Skip to content

Creating Content for Google’s RankBrain

Content-for-Google's-RankBrain

Google revealed in October that it uses artificial intelligence to help with 15% of search queries. Named RankBrain, the system analyzes vague, ambiguous queries and matches them with the most relevant results.

It’s a critical component to the search-engine giant’s algorithm. In fact, Google’s Greg Corrado told Bloomberg that RankBrain is now the third-highest signal contributing to a search-query result.

RankBrain is now the third-highest signal contributing to a search-query result says @google Click To Tweet

Google – and similar search-engine services – are getting smarter. As marketers, we no longer can rely solely on traditional digital strategies such as link-building or social-media signaling. We need to be far wiser. How? By first analyzing how Google’s current algorithm components work in conjunction with RankBrain.

Basics

Google uses more than 200 search signals to determine the ranking of websites and content. Given that RankBrain is the third most important signal, artificial intelligence has indeed revolutionized the way the search-engine giant understands long, complex queries. Users can now feel confident that they’ll find appropriate topics no matter how ambiguous their search terms are.

To give you a clearer picture of how good RankBrain is, here’s Google’s algorithm in a quick test, using a popular industry phrase, “conversion optimization.” According to AdWords, the term generates about 780 average monthly searches.

Google-AdWords

In a split second, Google displays results that mostly refer to the basic definition of conversion optimization (not bad). But let’s make the query more specific – “best conversion optimization tool.” Google immediately produces a list of articles on the top tools to use for conversion optimization.

Now, let’s use a query that a Google signal analyzes as natural sounding, “what is the top conversion optimization tool for marketers?” The results took 10 more milliseconds to produce the related results, but again, it’s impressive.

But to really gauge the capabilities of RankBrain, let’s turn the query into something vague like “shape of conversion optimization in the future of digital marketing and beyond.” This long-tail phrase is confusing and presents numerous possible answers. Of course, the gist of this complex phrase is to ask about conversion optimization predictions – and Google’s RankBrain does its job.

Google's-RankBrain-algorithm

Click to enlarge

It’s amazing how the search engine’s algorithm was able to take a bunch of complex words and present helpful results.

But how does this help content marketers generate content? What should they watch out for? What points should they keep in mind when crafting content that will satisfy both people and RankBrain?

RankBrain and content

RankBrain’s main purpose is to help Google analyze and produce relevant results to long, complex, and ambiguous search terms. This includes new “words” or jargon made up by industry experts or popular figures, queries that it has never encountered.

Unlike humans, machines can’t read for context and comprehension without relying on complex mathematical formulas. Remember the results for “shape of conversion optimization in the future of digital marketing and beyond”? We used the query to seek information about forecasts about online marketing, but we phrased it in such a way that it seemed vague for a machine. Still, RankBrain was able to take the context from the query and present us with satisfying results.

Another example comes from Jon Earnshaw who performed a voice query as detailed in Pi Datametrics. (Voice-based search queries are more likely to result in vague phrases because they are stream-of-consciousness speaking, not typed queries that you can edit more easily.)

He wanted to know about “the name of the fat guy with the beard in the movie with the four guys in Vegas?” which is pretty convoluted, especially for a machine. Not only are there countless films about Vegas, most of them have four guys.

Voice-query-results

But the machine does an amazing job at recognizing that Jon wanted results featuring The Hangover.

How should content marketers create content that will be valuable to this artificial-intelligence signal?

  1. Create content for a specific audience

As RankBrain is all about helping Google pull up more appropriate results, your content should be geared toward particular audiences to increase the chances of it appearing for related searches. If you sell digital-marketing services for instance, your audience could comprise entrepreneurs, brands, small businesses in almost any industry. Narrow your audience to a target and create content for that niche by:

  • Answering problems that your target audience may search online
  • Aligning your products and/or services to their unique needs
  • Grabbing their attention using specific topics, keywords, and recommendations

For example, say you have a buyer persona called “Roberto” who owns a suburban café. Tech savvy, he employs paid and organic methods to help his website gain leverage. Although he knows a bit about digital strategies, he wants to find out more and searches Google for “as a cafe owner how can seo help”.

