Google is increasingly ranking content based on relevance signals derived from algorithms designed to help them better understand search queries and webpages.
We are in a post-keyword ranking era. Researching keywords still matter.
But the way Google ranks websites, it’s largely not about matching words in a search query to words on a web page.
This article outlines a way to analyze the search results to understand why Google might be ranking a web page and how to apply those lessons to content writing.
Write About Topics?
A common idea in SEO is to set aside writing about keywords and begin writing about topics.
That approach makes sense. Until you stop and think about it.
What “topic writing” proposes is that instead of writing about the keyword Blue Widget, the writer must write about the topic Blue Widget and all the topics associated with blue widgets.
The old way was to think about Blue Widget and all the associated Blue Widget keyword phrases:
Associated keyword phrases
- How to make blue widgets
- Cheap blue widgets
- Best blue widgets
The topic method for writing content begins with outlining all the different topics to write about:
- How to make blue widgets
- Cheap blue widgets
- Best blue widgets.
As you can see, the writer ends up writing about keyword phrases, regardless if they are writing about topics or writing about keywords. There is no difference in what is being written.
Another approach to topic writing is to do a Google search and then review Google’s suggestions within the Related Searches and People Also Ask features.
They say to then take those keywords and write about those. But that is just rehashing the technique of writing for keywords.
Writing about topics is the same thing as writing for keywords. The only difference between topic writing and keyword writing is where the keywords are coming from. That’s it.
In Search of What Users Want
Something I’ve noticed about Google search engine results pages (SERPs) is that the webpages in the results read like they are answers to questions.
To use an obvious example, if you type a short phrase like “fried ribs” Google shows recipe pages.
To Google’s algorithm, when someone types “fried ribs” it’s essentially the same as typing, “How do I cook fried ribs?” or “Show me fried rib recipes.”
Google responds to all those queries with essentially the same sites.
When you search for “fried ribs” you are really asking “show me fried rib recipes.”
That is what I call the Latent Question. Latent means many things but one of the meanings is the sense of something that is hidden.
So when I talk about the latent question, what I mean is that every keyword phrase contains a hidden question and a deeper meaning.
The latent question for the search query, “fried ribs” is “How do I cook fried ribs?”
Almost every search phrase contains a latent question.
When you examine the SERPs, you might find that the first three search results may be similar. That’s because they share what is commonly known as the searcher intent. However, they also share the latent question.
The latent question is not the same thing as searcher intent.
- Search intent is a general understanding of what a user wants.
- The latent question identifies the more precise answer that a user wants.
Examples of Searcher Intent:
A popular way to define search intent is:
That is a macro-level view of what a searcher wants.
The latent question approach to search intent analysis is to identify the question that the user is really asking when they type a keyword in the search box.
The keyword phrase “fried ribs” becomes “show me fried ribs recipe” or “teach me how to make Chinese fried ribs” and many other variations. Google chooses the latent question variation that satisfies the most users.
The phrase latent question is a phrase I created to describe an approach to gaining a more precise view of why Google might be ranking a webpage.
I developed the latent question approach independently. Google has recently published a research paper describing a similar way to understand the hidden meanings within a search query.
Google Research on Hidden Need States
We know that Google uses technologies like stemming and natural language processing to understand what a search query is as well as to understand what a webpage is about. The article linked below explores what Google terms need states.
Google’s consumer research has identified what they call six need states that influence search behavior.
The six need states are:
- Thrill me
- Impress me
- Educate me
- Reassure me
- Help me
- Surprise me
Here is how Google illustrates those need states that are hidden (latent) in a search query:
“When a person realizes that they need something, so they might need to find things to do for their kids this weekend. They might turn to Google and search for ‘jumpy houses near me.’
The thing is, behind that, the need beneath that is, they actually want to be helped and help their family connect and have a moment together.
Those are the pieces that are actually motivating their behaviors. What they search is trying to solve that. But people don’t say that.
Educate me is a great example. It’s a big need state. People are coming to us to get information, and then move on.
You don’t want a sales pitch. You don’t want a big scene of creative. You want to know rank order. Give me the facts of what’s safest.
Marketers in that moment can be more efficient, more direct.
All along that, their needs are trying to be met. And they’re going to keep going in their journey until they feel like their needs are met.”
How to Identify the Latent Question
I created the concept of the latent question a few years before I read the above Google research. But it meshes perfectly with my concept.
The Google research is trying to identify what needs a user wants to be met when they make a search query.
Their example is to identify the phrase “jumpy houses near me” contains a hidden meaning of: help my “family connect and have a moment together.”
Using Google’s example you can ask, what do people really mean when they ask for “jumpy houses near me” and also what needs are they trying to have met?
So, when someone searches for “authentic Italian pizza Boston Ma” or “Boston Ma Pizza,” Google’s non-local part of the search algorithm treats those search queries as almost the same query and shows reviews of pizza restaurants in Boston in the top search positions.
The first three results have the words Best and Pizza in their titles and the fourth-place result is an article about the author’s experience of trying to understand what defines Boston’s style of pizza and discusses the variety of styles to be found.
The latent question for the first three results is clearly, “Show me a list of the best pizza restaurants in Boston.” The need in that query is a list of the best pizza restaurants to compare.
The fourth result describes a quest to understand the different styles of pizza in Boston. It’s an in-depth review of pizza restaurants and not at all a list of pizza restaurants or user-generated content about pizza restaurants in Boston.
This fourth result is different from the first three results and so the latent question must be different. It could be construed as “compare and describe the best pizza restaurants.”
Always let the search results tell you what users want to see. Google tries to show search results that users expect to see. So if the top results are lists of the “best” then that may be the kind of content you should consider creating.
In this example, the latent question can be said to be, “Show me a list of the best pizza restaurants in Boston.” and your content should clearly be about, “Here is a list of the best pizza restaurants in Boston.”
This concept of latent search queries is especially interesting for results that have featured snippets.
Understanding what users mean when they type a vague search query is what BERT is all about.
In those cases, Google will show the most popular search intent first then in the next block of results Google will show the next popular search intent.
But as I said previously, understanding the search intent is a general (macro) understanding of what is going on behind that search box.
If you can identify the latent question, you are now understanding the search results at a micro-level. It’s like the difference between seeing the forest (search intent) and seeing the trees (latent question).
Understanding the latent question isn’t a way to crack Google’s algorithm. It is simply a way to better understand what users want when they use a particular search query and then to create content that meets those needs.
- Google: 6 Need States Influence Search Behavior
- Content Analysis and Better Ranking
- How to Mine the SERPs for SEO, Content & Customer Insights
- Advanced Technical SEO: A Complete Guide
Featured Image: Shutterstock