Whenever trying to understand (or explain) the many, ongoing algorithm changes that search engines and apps make, I always keep one undeniable truth: no matter how they evolve, search engines are ALWAYS trying to deliver the best and most accurate results to searchers — often before the person even knows they want said information.
Even though Siri officially launched in 2010, only recently has voice become a dominant search method. (In fact, 85% of iOS users said in 2013 that they didn’t use Siri at all.) Thanks to increasingly busy schedules, growing impatience and a strong need for efficiency, voice search has now become more popular than ever before. During a recent conference, Google’s CEO Sundar Pichai stated that 20% of all queries on Google’s mobile app and Android devices performed by voice search. (Not seem like a lot? Perhaps the fact that Google gets 2 TRILLION daily searches will help put the sheer size of that statistic in perspective.)
The search engine giant has been carefully calculating their official voice search solution for years and finally has an answer to Amazon Echo. Say hello to Google assistant: a new digital platform designed to turn search into a two-way conversation and help people complete tasks in the everyday world. (And yes, the lower case “a” is intentional).
So what the hell does that mean, anyways? Search Engine Land’s Danny Sullivan explanation (seen here) is spot on. “Google assistant combines two things: Google’s expertise in extracting information from content across the web and from partners plus its machine learning smarts to understand what people are asking,” Sullivan said. “Put another way, Google search has been largely a way that people typed queries in a one-way conversation to get information that they themselves used to complete tasks. Google assistant is going beyond that, to a two-way conversation, one that aims to fulfill tasks as well.”
Basically in short: soon you will be able to make boozy brunch reservations simply by asking Google assistant instead of actually typing anything in. Convenient, right? Maybe for the every day joe, but marketers know we never get off that easy when the Internet of Things decides to update itself.
The platform’s infant status means that Google still has a lot of testing to do and kinks to work out before they truly understand its intricacies, let alone clearly explain it to marketers and every day searchers. Even industry insiders who’ve been keen to the inevitable over the last few years were forced to wait, only able to observe and react to the impact after the fact. After a week of research and trolling the blogs of marketing experts, I’ve found two blaring red flags with Google’s inaugural attempt at voice search.
No Data = No Accurate Learning Curve
Even though rumors of built-in voice search data are swirling, as of right now there is no way to accurately measure analytical results with voice search. Without metrics, it is impossible to learn how the platform works and adjust for its impact. Conversely, it will also be very difficult for marketers to explain to their clients their voice search results — especially when the data either isn’t available or not advanced enough to provide answers to analytical questions.With that being said, I have no doubt that Google will have some form of data analysis available when the Google Home device (with Google assistant included) becomes available for purchase this fall. It will likely have several bugs and third-party integration issues to be worked out, but only time will tell.
Voice Search Has Issues Recognizing Brand Names
While there is no denying that voice search has come a long way since Siri’s 2010 debut, it is not nearly as mature as it should be (yet). Not only does the platform need to account for different dialects and accents, it also must understand words that are common but not found in the dictionary. Search Engine Land’s Joe Youngblood pointed out that branded terms or acronyms are still a major challenge for voice search platforms. He used the following example: “”Two years ago, I asked Google a question about an NFL player’s stats. The answer Google gave me was originally written by Rotowire and published by ESPN. When it cited the source, Google called the website “Esss-Pen.com” instead of “E-S-P-N.” Even though Google now correctly says “ESPN,” the example still makes perfect sense (and is missing a solution.)