CafeOwnerSEOSearch

Google is able to drill down on what Roberto really wants – RankBrain recognizes café as a type of restaurant and interprets it as a “local” restaurant. A site called Restaurant Engine provides the first two results, 5 Local SEO Tips for Restaurants and Restaurant SEO: A Guide for Restaurant Websites. Note that Restaurant Engine sells website-development services for the restaurant industry. While it likely could design websites for many industries, it’s chosen to focus on the restaurant niche and deliver content valued by that audience, and its search engine rankings reflect that.

  1. Craft content that sounds natural

As search engines become more sophisticated through artificial intelligence, people are no longer bound by rules to search for answers. They don’t need to think about how to phrase something to get the most helpful results, they can just “talk” to the search engine and get the answers to what they want to know.

These are examples of natural language queries; and search engines like Google are getting smarter at deciphering them. To match your content with what RankBrain is looking for, include answers to problems written in natural language.

RankBrain is looking for answers to problems written in natural language says @alseoblog #SEO Click To Tweet

For example, a manager for a company uses a conversational sentence in the search box: “I rely on reviews to help my business.”

NaturalSoundingQUERY

RankBrain most likely interprets the word “help” in that context to reveal results that focus on how to get and use “positive” reviews.

Now, let’s look at the language used in the first result. Groove HQ author writes in a friendly, natural-sounding way. He uses second person: “If you want that help, you need to ask for it.” He doesn’t rely on formal language, “range from dirty and deceptive … to simply useless.”

NaturalSoundingExample

The author’s voice is clear, descriptive, and sounds like he is talking to someone in front of him. RankBrain loves this type of content because it’s easier for most audiences to understand – something Google knows its searchers find valuable.

  1. Make more detailed long-form content

According to Steve Baldwin of Didit.com, RankBrain uses co-occurrence to help deliver the most relevant results to users. Co-occurrence is the frequency in which terms or related groups of words appear in a given material. This means that if you’re searching for articles on “boosting sales using SEO” for instance, Google – with assistance from RankBrain – produces content results that contain related terms and treats them as significant to a person’s query.

LongFormContentRESULTS

Let’s analyze the first result from QuickSprout:

LongFormContentEXAMPLE

Using “boost sales” as reference, we can spot several words that are synonymous to the original term (i.e., monthly revenue, income, increase your sales, get a few more sales) within the first four paragraphs. This co-occurrence – along with other ranking signals – makes it easy for RankBrain to treat this particular content as relevant to the query.

Long-form content allows content marketers to naturally implement co-occurrence without sounding redundant. It doesn’t mean that thin content is out – but the longer your work, the more chances you get to sprinkle in related keywords.

  1. Incorporate references from case studies, white papers, and authoritative websites

As mentioned, one of RankBrain’s goals is to refine search so that users will get the most relevant results based on their intent. That means search engines will get better at reading between the lines. For example, Google’s complex algorithm signals alongside RankBrain can interpret whether a searcher who uses the word “consumer” means someone who purchases goods or someone who consumes food.

When you create content, be sure to include references from high-authority sources in case studies or white papers. Involving leading industry experts will help the search algorithm translate the intent of your content and reveal why searchers would find the content not only credible but useful to what they want to know. In brief, ensure that your content reflects that you:

  • Researched your topic well
  • Presented facts to support your statements
  • Boosted credibility as a knowledgeable source for a particular topic

If you look up the phrase “importance of content marketing in 2016” for example, you’ll see the top result is CMI’s B2B research report on content marketing in 2016.

Incorporate-references

Google recognized that the searcher using that query might be interested in a complete report showing pertinent data about content marketing trends, statistics, predictions, and samples.

Conclusion

Artificial intelligence or machine understanding is only going to grow in importance. Whether it uses RankBrain or a new AI signal, Google will continue to rank websites based on how people respond to them; and good content with good links are the basics of what most users are looking for. In other words: what people want, is what Google wants.

Stop thinking of SEO as a game where you act like a bot to beat the bot. Craft content for a specific type of searcher in a conversational tone. Leverage long-form content to incorporate multiple phrases and words in a conversational way and use authoritative sources to further delineate how valuable your content will be in relevant searches.

By developing and improving your output as content marketers with RankBrain’s effect in mind, you’ll begin seeing better SEO results.

Want to improve your SEO and other content marketing skills? Register today for Content Marketing World this September. Use code BLOG100 to save $100.

Cover image by Kevin Phillips via publicdomainpictures.net