What marketers need to know about Google assistant and Google Home
What started with Siri in 2010 is quickly leading to an age where consumers engage with the internet using only their voices, in much the way Captain Picard engaged with the computer on the USS Enterprise.
Google’s foray into voice search has been calculated and planned for years, according to Google CEO Sundar Pichai. It currently appears to be based on a closed system owned and overseen by Google, not on an open system like the trillions of websites that populate the internet are built on (i.e., HTML). I predicted this eventuality more than two years ago, after the Nest acquisition.
These are the problems and challenges brought by Google’s new assistant that marketers and SEOs alike need to be aware of.
Google I/O 2016 announcement
On May 18, 2016, Google announced Google Home, a speaker that houses the new Google assistant (Yes, it’s Google assistant with a lower-case a, not Google Assistant) platform and that resembles the Amazon Echo. The Home device seeks to help users complete tasks in the real world, leveraging Google assistant, which is designed to allow for two-way conversations with people and to actually accomplish tasks like booking reservations, when possible. Pichai said this is like building each user their “own individual Google.”
Consumer adoption of voice search itself started off sluggish, with 85 percent of iOS users saying they did not use Siri in 2013. However, this technology has gained traction in recent years as teens and adults alike began to see past the novelty and embrace the utility of Voice Search’s ability to offer quick answers to questions and directions. By 2014, Google had taken an extreme interest in voice search and released an infographic showing a large percentage of teens and adults felt that voice search was the future.
Today, voice search is bigger than ever. During the Google I/O keynote, Pichai announced that 20 percent of all queries on Google’s mobile app and Android devices are voice searches — and this number is growing.
Voice search problems
1. No data
The biggest problem with Voice Search is that we currently have no way of tracking it, gauging its impact on sales/conversions or understanding how it impacts organic traffic from Google. While there are rumors that Google will provide us with Voice Search/Conversational Search/Google assistant data in Search Console, there is as of yet no method of obtaining such data or gauging the effectiveness of efforts to optimize Voice Search.
Perhaps this will change and become more transparent by the time Google launches the Home device in the fall of this year. However, when this happens, it most likely will not be perfect or open data that can be easily picked up and recognized by third-party applications such as Adobe Analytics.
2. Lack of control
While Google’s goal with Google assistant is to help users complete real-world tasks while simply talking to Google, that may not mesh well with every business model. Already, Google is taking content from publishers and using it in voice search to provide answers without allowing control over the display and usage of that copyrighted content (except to opt out of Google completely).
Marketers have been okay with this so far, as it virtually guarantees them a top spot in Google for the question asked, leading to more traffic in the short term; however, with the growth of voice search and the possibility that it is replacing standard search, this could lead to a reduction of traffic to websites. To publishers who currently rely on advertising to survive, that poses a large risk to revenues.
During the announcement of Google Home, the video showed a young child asking about the number of stars in our galaxy. This answer was provided by Space.com, which Google happily told the child, who then proceeded to ask another question. At scale, this process slows page view growth or decreases page views — and therefore is most likely decreasing advertising impressions.
If your organic traffic is currently coming from offering stats or facts that are fairly unique and Google decides to give that answer via voice, then there’s a good chance that with the growth of voice search, you’ll lose website visits, which could likely bring down conversions/sales.
3. Bad branding
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.”
With conversational search and answers, Google may have issues saying certain brand names out loud and correctly, especially if that brand name is an acronym or a made-up word. (Note: At last check, Google says “ESPN” correctly now.)
Google has not yet announced a solution to this, such as allowing an app or website to provide a phonetically correct pronunciation of the brand name. They might think their machine learning will figure it out from other conversations users have with Google, or perhaps they just aren’t too concerned about it.
Sundar Pichai said during his keynote, “It’s not just enough to give [users] links — we really need to help them get things done in the real world.” Pichai noted that Google has been “laying the foundation for this for many, many years.” This has led Google to develop an end-to-end solution that means customers never have to open an app or visit a website.
One example Sundar gave was wanting to get movie tickets for you and your family. He used an example of engaging the Google assistant in a conversation that resulted in Google purchasing four tickets to “The Jungle Book” and providing the user with a purchase code. Another example was driving in an Android Auto-powered car and asking Google to help you get curry; in this conversation, Google understands somehow that you want take-out and orders your food from an unnamed restaurant, alerting you when it’s ready.
To me, it appears that Google assistant will accomplish its goal of having a two-way conversation and helping users complete real-world tasks by leveraging three sources: Google’s information and machine learning resources, information from other websites and third-party apps. That leaves many questions for marketers:
- Will Google require that apps process orders via Android Pay/Google Wallet in order to be included?
- How will Google assistant handle competing apps? For example, if I ask it to hail me a car, will it ask me if I want Uber or Lyft, or will it just pick one on its own? What if I have Curb or Truck installed?
- Will there be ways to include negative keywords for apps to optimize engagement?
- Will Google allow websites to compete in this environment and offer solutions to help consumers complete real-life tasks, or will they set up a different portal (e.g., Google My Business) just for inclusion in the Google assistant ecosystem? Will they ignore websites altogether?
- How will Google introduce paid listings into the Google assistant? Will marketers be able to pay to be the one item or listing Google provides, will Google assistant offer a paid solution first and then a more organic solution, or will Google take the Google Shopping route of making all listings paid?
This could also mean Google will cut out aggregation and listing websites such as Yelp. Further, it could mean complications for e-commerce providers like Amazon unless they agree to use Google’s system or, if required, pay Google to process the orders made via Google assistant.
VoiceXML, voice-powered browsing & The W3C
Using your voice to navigate information has been a dream for decades, and in 1999, a group of tech companies gathered to create a way for humans to talk to computers called VoiceXML. VoiceXML is now overseen by The W3C, the world’s governing organization of standards for the World Wide Web, which is still updating the protocols, along with what it calls Speech Interface Framework that includes other standards.
VoiceXML was designed to work with Natural Language Processing and to empower voice browsers to retrieve information and read it back. You likely interact with voice browsers powered by VoiceXML on a weekly basis by talking to/yelling at automated phone assistants your bank, utility, travel or phone company uses.
While VoiceXML 2.1 or 3.0 is far from perfect for voice search, it does show that we have a long history of preparing for voice as a user input type. That and the fact that it is an open standard published by the W3C makes me wonder why Google did not work with the W3C to make a standard for voice interfaces using XML or HTML5 that could execute commands on websites and instead are choosing to use what appears to be a closed system that likely requires Google being more deeply involved in a third-party app.
Voice Search and Google assistant are exciting technologies that promise to make searching at home a more immersive and useful experience; however, they come with challenges we as marketers must be ready to face and find solutions to.