originally published in Ad Age
At the Consumer Electronics Show earlier this month, it seemed like every device that debuted was voice-activated. Manufacturers added voice activation to televisions, refrigerators, cars and toys, with Amazon’s Alexa emerging as the predominant platform powering them. Even a hairbrush produced by Withings, Kérastase and L’Oreal has a microphone (to detect split ends), and you can talk to an internet-enabled adapter for trash cans — telling it to reorder discarded items.
Yet calling this year’s CES the “Year of Voice Interaction” or some such misses a critical distinction. Are these devices listening, or are they speaking? A better way to describe this CES is that it was the “Year of Passive Voice”: these devices are passive when it comes to responding to people’s queries. What manufacturers and consumers must determine is how much they want active voice instead, whereby devices communicate with their human owners proactively.
Marketers will benefit far more from active voice, but only if consumers embrace it. In the meantime, a third state has emerged, best described as “engaged voice,” where the consumer triggers the interaction, but the device responds to qualify the request. Today, the term “voice control” implies that people’s voices control the device. Yet passive, engaged and active voice center the narrative around the devices and their software.
To illustrate the three types of voice, consider a simple scenario: Lenny is straightening up around the house, and he is running low on paper towels. He has a voice-activated assistant, dubbed “Device.”
Passive voice: Lenny asks Device to add paper towels to his shopping list. Device responds that it did, and paper towels show up in the list on Device’s companion mobile app.
Engaged voice: Lenny asks Device to order paper towels. Device responds, “You last ordered a 10-pack of Strongwipes for $20. But Ecowipes now has a 10-pack on sale for $18. Which would you prefer?” Lenny makes his choice, Device confirms the order, and the product ships.
Active voice: Lenny is at home, which Device knows because it detects his phone with the mobile app nearby. Device says, “Good evening, Lenny.” Lenny says, “Hi.” Device says, “It has been one month since you last ordered Strongwipes. Can I tell you about a special offer on paper towels?” Lenny tells it to go ahead. Device says, “Ecowipes, rated 4.8 stars, has an offer for a 10-pack of paper towels for $18, which is $2 cheaper than Strongwipes right now. Would you like to try Ecowipes?” Lenny then selects an option.
All of these scenarios are good for marketers, as they reduce friction for commerce. Passive voice is the least intrusive, but it also limits the kinds of value that can be offered through the device. Manufacturers will likely be slow to test boundaries here, worried about the backlash from annoying consumers, or coming across as constantly spying on them. Still, we’ve seen intrusive models on other platforms prevail when consumers feel they benefit. When Amazon experimented with exclusive holiday deals only offered through Alexa, I kept forgetting to ask what they were, but I also wondered if I was missing out. A daily active voice prompt to hear those offers would have remedied that.
This distinction among passive, engaged and active interfaces will extend far beyond voice control. It applies to anything that can be enhanced with machine learning and artificial intelligence. Facebook and Google could proactively tag all your photos with people and locations, and, ultimately, even the objects and brands in those photos. Autonomous vehicles will be able to proactively make assumptions about where its passengers are going, and could recommend other points of interest. Televisions could automatically change channels during commercials (on-screen features powered by Samba TV prompt viewers to do this today) or when there’s some action happening during a sports game.
The age-old question “What’s for dinner?” may well be answered by Instacart, Seamless, Pinterest, or a smart refrigerator proactively making recommendations. We’re not too far off from a smart freezer paging one’s flying, Roomba-like home drone, shuttling the food to the smart microwave, and preparing dinner all by itself. Those smart appliances could probably predict what my toddler will eat better than I can anyway. (If they can get her to eat her vegetables, I’ll give the appliances a raise — or at least a software update.)
For now, our devices aren’t as dumb as they are dormant. Software will keep improving far faster than people will learn to trust it and accommodate it into their lives. When the user experience shifts toward engaged and active approaches, marketers will encounter new bidding engines that build on models from search engine marketing and programmatic advertising. What’s old will be new again.
None of this was overtly on display at this CES. But the new models incorporating active voice will give marketers even more reasons to keep congregating at CES in the years ahead.