The new buzzword in SEO is information gain. And like all new buzzwords, SEOs are throwing it around like we just discovered fire.
But there is a huge problem.
Information gain means different things to different people.
In this article, you will learn about obtaining information and how to use it to your advantage.
The 3 schools of information win: Humans, machines and search engines
Information gain can be used in three ways:
Machine learning. Google patent. Information search theory.
Information gain is used to train decision trees in machine learning. And unless you’re a computer programmer, we can leave that can of worms unopened (for now).
When SEOs talk about information gain, they are mainly referring to Google’s patent.
Google received a patent in 2022 on an information gain score that applied to documents.
This patent showed that Google had developed a way to measure the “similarity” of content and promote or demote it accordingly.
This is a great way for Google to treat content that is not essentially original or simply copied from another source and rehashed.
But what about obtaining information in relation to information seeking theory?
The theory of information seeking was documented in the book of the same name, written by Peter Pirolli.
Apply the models of how animals look for food (optimal feeding theory) to how humans search for information (which we will discuss later).
As you can see, we have three different meanings for the same term.
When it comes to SEO, Google’s patent is mostly easy to understand: just make your content unique.
However, the search for information is more complex, so we need to examine it further.
Why Searching for Information is Important for SEO
Recently, Google began to discuss the theory of information seeking in its report of decoding decisions (the messy half).
In fact, information search theory seems to be the direction Google is heading, and to quote directly from their report:
“An explosion in product choice and information has made it harder to feel confident that you’re making the right decision.”
Or, to put it another way, there is too much information.
If we have too much information, the time to make a purchase decision increases, and that is not good for anyone.
You can see why Google SGE can help things if you think about it.
By providing a generative AI response to a search query, a search user immediately captures the topic without the need to click through to a website.
This initial information should help the user make their next search decision.
Take this search result on Perplexity.
Within seconds, my knowledge of the best gym shoes for bad knees has increased, and there are many links and suggestions.
My next click will be to look at suggested shoes, not read another five blog articles.
If SGE works the same way, you can see how trading will radically change.
We are no longer optimizing for Google. We are optimizing the AI.
Dig deeper: LLM optimization: Can you influence generative AI outcomes?
Get the daily search newsletter marketers trust.
From SEO to information gain optimization
Google has been involved in AI for a long time, and AI is part of many of its systems.
They used BERT to improve their language comprehension, and I’m sure many more systems are used.
The point is that Google is trying to understand content to better serve search engine users. So Google itself is reading your content.
Sure, not like a human does, but they’re reading it.
So it makes sense to apply a similar approach to increasing Google’s information gain from content, just like humans.
In essence, we become information optimizers.
Our job as SEOs is to continuously increase the rate of information gain.
The rate of gain, explained
Rate of information gain, when it comes to information search theory, is described as:
Earning rate = value of information / cost associated with obtaining this information
You see, while search engines incur a cost to index and retrieve documents, so do humans.
When we use our brain, we consume calories and the body is very efficient at not wasting them.
We use heuristics (mental shortcuts) to filter the world and make decisions.
Information-seeking theory suggests that we seek to do the same. We try to get as much information as possible from a source in the shortest possible time.
To do this, we go through a five-stage process.
Goal
What information do we need?
patch
We decide which source of information would best achieve our goal. This could mean we go to Tripadvisor, TikTok, YouTube, or any website/search engine you can think of.
fodder
Here, we look for the information we need on the chosen platform. For this example, we’ll stick with Google. Type the keywords in the search engine to try to find the information you need.
smell
When we turn to search engines, we are looking for the smell of good sources of information; signals such as reviews, higher rankings and page titles that encourage clicks. We click on sites, scan information, and decide whether to spend time reading the resource.
diet
We consume information from various sources before making decisions. This is what Google refers to as the cluttered center of search. For brands/sites, being a part of your consumers’ information diet increases the likelihood that they will depend on you for information and trust you.
As we know, that trust leads to purchases or increased clicks (which can generate advertising/affiliate revenue). This means that SEO should include optimizing around the smell of information.
But if you’ve read above, you can see that Google search works similarly, just a machine version.
Optimizing information: the new science
If we want to optimize information retrieval, we must understand that it requires a greater understanding of two factors:
Machine learning. Human learning.
We already know that Google wants original and expert information from the best sources.
They also want to reduce the cost of extracting this information.
Yes, Google wants an easy life. So how do we do this on a practical level?
Simply put, we make information extraction easy for both machines and humans (at the same time), and here’s how.
The optimal website maximizes the value obtained per interaction
Contrary to popular belief, fast websites can matter, but if a website’s information gain rate is low or has a high perceived cost, then the person will leave.
Here is an example.
I asked ChatGPT for information about a hotel in Paris. It gives me the information in the best possible way.
It provides a lot of information that I can easily extract at a low cognitive cost. But how should a website deal with this?
Tripadvisor has an entire page dedicated to the hotel. See how they optimized a section for information gain rate.
The content, which uses symbols, dashboards and lists of room types, is designed so that humans (and machines) can get the most information in the least time/cost.
And that’s what we have to move our heads to help users search.
But we need to destroy some myths surrounding the content.
Good content is based on context
I read a lot of great content, but most of it is in my inbox in the form of blogs that people have written that aren’t designed to gain search traction.
Good content for SEO is very different.
When we search online, we have a state of emotional need that needs to be resolved.
Kantar and Google did some research a while ago
In this study, the above need states were used by searchers, who went to search engines looking to be solved.
Some words that stand out in front of each state of need are:
quick Laser focused. Specific phrases. To the point Simplicity No complications. trust Ratings reviews competition Location.
It is these attributes of information that users look for in online content.
Amazingly, we can see how Tripadvisor content exhibits these attributes, and we can also see how applying them to content would increase the rate of information gain for humans and machines.
But how can we begin to take the information-to-content optimization approach?
Well, here’s a four-part process to get you started.
Part 1: Structure of the content
See how your page should be structured for search to increase your bounce rate.
A good example is the Tui website:
They have used faceted search “buttons” to help users find what they are looking for.
Think about the best way to design your page for humans and search engines to increase information retrieval.
UX is important, as is the information on the page.
Part 2: Information architecture
Think about how you want your information to be structured to get the most information.
You may consider giving information early and quickly, for example:
“When is the best time to travel to Jamaica?”
“March is the best time to travel to Jamaica.”
Look at your content and try to add some, if not all, of the following attributes.
Quick Adventure Laser focused. Specific phrases. To the point Simplicity No complications. trust Ratings reviews competition
Part 3: Content design
The final impact is content design.
Think about the best way to add value, such as using unique images in your posts to help explain information or data.
Backlinko use images like the ones above to transmit data in an interesting format.
This brings us to the final part.
Part 4: Difference in content
If you do all of the above, you should have very different content than what already exists.
But if you don’t, make sure you do.
There are 1,000 different ways to say the same thing, but it requires creativity and consideration on how best to show your unique angles and points of view around it.
But here is a small challenge.
Go to a site like Backlinko or HubSpot and check out their content.
Find an article and apply the four-part system above and think about how you would improve it based on your unique opinions or experience.
This could serve as a suitable workshop for agencies and internal staff to consider obtaining information and how best to apply it.
Because in the age of generative content, information gain is king.
The opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.
