Posted on March 27, 2018
“Alexa,” Alex Jones, the conspiracy radio host, growled while holding Amazon’s smart speaker at eye level. “Do you work for the CIA?”
“No. I’m not employed by them. I’m made by Amazon.”
My brother sent me the YouTube video a few days after I got my own smart speaker — a Google Home. “Say hi to the NSA for me,” he added.
As with many new users, the clip hit uncomfortably close to home. I was struggling to answer two key questions about the product: first, how exactly did it work, and, second, how much of my information was being monitored and analysed? And by whom?
The idea that a digital personal assistant like Amazon’s Alexa or Google Home collects and analyzes user data shouldn’t come as a surprise. Every aspect of our digital lives—from Facebook, to YouTube, to Spotify and our banking apps—generates a wealth of personal information, some of which is used to target us with relevant advertisements.
Despite recent fears about data security as a result of the Cambridge Analytica scandal, you can make a strong argument that our present online environment isn’t that scary. We voluntarily give up personal information so that a genuine and accurate profile of ourselves can be created online. This lets us make better use of these tools and provides us with (usually) contextually-relevant advertising.
While it may be unnerving for some users to Google “veterinarians” and then have Google later serve ads for Kibbles ‘n’ Bits, it’s less frustrating than seeing ads completely unassociated with our interests or needs.
While many people are willing to give up some privacy for an easier digital life, Alexa and Google Home raise worries that they collect far more data than users would otherwise wish to give up.
This fear seems unfounded, though. Yes, smart speakers are always “listening” in the sense that they are awaiting their commands (“Hey Google…” “Alexa…”). While they are “listening” in the most basic sense, they’re not analysing or storing every conversation you have within earshot. From a practical perspective, it would be a massive and pointless monetary drain on Amazon or Google to collect and store every word said by users.
In fact, as we can see from ads that have made use of smart speakers, when issues arise, they’re not caused by Google or Amazon making too much use of personal data. Instead the problem is that they are not making enough use of it. So far, we are not seeing evidence that these new digital assistants are having a noticeable impact on ad targeting.
Last year, when users asked their Google Home Minis about their day (a feature that recaps weather, traffic, top news, and your personal calendar), Google included a short teaser for Disney’s Beauty and the Beast. Google claimed this wasn’t so much an ad, as it was a way to keep users up-to-date on recent events. However, the idea that the release of Disney’s Beauty and the Beast was one of the top news stories of the day didn’t wash.
Instead of targeting these ads to users who might actually be interested in seeing Beauty and the Beast, Google served the ad broadly to its users, upsetting many people who found the teaser’s presence jarring. Google’s first misstep with its smart speakers wasn’t the result of “listening” or “monitoring” users, collecting their data, and using that data to target ads towards them — it was because they didn’t.
Had Google used its vast user-specific data to target the ads, it’s conceivable that users may have considered the ad more appropriate and not an intrusive attempt to earn marketing dollars.
The Beauty and the Beast spot didn’t follow the approach that has been so successful for Google’s search advertising approach. Nor did a recent foray by Burger King.
Burger King ran a television ad in which an actor claimed there wasn’t enough time to list all the ingredients in a Whopper, so instead called out “Hey Google, what’s in a Whopper?” Viewers who had Google Home close enough to the TV were inundated with the digital assistant’s robotic voice reading off the Wikipedia article for Whoppers. This, understandably, frustrated some users. Google quickly patched the Google Home software to stop the readout.
Again, the furor was caused not from too much use of personal information, but by using no personal information. Users who had no interest in learning about the Whopper—no matter how clever they found the ad—were frustrated at having their homes and smart speakers hijacked by advertisers.
There are a lot of ways to effectively and inoffensively target advertisements towards potential customers. Key among those is ensuring that ads are useful to those seeing them. While it is a common refrain that people are “sick of advertising”, no one appears to be sick of consuming. As a result, no one minds ads for things they actually want to buy. Google already knows this and has shown how successful targeting ads in this way can be.
Companies risk further frustrating customers if they don’t integrate personal information into smart speaker advertisements. Simply put, less information means less accurate ad targeting. If, instead, companies make full use of the information they already have, they can make advertising less intrusive, getting the products users want in front of the customers who want to buy them.
Views expressed are those of the author and may not represent those of Navigator or its affiliates.