Keyword Research for Search Intent

by Geoff Hineman

In the early days of SEO, rankings were often determined by the number of times a particular keyword was repeated on a page. You could even goose the relevance of a term by putting it in header tags or bolding it. Pretty quickly, keyword stuffing became the quick path to good rankings. The problem then became SERPs full of often-unreadable content.

This not only posed a problem for searcher, but for Google as well. For Google's advertising model to work, they MUST deliver the best organic results first. If people can't find what they need on Google, they look elsewhere, and Google loses out on the potential revenue that comes from lost ad clicks.

To help deliver better results, Google has taken several steps to better understand content on a page, from Latent Semantic Indexing (LSI) to Bidirectional Encoder Representation from Transformers (BERT). Paired with newer search technologies (e.g., voice search, smart TVs, etc.), the ability for Google to discern the meaning of a page has never been better. The result is new opportunities for SEOs and other content publishers to create content to meet searchers at every step of the funnel.

In this post, we are going to discuss strategies for keyword research for search intent, so you can meet prospects where they are on their search journey.

Latent Semantic Indexing

As an early step in getting away from keyword density, Google started leveraging Latent Semantic Indexing (LSI). It worked by identifying conceptually-related terms to infer page's meaning. This is different from simply identifying synonyms for keywords.

For example, if we are looking at a page that deals with coffee, synonyms might include: joe, java, or espresso.

If we take an LSI approach, however, we start to get a deeper understanding what a page is about.

Page #1:

Coffee
Arabica
Ethiopia
Goats

Page #2

Coffee
Cold
Grind
French Press

We can infer from the differences in these conceptually-related terms that the first page is likely to deal with the origins of coffee as a beverage for human consumption. The second page is more likely to be about different ways to prepare or serve coffee.

Hummingbird, RankBrain, and Voice Search, Oh My!

Not one to call it a day and rest on its heels, Google continued to refine the process of discerning user intent. It did so through a couple of very significant algorithm updates known as Hummingbird and RankBrain. Along the way, the rise of voice search began exerting a very strong influence on the way these systems worked and would evolve.

The Hummingbird update came in August of 2013. In true Google form, however, Google didn't officially announce the change until a month later. The Hummingbird update placed more emphasis on natural language search queries as an evolution to LSI.

In 2015, the RankBrain update rolled out. RankBrain implemented AI technology to monitor and analyze search results. The goal was to continually refine results based on how well those results addressed the search queries that triggered them.

Along the way, a bit of a blessing fell into Google's lap. The rapid increase of voice search via mobile and home assistants revealed that users spoke inquiries in a much different way than they typed them. As it turns out, people speak questions into search considerably more often than they type questions into a search box.

For instance—and in keeping with our coffee theme—someone looking to know about the origins of coffee is more likely to simply type "coffee" into a search box, look for a Wikipedia result, then scan the page to find what they want. With voice search, however, users are more like to search/say, "When did people start drinking coffee?"

By comparing these two different types of search and the user interaction with the search results, Google was able to leverage AI to further establish connections between different search mediums and use that data in an effort to serve the best result. The rise of voice search, then, has given Google's AI the ability to correlate searches across different mediums to better understand a user's search intent.

Enter BERT

In October 2019, Google launched BERT. Undoubtedly influenced by the longer search queries that come from voice search, BERT's goal was, in part, to better understand the relationships between words on a page beyond their conceptual relationships. Google estimated that this algorithm change would mainly affect longer search queries and impact roughly 10% of search queries, which is a statistically massive number for an algorithm update.

What does this mean on a practical level, though?

One simple example is that BERT offers an improvement regarding the role prepositions, for instance, play in a sentence. After all, one preposition can change the entire meaning of a sentence.

"I just heard this rumor from you," is a distinctly different from, "I just hear this rumor about you."

Top of Funnel Keywords

The top of the funnel, sometimes referred to as TOFU, is the initial information-gathering stage. During this discovery phase, searchers often have a problem that needs an answer. Blog posts, eBooks, and overview videos are a few good content formats for this stage of the funnel.

Search queries at this stage will often include words like:

These types of keywords and keyword phrases can act as a signal that someone needs information about a topic before moving further. In some instances, you will leverage these terms into your content in a way that satisfies the search intent but doesn't yield a conversion. That is okay. When done right, these pieces of content either help move users to the middle of the funnel or makes you a valuable resource the next time that user is looking for something related to your products/services.

The goal here is to drive traffic. Driving conversions comes later in the funnel.

Middle of Funnel Keywords

The middle of the funnel (MOFU) is the stage where users have gathered their information and are now considering the acquisition of a product or service. Pieces of content well-suited to this phase can include detailed expert guides, podcasts, webinars, and in-depth whitepapers.

Search queries at this stage will often include words like:

At this stage, you can see the scope of the search narrowing. Initial information has been gathered, but now more detailed information is in play. As expected, your content should start including more conversion-related lingo, as well.

Bottom of Funnel Keywords

The bottom of the funnel (BOFU) is where the rubber meets the road. Searchers have likely decided they are ready to make a purchase or request a service. Content that works well at the bottom of the funnel includes demos, case studies, in-depth product literature, and even incentives to convert now.

Search queries at this stage will often include words like:

Buy
Near me
Ratings
Prices
Coupon

At the bottom of the funnel, you really want to be driving home your unique value proposition and pulling out any stops that might hinder the conversion.

Putting it into Practice

Information and concepts are fine but putting them into practice is when you can start making observations that can improve your efforts. An easy way to start is to look at your existing keywords and see which modifiers would work best with those keywords at each stage of the funnel (e.g., What is organic coffee, organic coffee brands, buy organic coffee).

Then choose a couple of different content formats for each stage. Develop them and publish them.

Finally, keep an eye on your landing page analytics to monitor both traffic and conversions. If the numbers are not hitting your goals, revise the piece or try a new piece in its place.

In Conclusion

As Google's search algorithm evolves to deliver search results more in line with the intent behind a search query, it behooves content creators to keep this in mind when selecting keywords. That said, the ability to craft content that directly addresses the need of searchers and moves them through the funnel has never been more in demand. Shortcuts like keyword stuffing or eyeballing low-competition keywords just isn't enough to yield desirable results. By keeping your organization's unique value proposition in mind and crafting meaningful content, you'll be sending signals to Google and users that your content is the best match.

If you need assistance with keyword research, funnel development, and/or content development, contact Lett Direct or drop me an email directly at geoff@lettdirect.com.