Why Keyword Research Doesn’t Help Create Better Content (and What to do About it)

There is no shortage of keyword research guides promising to help find those untapped, high-volume keywords for which your site can rank. Finding a keyword phrase like that is the Holy Grail of content marketing. The rationale is that keyword volume is a good indicator in deciding whether or not to write an article on the topic.

Keyword competition (based on link metrics like domain authority) is another criterion often employed to determine the likelihood of ranking for a particular term. If high ranking pages have a large number of inbound links, then keyword difficulty is judged to be high.

These two variables are plotted on a graph, as shown below. Those who use this process believe that the sweet spot lies in the space where monthly keyword search volume is high, and competition is low.

Granted, it’s a simplified view of the keyword research process, but that’s it in a nutshell. Some people prefer to use Google’s Keyword Planner to come up with keyword ideas, while others prefer using a third-party tool. Regardless of how you create your keyword list, one issue remains.

The approach of using monthly searches as a north star metric is problematic. Here is why.

Pages don’t rank for only one search term; that’s not the way search engines work. Thanks to Google Rank Brain, the search engine can match in-depth content to the appropriate user intent profiles. Thus, there is a hidden potential in keyword with little organic search volume, including those with no volume of which to speak.

An article targeting a low-volume keyword can rank for hundreds of other keywords, bringing in substantially more traffic than your favorite keyword tool would indicate. Here are more examples of posts written for long-tail keywords that generate lots of search traffic.

Ignore keywords with low monthly search volume, and you risk missing out on some serious traffic. On the other hand, if you just focus on long-tail keywords with little competition, chances are your website will provide visitors with a fragmented experience.

Think of your site like a puzzle; every piece of content you publish fills in another part of the puzzle. Creating content with close semantic relations, and linking them to form topic clusters, fills in the puzzle faster than random posts created with little forethought.

Here’s something else to keep in mind.

The Failure of Keyword Research to Improve Content Quality

Keyword research isn’t beneficial when it comes to creating content. This has nothing to do with your research skills and everything to do with the nature of the keyword research tool itself.

These tools aren’t good at establishing semantic relationships among topics. So, when you type “content strategy” into a typical keyword tool, the relevant keywords it comes up with leave much to be desired. Here’s an example from a free keyword tool.

Unfortunately, these keyword suggestions offer little insight into the concepts that need to be covered to create an in-depth article. Therein lies the crux of the problem.

Contrast that with output from MarketMuse Compete application.

The topic model is built by analyzing thousands of web pages for the focus topic. The list of related topics is sorted by relevance, so it’s immediately clear what the must-have topics are in a post about content strategy.

The Suggested Distribution offers an idea of how frequently a topic is mentioned, based on its occurrence in the top 20 search results. Take a look at Topic Research vs. Keywords for Content Strategists if you’re not clear on the difference between topics and keywords.

Google Doesn’t Rely on Keywords

If keywords are driving your content, then you’ve got it all wrong. Here’s an example (courtesy of Bill Slawski) to show you what I mean.

Type in the search term “who said they loved napalm?” and see what you get.

Pay attention because this is important.

None of the URL’s in this SERP use the keyword phrase that I employed in my search. There’s no search engine optimization occurring for that specific search term.

So, how does Google serve up the most relevant results for those searches? It relies on semantics to find attributes of facts and related topics. In addition to a Knowledge Graph, the search engine also has a Topic Layer so that it can understand both entities and topics.

Over the years, Google has filed numerous patents related to semantic search. SEO by the Sea is an excellent place if you want to learn more about the subject.

Using Topic Modeling to Craft Expert-level Content at Scale

Keyword research doesn’t help when it comes to improving content creation. But if you think that “sprinkling a few LSI keywords will work,” think again!

Topic Modeling

So, stop pondering keywords and start focusing on topics. Don’t worry about search volume. Instead, be concerned about user intent. Use topic modeling to better understand concepts that need to be covered to ensure your content matches the intention and that it’s comprehensive. Here’s a good explanation of topic modeling for SEO.

To take that one step further, topic modeling enables you to determine where gaps exist, not only in your blog posts but that of the competition. That’s how Social Media Sharks and Tomorrow Sleep used MarketMuse Suite to:

• Grow organic traffic from 4k per month to 400k with a year.
• Outrank their largest competitors.
• Hold multiple positions in a single search engine results page.

Read the case study here.

Remember to take advantage of your content inventory. Often, existing pages can be expanded or re-worked to provide enhanced coverage of a chosen topic.

Once more, apply topic modeling to pages that already exist to determine how to make them better. You can do this with MarketMuse.