As more brands test out voice technology, voice data will be key to measuring success. Columnist Chris Wareham breaks down the metrics to use and explains why they should complement, not displace, your existing data sets.
Despite all the advancements we’ve seen in computing, the way consumers interact with their devices and the internet has remained somewhat static. The keyboard gave way to touchscreens, but we’ve consistently relied on touch as the primary interface. This is poised to change in a dramatic way, however, as voice technology — one of the natural forms of human communication — begins to play a major role in a consumer’s life.
The numbers are encouraging: Since last year, online sales of Amazon’s Echo devices and Google Home have grown 39 percent. At the same time, we’ve seen stats showing that 20 percent of mobile queries were made via voice in 2016, with accuracy now at about 95 percent.
As consumer habits change, brands will need to stay ahead of the curve to survive and thrive. If mobile and video are any indication, one thing is for sure: Technology trends that catch on often spread faster than a brand’s ability to effectively address them.
Those that deliver the best experiences early on will have a significant advantage. Consider pizza, for example. If getting delivery is as easy as saying a couple of sentences, it makes that experience much more tempting than one requiring an app or getting onto a computer.
For brands currently experimenting with voice interactions or exploring the possibilities, the key to success lies in data. Those brands that properly capture and analyze customer behavior, while leveraging it to drive action across all channels, will have a significant advantage over the competition.
Here are a few guiding principles to keep in mind:
Voice data is inherently more complex
Organizations accustomed to measuring things like clicks and visits will have to reorient their thinking and make sure the proper measurement framework is created. Voice data may appear unwieldy and unpredictable, but it can still be broken down and quantified:
- Intent and parameter: Voice interactions can be broken down into intent (“show me the cheapest flight”) and parameter (“to Los Angeles in August”), providing a baseline into customer behavior.
- Error rate: Given that voice technology is still maturing, creating an “error rate” metric will let you see where users are hitting roadblocks.
- Pathing: Because of the importance of great omnichannel experiences (we discuss this in more detail later), the idea of pathing allows brands to see what users (tied to a specific ID) do following voice interactions.
- Other metrics to capture: For a full view on how customers are engaging via voice, other data points to capture include session length and frequency of usage. Additionally, there are more macro-level stats regarding the number of users and sessions that can be tracked to determine which skills are getting the most attention.
Voice data has limited value on its own
One of the biggest challenges facing brands is how they can engage customers as people and not devices or channels. The best desktop experience can’t make up for a poor mobile experience, and customer loyalty will suffer as a result.
In this context, it’s critical for brands to ensure that voice data is not captured and analyzed within a silo. Voice insights have to complement and inform existing data sets, from desktop to mobile app and web.
When done correctly, the output is a seamless customer experience that is intuitive, more contextually relevant and better personalized. For instance, take a travel company that enables bookings and price comparisons via a voice-enabled device or digital assistant. If they know a customer’s points status, they can offer it in real time to cover costs for flights that are being brought up. And instead of mailing free drink coupons, they can offer them upfront to the most loyal customers, as a mobile email screenshot that can be used in-flight.
Personal finance is another area with great potential. As consumers increasingly move online for their banking needs, it’s diminished some of the ability for brands to offer custom services and promotions that were easier to deliver in person.
Let’s consider a brand that allows their customers to complete tasks via voice, such as checking balances or making payments. When tied back to the customer’s existing profile, the voice device can provide the option for users to hear customized promotions they qualify for. Or, if they see that a particular customer is prone to overdrafts, they could offer protection services that best suit the user.
Experiment and iterate
All practitioners fear missing the boat on an emerging trend. The last time this happened? Mobile. A lot of brands made the mistake of rushing to put out a mobile app and hoped it would catch on. As we’ve seen with conversion rates in retail, for example, the strategy didn’t pan out for many.
The opportunity in mobile was to engage customers in ways that were unique to the new medium. Those who simply ported over their web experience to a mobile app missed all the richness of the platform, such as the immediacy and contextual location data that apps provide. Brands can avoid that by taking on a spirit of experimentation and iteration.
The goal post is not reached with major leaps, but small tweaks driven by good data. When it comes to voice, don’t make the mistake of porting over the most popular features on desktop web and mobile.
Take a moment to consider the form factor and the context, identifying the one or two hero features that make sense for these devices. Push out the skill and measure constantly. Set up a cadence where constant tweaks can be made, and be at peace with skills that have to be canned. No matter what, the data will be the guiding light.
Some opinions expressed in this article may be those of a guest author and not necessarily Marketing Land. Staff authors are listed here.