How To Write Helpful Content With Topic Modeling & Topic Clusters

How To Write Helpful Content With Topic Modeling & Topic Clusters

Senior Content Strategist at MarketMuse
Bio
Follow
This post was sponsored by MarketMuse. The opinions expressed in this article are the sponsor’s own.
Wondering how to tackle Google’s helpful content update?
At a loss for how to recover your organic traffic if your keyword strategies are no longer effective?
Instead of just ensuring that keywords are present in each individual piece of content, try focusing on overall topic modeling and topic clusters to drive your SEO content strategy .
Topic modeling and topic clusters go hand in hand when it comes to creating helpful content. It’s like peanut butter and jelly, baking and pies. You get the picture!
Here’s a recipe for making it work.
Use Topic Modeling & Topic Clusters, Not Just Keyword Research
Topic modeling is how machine learning detects and understands related topics and themes across multiple pieces of information. This can help you understand what pieces to create as part of a topic cluster.
Topic clusters are the actual groups of linked related content pieces that have a similar theme.
As Google’s algorithm continues to evolve, it’s becoming clear that it favors topic clusters because they better match the user’s full search journey and user intent.
The Difference Between Keyword Lists & Topic Models
When conducting keyword research you’ll typically end up with hundreds of different terms and still not have a clear understanding of what’s required to adequately cover a subject. It’s a case where more isn’t always better.
With topic modeling, you have a more clear roadmap to a more successful piece of content.
Topic modeling reveals how experts write about that subject, which directly reflects how readers typically journey through a subject.
Here’s a short example using the keyword phrase and topic “how to grow tomatoes.” As you can see, comparing keyword research to topic research is like comparing apples and oranges. They’re fundamentally different approaches.
Ideally, our MarketMuse topic model recommends 50 semantically related topics, but even the small excerpt shown below provides critical insight into the necessary coverage.
Image by MarketMuse, October 2022
In SEO, a topic model serves dual roles – prescriptive and descriptive.
Prescriptive means that it indicates what topics need to be covered.
Descriptive means that you can verify whether an article covers those required topics.
Google loves this.
How Do I Discover My First Topic Model?
In the case of MarketMuse, our patented topic modeling technology analyzes thousands of pages to help you discover those topics faster.
How Topic Modeling Produces High-Quality Content
Topic modeling serves as a map for producing comprehensive content.
In the case of “how to grow tomatoes,” following the map reveals:
There’s a particular disease (blossom end rot) that many tomato growers have to deal with. At a minimum, you should acknowledge that in your article, but it would be better to offer a solution.
There’s a pest (tomato hornworm) that also needs to be addressed, just like the disease.
Soil is an important factor in tomato growing success.
You’ll need to account for growing different varieties (determinate and indeterminate).
Tomato cages are a useful and popular accessory for growing tomatoes.
You’ll want to discuss buying tomato transplants, in addition to growing them from seeds.
Plus many other related topics that help make the article more relevant.
Screenshot from MarketMuse Research, October 2022.
How To Use Topic Modeling To Create Original Content
Original content isn’t simply rephrasing what everyone else says.
Original content goes beyond what’s already out there, contributing added value to the conversation.
You can achieve this quite easily with the help of a robust topic model, one based on its analysis of multiple data sets.
Step 1. Compare Your Topic Model To Your Competitors
Now that you’ve gotten your first topic model built , you’ve seen that our optimization screen ensures that we cover all the related topics, just like an expert.
Next, you’ll need to map the top 20 SERP results against your topic model to determine any deficiencies.
Run models on each of their content URLs to understand their topic models and topic clusters, then determine where content gaps lie.
The Easy Way: Do this step automatically with MarketMuse . Simply visit the “Compete” tab & build content around the red squares.
Screenshot from MarketMuse Compete, October 2022
Here, you’ll see a heatmap with the top 20 results across the top along with their URL and content score. Down the side are all the related topics in order of relevance.
The colored squares show how well each URL covers the topic, according to the number of mentions.
The red squares show that these topics are mentioned zero times by competitors; these are the ones to build content around.
You now have a verified topic model to create content from.
Not yet convinced of the need to go above and beyond in differentiating your content?
Keep in mind that Google has a patent (US20200349181) that can analyze any document to determine whether it provides additional information. Based on the patent’s filing history, they’ve been at this type of algorithm for quite some time.
Step 2. Create Topic Clusters With Your Verified Content Models
So far, we’ve used topic modeling to discover individual directions that are high in quality, are comprehensive, and cover gaps in your industry.
Now, let’s see how it can help in developing topic clusters.
Let’s return to the topic model for “how to grow tomatoes.”
How do we create a content cluster, a section, or an entire site around the same topic?
Instead of briefly mentioning these topics on one page, we’ll create dedicated pages for each topic. That offers a much greater opportunity to explore these topics in depth.
Start with “How To Grow Tomatoes” as your pillar page , then create separate pages for each supporting topic, such as “blossom end rot.”
 
Image by MarketMuse, October 2022
Then, create separate topic models for each of those pages and build out their high-quality supporting content.
Finally, add another layer to the cluster using topics from the model for the term “soil stream” as individual supporting pages.
Image by MarketMuse, October 2022
Repeat this process as required to explore the subject, building out the cluster to an even greater level.
Step 3. Link Your Topic Clusters At Scale With Topic Modeling
The last piece of the puzzle is internally linking the pages to build authority and provide a path for your audience to explore the subject further.
Once again, we use topic modeling to help.
On the pillar page “How To Grow Tomatoes,” where you discuss blossom end rot, link to the new page about that topic using “blossom end rot” as the anchor text.
Doing it in this manner puts all the SEO stars in alignment, so to speak.
 
Image by MarketMuse, October 2022
Both your pillar page (about growing tomatoes) and supporting content (about blossom end rot) are semantically connected because “blossom end rot” appears in the topic model for “how to grow tomatoes”. The link itself reinforces the connection, as it also uses the term “blossom end rot”.
Often, you may have pre-existing content that fits within a topic cluster that you might not know exists.
MarketMuse Connect helps discover this content by analyzing your site for pages that match the topics within the model.
 
Screenshot from MarketMuse Connect, October 2022
It can find up to 10 topics for which there is matching content on your site. In addition, MarketMuse Connect offers up to 10 unique suggestions for each of those topics. Just make sure to cover that topic in your article and link to the suggestion.
Start creating authoritative topic clusters and better content with MarketMuse.

Images Powered by Shutterstock