Amazon’s Alexa and similar voice-activated assistants have been a big hit, but few consumers use them to shop. Here are some reasons why, and some steps that retailers and technologists are taking to overcome the obstacles to conversational commerce.

Dan Levy, editorial director, Smooch

Dan Levy, editorial director, Smooch

You’ve seen the stats about voice assistants. Nearly one in five U.S. adults had one in their home last year. Half of all searches will be voice-based by 2020. More than 100 million Alexa-enabled devices have been sold. Sounds impressive. But despite the hype about voice-enabled devices powering the next wave of conversational commerce, the question remains:

Why aren’t people actually using them to buy stuff?

Opening Pandora’s Box

 Less than a third of voice-activated speaker owners used them to make purchases last year. As shown in Smooch’s “2019 State of Messaging” report, consumers use their devices to play music, check the weather and get their news, but few are using them to shop.

They could just say “milk” and the assistant would know they mean the “1 gallon of 1% Great Value organic milk” they ordered last time.

Some of the largest music streaming services are hoping people will change their tune. In April, Pandora announced it would be launching voice-enabled ads that will let listeners ask for more information, place an order, or skip past the ad via voice commands.


The idea, according to TechCrunch, is to allow advertisers to target people “at times they aren’t typically able to respond—like when they’re out running, at the gym, driving or cooking.”

Now Spotify is joining the fray, with voice-powered ads by Unilever set to debut on the platform this month. By encouraging consumers to engage with an ad’s call to action in real time, Pandora and Spotify are hoping to help marketers overcome one of the major obstacles of voice-based advertising: Unlike web and mobile ads, which can be measured through things like impressions and clicks, traditional audio ads aren’t clickable.

OK Google, Got Milk?

Walmart is taking a more direct-to-consumer approach to voice commerce. The retail giant announced a new grocery-ordering capability on Google Assistant-powered devices.

After initiating a conversation by saying, “OK Google, Talk to Walmart,” users can order items immediately or add them to a shopping list as soon as they realize they’ve run out—an experience that better reflects how people shop for groceries. The bot maintains context by keeping track of previous orders so customers don’t have to repeat a product’s full name, as TechCrunch explains: Instead, they could just say “milk” and the assistant would know they mean the “1 gallon of 1% Great Value organic milk” they ordered last time.

The Diversity Dilemma

While these developments are promising, voice commerce faces many of the same challenges as the uptake of business messaging:

  • Disappointing AI (the key to scalability).
  • Questions around discoverability (those interactive ads should help).
  • Growing concerns around consumer privacy (apparently Amazon may be listening after all).

Voice assistants have also been accused of perpetuating the inherent biases of their Silicon Valley-based developers.

For a Washington Post study, Researchers tested both Alexa and Google Home to see how users with different accents interacted with the devices. They found that people with Southern accents were 3 percent less likely to get accurate responses from Google Home than those with Western accents.

Alexa, meanwhile, accurately responded to people with Midwestern accents 2 percent less than those from the East Coast.

Non-native speakers had the most frustrating experiences. People whose first language is Spanish, for instance, were understood 6 percent less than those from the West Coast—which happens to be where Amazon, Google and most other tech giants are based.

Although the gap may seem small, the report suggests that as voice technology spreads, this can become a significant handicap, exacerbating issues like wealth inequality and social isolation.


Right On, Q

Software makers have also come under fire for programming their voice assistants with women’s voices, reinforcing perceptions of “female servitude,” as The Next Web puts it.

To fight back, a team of linguists, sound designers and activists have created Q, which they’re calling the world’s first genderless voice. They recorded the voices of two dozen people across the gender identity spectrum to arrive at a frequency range most people perceive as gender neutral (take a listen, it’s pretty trippy). The group is lobbying Apple, Amazon, Google and Microsoft to integrate Q into their voice assistants.

It won’t solve all our Big Tech problems, but it’s a step toward making automation more inclusive.

Smooch provides a platform that connects business software to a variety of messaging systems.