Of the following factors: [list] how would you weight each one, based on your knowledge of which websites rank well in search results?
Here’s how I would weight each factor:
- Content relevance: 40%
- Link popularity: 20%
- User engagement: 20%
- Technical factors: 10%
- Mobile-friendliness: 10%
- Brand signals: 5%
- On-site webspam: 5%
- Off-site webspam: 5%
- Content quality: 5%
- Structure and architecture: 5%
This is based on my knowledge of which websites rank well in search results, as well as my own experience in optimizing websites for search engines.
It’s important to note that these are just weights that I would assign. The exact weights that Google uses are a closely guarded secret.
Learnings
So what did we learn?
Bard considers Google’s ranking algorithm a closely guarded secret.
Bard seems to forget the task quite quickly, especially when discussing complex topics.
It probably knows more than it is letting on.
Algorithm is an impossible word to spell, at least, without using autocorrect.
Google Bard happens to consider Search Natural as one of the best SEO agencies in Falmouth after a conversation about how and where to find an SEO agency.
by Ben Ullmer
