What You Need to Know About RankBrain and SEO
Over the last few years, shrouded in the company’s typical mystique and under the cover of darkness (or just the ceiling at headquarters, really) the squad of geniuses at Google have been working on a tool that would help improve search results. In October 2015, Google confirmed the existence of their new machine-learning artificial intelligence system rather whimsically called “RankBrain.”
The powers that be haven’t been too eager to cough up details about their newest innovation, but thanks to a few well-placed sources, some expert analysis, and a video from Google itself, we know the following:
What is RankBrain?
RankBrain itself isn’t the algorithm behind Google’s search results but rather one of the more than 200 ranking factors Google uses to help determine how high websites ultimately rank. There are other pieces that make up the algorithm puzzle, too; from the spam-fighting Panda and Penguin to Pigeon, which specializes in localized search results, to Pirate, which helps unearth copyright violations, Google has a whole host of tools at its disposal. RankBrain is just the newest.
How RankBrain Works
RankBrain was designed with the ability to learn as it goes. When it encounters unfamiliar words or phrases (which make up some 15 percent of the daily queries entered into Google), RankBrain makes intuitive guesses as to what those search words might mean. For consumers, that means more accurate search results even if what they’re typing into the search box is wacky, misspelled, or phrased in an unusual way.
As outsiders, our best guess is that RankBrain is also a cooperative process. As it processes more and more unfamiliar terms and associates them with words already in its vocabulary, Google’s tech team double checks accuracy, makes any necessary tweaks, and then releases the updated system to go learn even more.
Why Does Any of This Matter?
If you care about SEO – and if you have a website or any kind of online presence associated with your business, you really should – you need to know not just your rank but how you got there so you can also figure out how to get better.
As we mentioned earlier, Google’s overall search algorithm contains hundreds of parts. Google calls the interpretative data gathered by these parts or tools “signals,” and they have consistently refused to disclose what those signals are. They have released a few breadcrumbs, though, like the following verified “signals”:
- Bolded keywords
- Keywords within a header tag
- Keywords within an HTML title tag
- Keywords in the main HTML body of the content
- Keywords in the anchor text of a hyperlink
- A site or webpage’s PageRank (an overall value assessment, basically)
And that’s just a hint at the 200+ signals (each of which apparently has 50 some variations for a total of 10,000 or so factors influencing your ranking) that determine your website’s visibility within search results.
Embracing a More Human Approach to SEO
Long-time Google employee Matt Cutts once told an audience at PubCon that instead of chasing Google’s current algorithm, they should be predicting where that algorithm is going to go. With innovations like RankBrain on the loose, Google’s algorithm can now predict what consumers are looking for without having to rely on content stuffed with awkward keywords. It’s the rise of humanized content, and that’s pretty special.
So, be like RankBrain. Learn from the signals your consumers are sending you and keep updating your product, your content, and your overall communication to reflect your findings. Build out your vocabulary. Read between the lines. Write content that showcases your message instead of your target keyword. Connect so you can convert.
Show Google and all your customers that human intelligence can keep up with the artificial kind – at least for now.
http://searchengineland.com/faq-all-about-the-new-google-rankbrain-algorithm-234440, https://en.wikipedia.org/wiki/RankBrain, http://www.bloomberg.com/news/articles/2015-10-26/google-turning-its-lucrative-web-search-over-to-ai-machines, https://www.mattcutts.com/blog/,