Digital Assistants and the Rise of Voice Search: Shopping in the Future
This is the fifth installment of a blog series that explores the current state of digital assistants and their impact on the e-commerce industry.
Predicting the future is inherently challenging, particularly in the environment of rapid technological change that we live and work in. That said, actively planning for the future can be a useful exercise to spur progress, even if things don’t turn out precisely like we predicted (because they won’t). In this post, we envision what the future shopping experience might look like in a world where voice search and digital assistants play a significant role.
While this series is about voice search, that doesn’t mean that voice ALONE will necessarily drive these shopping experiences. Rather, as we discussed in our last post, voice will complement other technologies that already exist.
We also talked earlier about the importance of screens and how they are critical to most shopping experiences. In the same way you wouldn’t shop for pretty much anything sight unseen (with the Woot mystery box being a notable exception) the introduction of voice alone isn’t enough to dramatically change consumer behavior. Two primary vehicles exist for voice search — through use of a phone (leveraging something like Siri or Google Assistant) or via a digital assistant (Amazon Echo, Google Home, etc).
When voice search is performed on a mobile phone, that device could fairly easily leverage its screen. When done in a home using a digital assistant, other options are available — phones, tablets or connected televisions. Digital assistants are essentially gateways to the internet and could leverage these other screens to create new shopping experiences. For example, that smart TV that you own could provide the visual aspect of shopping using a digital assistant. In many respects, the rise of digital assistants commodifies televisions, which for the most part have not done much with the internet beyond connecting to entertainment services like Netflix, Hulu or Amazon Prime. In fact, Amazon has already started offering TVs that have Alexa integrated.
As a result, shopping might begin with voice search but could ultimately end up heavily engaging other screens. Let the attribution debates begin now!
One of the interesting aspects of the Echo Show is the digital concierge, available as part of the video chat feature.
Imagine “Alexa” is actually a real person who pops up on your screen to assist you with shopping. Essentially, this concierge is personalized, voice-driven live video chat. This feature isn’t appropriate for all companies or categories and clearly has a level of cost associated with it (real “Alexas” require salaries and benefits, whereas virtual ones just need computing cycles). But for categories like health and beauty or apparel (particularly high-end products), this makes a lot of sense. In fact, Amazon set a precedent for this — the Mayday button on the Kindle Fire is essentially a support concierge.
Any service that you would expect from an in-store associate would now be available in your home. Need a pair of shoes to go with a formal dress for an upcoming wedding? Ask Alexa. In this example, the concierge that appears on your screen could ask a few questions about what you are looking for and even view the dress to find an appropriate color match for the shoes. Clearly, this phenomenon would take away one of the value points that brick-and-mortar stores have held onto, which is customer service. While we wouldn’t expect this feature (or even find it useful) for many purchases, it could be an interesting selling model for bigger ticket items.
Browsing by Voice
Another difference with voice search is that consumers search with very different terms from those currently included in retailer’s websites. Retail websites have traditionally been set up to browse and filter based on categories and other product attributes (size, color, gender, etc). Some of these would be easy for voice to interpret but others are much more difficult.
For example, consider the following search command:
- “Show me skirts under $150.” This is fairly straightforward and voice search could interpret that query and return a set of results that match the category of “skirts” and fall under $150 in price. Those results might be returned to a screen attached to a digital assistant or mobile phone.
Refining the search is where it gets harder:
- “Show me more like the first one.”
- “Show me more with a floral pattern.”
- “Show me brighter colors.
- “Show me ones like this but longer.”
What you realise is that voice search is actually very similar to how you would interact with a knowledgeable sales associate in a store. He or she would have visual knowledge of the products and be able to produce a good match. Conversations tend to use more adjectives (“brighter”, “longer”, “floral”) than nouns (size, color, brand). Today’s e-commerce platforms don’t have this capability to make these connections (and retailers don’t have the data) but this capability will be necessary if such usage patterns emerge. It will no longer be enough to have metadata that describe the product (UPC, size, color, brand); there will be a need for data that reflects how a consumer would describe the product.
Predictive Commerce – “You didn’t know but you needed shaving cream”
Finally, as we touched on in an earlier post, the rise of voice search and digital assistants creates a whole new set of data that can be leveraged to predict purchase patterns. We believe that predictive commerce will be enabled and those companies that control the gateway to purchase intent data and have a wide selection of products have a huge leg up here. A significant proportion of most consumers’ spending money goes to recurring items (mainly groceries). As consumers adopt the shopping list features of voice and digital assistants, those lists become gold mines of purchase intent and history. Rather than reminding a user to place an order, the system will detect the need for certain products and just automatically include those products in an upcoming order.
Now even a great predictive algorithm won’t be 100% accurate (human beings are fickle after all) and so returns become a critical part of the solution. Apparel retailers already know that easy, hassle-free returns are critical to meeting consumer expectations these days. E-commerce will evolve to the point where it is just as likely that Australia Post/Toll drivers are picking up return packages as they are dropping new packages off. With return logistics better managed, the marginal cost to a retailer of predicting future purchases drops significantly.
So, voice and digital assistants will drive significant changes to how consumers shop — these patterns will emerge quickly and those who can adapt quickly will be well positioned to win. In our final post of this series we will recommend what brands and retailers can do to prepare for this future. Stay tuned!