Researchers estimate that voice shopping is a market that’s going to explode in a big way over the next couple of years, hitting more than $40 billion in 2022, up from just $2 billion today. That’s according to a market research report released in March from OC&C Strategy Consultants.
The Information, citing two people briefed on internal Amazon figures, has reported that only a very small fraction of people who use Amazon’s voice-based assistant which powers the company’s Echo line of speakers actually use their voice to buy things. That number is about two percent of people who own devices powered by Alexa, of which there have been apparently about 50 million that Amazon has sold.
Of the people who did buy something using Alexa voice shopping, about 90% didn’t try it again, one of the people said. A larger number, 20%, have engaged more broadly with Alexa voice shopping by using commands like “What are my deals?” and “Where is my stuff?” to track orders that were likely made on other devices.
Why voice shopping shows high abundant rate?
While voice is a natural communication method for humans, it is a new medium for marketers, product managers, and developers. A common mistake businesses do when they try to make the transition to voice, is to mimic the web experience they provide for their users into the voice channel. The well-known web – user experience patterns the industry developed for the last twenty would not work in voice, and we need to develop new ones. The best way to do that would be to look back on how we would build a similar experience when we train our employees (sellers in this case) to talk with our customers.
How do Amazon Alexa and Google Home sell you today?
Selling using pure voice has many challenges, one of them is the lack of visual. How can I buy a new dress if I don’t see it? But what about all the day-to-day cases where we buy things where we know well what we want to buy and usually case less on visual? Let’s run a simple example where I want to buy AAA batteries. In my case, I don’t care much about the brand, but since I don’t use those batteries often, I might be sensitive to price.
I told Amazon Alexa that I want to buy batteries. Note – All screenshots are taken from Amazon Alexa simulator.
As you can see, Alexa offered an option, when I said it’s not a good fit for me, it offered the second best option. This behavior mimics one to one the experience Amazon provides over the web. Basically using a sophisticated search engine, that analyses my request and provides me with a list of options to choose from.
Amazon is probably the best at guessing what their users want to buy, however, while this approach provides a great web user-experience, it provides a poor user experience on voice.
Let’s see another example, in this case, I will ask explicitly for AAA batteries
As you can see, providing a simple extra information changed my experience from one extreme to the other. Now I quickly got exactly what I wanted, and was able to complete the purchase in a few seconds.
Using “Guess Who?” strategy to master voice shopping
“Guess Who?” is a two-player character guessing game. Each player starts the game with a board that includes cartoon images of 24 people and their first names with all the images standing up. Each player selects a card of their choice from a separate pile of cards containing the same 24 images. The object of the game is to be the first to determine which card one’s opponent has selected. The game has a simple winning strategy, that can be easily learned by 4 years old. Ask questions that with any possible answer will exclude as many possibilities as possible. For example “Does your person is a man?” or “Does your person wear a hat?”
How is this relevant to voice commerce? In the web scenario above, we searched for “battery” and Amazon provided a list of options for us to select from. Think of an opposite definition of the problem – the users “think” of an item they want to buy (“AAA batteries” in this case), and Amazon should “guess” what they wanted using the least number of questions. This can be achieved with the following simulation:
Instead of providing the next option, we use the “Guess Who?” strategy and follow with a question that will significantly reduce the set of options. “What type of batteries are you looking to buy?”
A simple followup question changed the experience from a poor one to an excellent one, that ended with a successful sale. This requires some deeper domain expertise than a simple search, in this case, the different types of batteries, however, as you can see, a deeper domain expertise, is a key success factor in voice commerce. When approaching Voice UX, we must understand that the “old” web methods are not applicable. We must reach out to the basics of human-to-human conversatons, and try to created an automated conversation that would mimics them.