In a frustrated early testing of Alexa, Bezos allegedly told the prototype to “shoot yourself in the head”. (Photo by Business Insider)

Just this week, Amazon began another round of layoffs as a part of its plan to cut 18,000 jobs. One of the main casualties? The Alexa team — with over 2,000 employees in this division losing their job.

Alexa was a calculated risk taken in order for Amazon to tap into the world of voice-assisted technology. We’ve seen these kinds of leaps into unfamiliar tech fail among consumers, such as Meta’s recent launch into virtual reality, but the Alexa story is different.

When Alexa was first launched, it performed better than ever — indicating that adoption wasn’t the issue with customers. By 2019, the company had sold over 100 million devices.

The question remains: If Alexa sales are high, how come it’s never reached profitability?

The Need for User Testing

Let’s take a look at Alexa’s business model.

Alexa creates revenue by responding to “requested actions”, specifically those centered around closing a sale. (Photo by AV Magazine)

Amazon’s Alexa is sold at around the same amount it takes to manufacture the device. This was a strategic attempt to make Alexa’s accessible to every household, but one that relied on a user experience that has yet to come to fruition.

“Alexa order me a Dominos pizza” — is exactly how the company foresaw user interactions to take place. Unsurprisingly, this type of requests aren’t very frequent. The top requests of Alexa are much more simple — to set an alarm, play music, or check the weather.

Alexa’s business model was crafted around the expectation that the ease of voice-assisted technology would motivate users to spend more. Instead, some have called the product “a glorified alarm clock”, mocking how plain the functionality of the device is in comparison to its price tag. In theory, ordering a pizza with Alexa is probably much quicker and easier than searching online, calling a store, and listing off your credit card information and yet this is exactly how consumers have been conditioned to think for decades. Trusting a device to do such a simple job has more risk than reward.

This appears to be the issue with all voice-assisted devices. The top performer, Google Home, has seen sales decline at a yearly rate of nearly 20% since its release.

As we’ve seen in chatbot technology, consumer spending is highly dependent on how users perceive the assistant’s reliability. In a study by Nielsen Norman Group it was discovered that many Alexa and Google Home users assume the devices have “a low level of competence for complex tasks”, even finding them “socially awkward” to interact with.

The same study found that users felt Alexa failed to learn anything about their behavior. AI is often cherished for its ability to become smarter with more and more user interaction, and yet very little Alexa users noticed increased personalization despite regular use over a period of months.

Often users complained of the consequences of Alexa’s misinterpretation of their requests, such as inappropriately calling the wrong person or making an incorrect purchase.

There may still be hope for Alexa

Despite recent layoffs, Amazon still holds high expectations for Alexa.

Senior Vice President David Limp has argued that Amazon’s CEO isn’t threatened at all by the loss of revenue, stating: “I’ve yet to be in a meeting where he doesn’t call out those as big inventions and big bets”.

It’s tough to predict where the Alexa project will be in the next decade, but it may have negative implications for the AI space. Such a huge investment in low-functionality AI may cause consumers to be wary to buy any more developed technology in coming years, especially considering Alexa’s user feedback.

Amazon will have to be more creative in incentivizing users to spend via their Alexa, perhaps with a rewards system in which benefits trump the risk of trusting the device with such sensitive information.

For more info on the recent Alexa turmoil, check out the following articles: