Voice Analytics: Nothing to Shout About (Yet)
By David Sprinkle
Last month, Adobe announced Adobe Analytics Cloud, which features new voice analytics capabilities that allow brands to deliver more tailored customer experiences and build brand loyalty through voice.
We asked our Chief Analytics Officer, David Sprinkle, what his thoughts were on the new advancement and whether it was a ‘eureka’ moment for voice in analytics, here’s what he had to say:
I’ve been reading a lot lately about the growth in voice-enabled devices like Amazon’s Echo, which seem to be at the cusp of strong growth. Recently, Adobe announced new functionality in the Adobe Analytics Cloud geared towards helping companies track voice-assisted interactions with apps, with an eye towards eventually helping brands personalize those interactions.
With our background in intent-based marketing, we at Acronym have long been intrigued by the possibilities behind voice-driven digital interactions. But, having dug a bit deeper into what Adobe’s announcement really entails, my excitement about this emerging space has cooled off a bit. While I see value in tracking and analyzing the kind of data this new integration will provide, I’m not convinced it’s as revolutionary a step as I initially hoped it would be.
To understand why, one needs to understand what the flow path looks for a voice-driven interaction and how this will show up in the data.
The first link in the chain, of course, is someone making a vocal request of their device, such as, “Hey Alexa, I need to book a hotel in Boston next Tuesday.” From that point, algorithms within the digital assistant parse the request and convert it into a structured request that can be fed into an app, for example, the Kayak app. Next, that app responds, and in this case, the app would respond with some hotel suggestions and rates. Finally, the voice device would then vocalize the response from the app to the user. Neat!
Considering what happens when a person uses voice activated search to book a hotel, a question arises regarding how a hotel brand can ensure their property is in the mix of responses. Do search engines only go to booking engines online for those queries, which means hotel brands must already do well within those engines?
This is a very interesting question and one that I can guarantee is going to be a thought-provoking subject for years to come, and maybe even lawsuits! However, I’m not sure anyone really knows yet. In the travel example, would Google be going to Kayak for that, or to their own Google Travel results? If you say you want to book a hotel room at the Four Seasons, would it go to FS.com or TripAdvisor (Or Orbitz, etc.)?
The basic question here, I think, is whether Google/Amazon/Microsoft/Apple are going to end up being the gatekeepers of this voice technology because of their already dominant positions, or if separate voice capabilities are going to be taken on by every app out there. Realistically, I’m not going to pull out my phone, then click on an app, then talk to it—I’m going to talk to Android or iOS directly and expect them to sort all of that out for me. I don’t think the landscape is clear on how this is going to play out yet, that might it’s own separate piece!
But, anyway, what’s important to keep in mind here is that most of what my colleagues in the search marketing space would call “the fun stuff”—divining user intent from an unstructured request—is going to be taken on by the device’s algorithms. And let’s face it: with Google, Apple, Microsoft, and Amazon already staking claims on this piece of the puzzle, most marketers aren’t going to see what happens in those black boxes.
We will have access to the structured data requests that these algorithms spit out to our apps, and these are what Adobe’s new solution plans to track, along with app responses and outcomes. This is unquestionably valuable information; for one thing, I have a feeling the interfacing between the device algorithms and apps might be error-prone, and we may need to be able to monitor that. Once Adobe progresses beyond simply capturing this information and moves into personalized responses, that’s where things could get really interesting.
But for now, it looks to me like we’re just going to be getting a few additional data points around app performance, and not a world-changing new source of customer intent data.