For years Google has been working to provide better search results for their users. Better search results means a better overall experience, which can directly translate into bigger profits for Google (people keep coming back to a search engine that provides helpful results, which means Google can sell more ad space at a higher price around these results). We’ve seen algorithms such as Panda and Hummingbird rolled out, tweaked, and refreshed over the past several years–and these algorithms have been more and more integrated into the overall ranking algorithm that Google utilizes to rank pages based on search keywords and phrases.
The search giant reported earlier this year that it handles over 100 billion searches per month… That’s a ton of volume! However, Google’s algorithm can stumble a bit when people type in a completely unique phrase or question that Google has never received before. According to a recent report, around 15% of the 3.5 billion daily searches that Google receives every day have never been seen before (that’s 525 million searches per day!), so Google needs to figure out how to best serve up results for these new searches.
That’s where RankBrain comes in. Google has been working on developing and integrating artificial intelligence into its algorithms, and RankBrain is a step in the right direction. The way people search online has evolved over time, and more and more searches are taking the form of questions (i.e. “What are the best places to eat around here?” instead of “best restaurants San Marcos, TX”), so Google has been working on a way to parse these queries to understand search intent and serve up helpful results. When analyzing these search queries, if the algorithm sees a word it doesn’t recognize, it attempts to take an ‘educated guess’ at what the word might mean, based on other words and phrases that may have a similar meaning. It then serves up the results that it believes to be most relevant, and the monitors how the person browses and responds to the results offered. If the user doesn’t appear to find the information they’re looking for (a simplistic example: they click the first few links, leave the websites quickly–indicating the website likely didn’t have the information the user was looking for–and search again using a different phrase to find the information they want), the algorithm makes adjustments accordingly.
We find this fascinating at tekRESCUE, and we’ve been waiting for Google to take a true stab at semantic search, because we know it’s been coming for a long time. We’re excited to see these AI changes roll out for more applications such as backlink analysis and better spam detection. This will make “cheating the system” much more difficult for spammers and black hat search engine optimization techniques. Every day we come across competitors of our clients who have spammy backlink profiles, yet they are still ranking well. Google has attempted to thwart this kind of manipulation for years, with varying degrees of success. We don’t use spammy SEO tactics for our Austin and San Marcos, TX SEO clients’ websites, so we are confident that these changes (as they slowly roll out) should only bolster client rankings as high-ranking competitors using spammy tactics are weeded out.
Honestly, there’s much more that this will revolutionize than just search engine results. Seeing this type of artificial intelligence rolled out into a multi-billion dollar company’s core framework shows that we’re taking steps in the right direction to utlize sophisticated AI solutions to make the world a better place for us all.