60% of owners of these devices say their results to spoken queries are often inaccurate. But nearly two-thirds of owners 18-49 year olds use them to research products and make purchases.

Google Home and Amazon Echo have introduced voice search into homes and with their popularity growing, it’s only a matter of time before voice search becomes just as popular, if not more so, than search in written form. There is seemingly less effort involved when asking Google Home for a recipe and hearing it than looking up that same recipe on a screen. And while there is little question that voice search will inevitably be adopted by the masses, its implications for retailers and e-commerce, at least in the near term, are still unclear.

To gain some understanding into consumer behavior with virtual assistants, we conducted a study asking 500 owners of either Amazon Echo or Google Home how they use their devices. The first insight we uncovered was fairly telling. Just 60% of those with a virtual assistant found them to yield accurate results. This means that a little over half of those who own a virtual assistant find them to be useful. Clearly, Google and Amazon have a lot of natural language processing work left to do.

However, there is no doubt that they will do the work and that they will do it well.

Get in Early

For retailers, this statistic is good news. The nascent nature of voice assistants and their current limitations present retailers with a great opportunity to get ahead now, while the barrier to entry is relatively low and they can set the rules.


One key recommendation for retailers as they start exploring entering the voice search field is enhanced structured data. Structured data is important for desktop and mobile websites as well, but it will become very important when it comes to voice. The user will be relying on the virtual assistant products like Echo or Google Home to relay the correct information to them when asking a question or giving a command. The search engine needs to pick up on contextual clues and keywords in a website’s content in order to provide the most relevant answer. Think of structured data as those blatant clues in your code around your content that make it easy for Google to discern the information you’re after.

Become Familiar with Conversational Search

Aside from discovery, users are also beginning to make purchases on their voice assistants. According to our study results, 51.5% of those with a device use it to research or make a purchase. When looking at age demographics, the number rises to 65% of 18-49 year olds using voice search to research products and make purchases.

I am learning how to communicate with these devices. Retailers should do the same.

These numbers will only rise over time and, in the meantime, Amazon Echo and Google Home are training users to search in a conversational way. When using one of these devices, I’m very conscious of how I phrase my questions so that the device understands me. I am learning how to communicate with it. Retailers should be doing the same.


To get ahead of the curve, retailers need to get in front of voice search by creating conversational content. For example, phrases, questions, and sentences—not just keywords. With desktop searches, people would type in “chocolate chip cookie recipe.” However, with voice search, someone would ask their assistant, “How do I make chocolate chip cookies?”

Since advertising does not yet exist on voice search, it’s hard to yet say how it will manifest. The text search advertising models of impressions and PPC [pay per click] are not easily translated to voice search and virtual assistant products. It’s possible that new monetary attribution models, such as Pay Per Transaction, will be developed and retailers will need to adapt their SEM [search engine marketing] practices to these new models as they arise.

Focus on Apps and Local Search

Finally, the question about what people are buying. In our study, we asked for a breakdown of purchases on Amazon Echo. Unsurprisingly, 25% of purchases are going towards entertainment and media. Since this is the Amazon model—Echo was created as a home entertainment device—the most popular purchases are for just that purpose. The other purchases mostly relate to the popular items that are bought through Amazon that seemingly don’t require much browsing or research. Twenty-one percent buy household items and 18% to apparel.


Given the nature of voice search purchases and the lack of screens for fast price comparisons, we predict that big brands will likely profit most from initial voice advertising efforts. A complete breakdown can be found in the chart below.

Since advertising is still not an option on these devices, marketing efforts should be focused on apps. Amazon Echo is very app-reliant, and has already partnered with third-party vendors such as Starbucks, Uber and Domino’s. Consumers are already getting and responding positively to push notifications on their phones for things like coffee. As these devices become more mainstream, more retailers will be looking to integrate their apps.

And while the bigger brands have the budget and resources to get in with Amazon at the moment, there is still an opportunity for smaller business can take advantage of this market. Smaller brands can be at an advantage by having a presence for “near me” searches. By optimizing for local search on their sites and local business listings and managing customer reviews regularly, smaller retailers put themselves in a position to stand out when more consumers start asking their Google Home device questions like “Where should I eat dinner tonight?

Search went through some big changes last year. With the arrival of voice assistants, even bigger changes are on the horizon. Retailers need to start paying attention now, before these new practices become mainstream and become a big driver of consumer behavior.


NetElixir provides search marketing services to nine of the Top 1000 online retailers in North America, according to Top500Guide.com