According to Stone Temple’s recent study on mobile voice usage, people are increasingly comfortable using voice search in public—despite 45% of the surveyed audience agreeing that they feel annoyed when they hear someone else talking to Siri! So, it looks like as a species, we aren’t quite ready to concede that voice search is becoming a way of life for everyone, and not just fulfilling our personal needs to set alarms or direct us to the nearest caffeine source.
But a tipping point is coming that all companies must prepare for. In late 2016, Gartner predicted that 30% of web browsing sessions would be performed without a screen by 2020, and comScore predicted that 50% of all searches will be performed by voice search by that same year. This means that in less than two years, people will be using voice search to such a high degree that they fully expect to receive one, and only one, great answer from Siri, Alexa, Cortana, and friends, regardless of the complexity of their questions.
Are companies ready to meet this opportunity and lay claim to being the one and only return when people ask questions related to the product or service they provide? Rob Pierre, CEO of Jellyfish, and Rebecca Stone, VP of International Marketing and Demand Generation of LiveRamp, weigh in on what’s important to work on now, and the ethical responsibilities Google, Apple, Amazon, Microsoft, and other makers of digital voice assistants bear with as they build increasingly powerful capabilities in voice search.
RampUp: I’d like to introduce who we have here on the RampUp Podcast. We’ve got Rob Pierre, who is the CEO of Jellyfish; and Rebecca Stone, Vice President of International Marketing and Demand Generation at LiveRamp. I’m excited to have both of you here to talk about the market for voice search, and the importance of this channel for all sorts of brands across verticals.
Rob, since this idea came from your end, I’d like to hear your perspective first.
Rob: Thank you. I think what’s interesting about voice search at the moment is that historically, whenever you looked at search, it provided several results to your query. Essentially search engines asked, “Which do you prefer? Which of these results do you think is most relevant?”
Where it’s changed with voice search is we’re normally looking for one specific answer. This requires more signals than historically has been available with your traditional search. It will use your location, your history, your devices, and your behaviors online. There’s multiple factors that it would use to then come to the answer. So instead of a page with listings, which you would have historically had on desktop or mobile, it’s giving you the answer and quite often returning the answer in voice.
I think that’s the bit that’s really interesting. It’s going to be how these devices are going to provide the results, how they’re going to narrow it down to one thing and how can marketers and the brands mark up their data and their information and services to make it available to the search algorithm so that these decisions can be made.
Rebecca: In terms of that, where do you think we are on that spectrum today of getting the right answers?
Rob: I would say we’re not quite there yet. I think that is down to the fact that the way in which we mark up our data hasn’t really catered for voice search. I think also that the granular nature in which you search with voice—the way we speak—is very different to the way we would type and search. It may not be that we’ve got the variety of keywords and the granular data to point people in the right direction.
I think the technology’s there. I think we’re more than capable of providing the right answers, but I think until it hits that tipping point where there’s enough voice search for every brand to dedicate the speculative resource and investment to make it work it well, we’re kind of in that catch-22 where it’s possible, but the volumes aren’t at the point. The critical mass isn’t there for the right level of investment.
Rebecca: To your point about using location data and a whole bunch of different points of information in order to be able to really figure out what a person is asking, do brands need to build out the search results using AI or machine learning? Or is it just a matter of more people needing to do it so that more companies are addressing it?
Rob: I think definitely a lot of AI is needed. Machine learning as much as AI powering this stuff now. The natural language processing that happens with voice search is incredible. The amount of data points that they’re taking in and allowing us to handle with cloud processing is incredible.
What’s also interesting is the way in which we’re going to have to determine what the intent of the searcher is and how we provide the right information back—it’s going to require a lot of processing support.
Rebecca: As a marketer, it’s scary to hear that I’m going to need to learn machine learning and AI. I have been thinking about both in terms of analytics, but not just for the simple fact of channel marketing. How does a marketer even start to think about getting themselves to the point where they’re knowledgeable enough to build a campaign around voice search, using AI and machine learning?
Rob: I don’t think it’s going to be down to us. If you think of the algorithms that are in play for the auction within AdWords and search as it stands today, most of what we’re talking about is being taken care of by the actually search platform providers. I think we can rely on Google and Bing, and I think all the AI and the modules are going to be produced for us. It’s a bit like Google’s natural language processing API—it’s all there. We can actually leverage it and use that today, like their image recognition.
The fact that you could have the Google Photo app on your phone and you can search by person or by location—you don’t even have to have those things tagged. AI and machine learning will actually look at the background and figure out where you are based on what it’s recognizing. It’s even got the video API, so you can upload a video and their API would tell you what’s in that video. It will even give you a timeline, telling you that three minutes and two seconds in, there’s a dog on the beach.
This is all incredible stuff that’s being made available by the big search platforms. So we just need to know how to make our data available and mark it up so it’s accessible and we will be running campaigns in a very similar way. The only difference is, we have to cater to things being more granular so that we can point people in the right direction once we’ve identified what that search is.
Rebecca: It’s almost like the next generation of SEO, so all the SEO firms are going to be the ones who are really the experts at, to your point, marking things up.
Rob: Exactly. The big thing is that when mobile came along, it was the first three results that were important, because they’re above the fold. Now it’s just one result, so that’s going to change the dynamic immensely.
What really excites me about voice search is when you’re talking to a multitude of devices and it provides you with the one answer you’re looking for. It’s sort of preempting what your needs are based on your circumstances, the time of day, your personal demographic, your behavior online, and your location. It’s just incredible the amount that they’re going to have to process in that millisecond to provide you with the answer that you’re looking for. You’re right; it is marking up data and making sure that you’ve got all the fundamentals that is going to be extremely important.
Rebecca: Taking it from the consumer side for a bit, I was interested in voice search because I have a now-five-year-old daughter who was first introduced to Siri probably two years ago by my brother. He was just joking around with her, showed her Siri, and started asking some questions like, “Hey Siri, can you rap?” All these questions provided more humorous answers but it got her to the point where she was comfortable and recognized that she can ask Mommy or Daddy’s phone a question to get an answer.
So, even before she can type or anything, she can go and ask questions without having to know letters or words even—she just knows how to speak. The one place where voice search is falling short is exactly to your point— it comes back with a whole page of results. My daughter knows how to ask a question, but because she can’t read, she can’t actually get the answer yet.
Rebecca: Once voice assistants start talking back to her, she’ll be able to do it. These kids are growing up learning how to use mobile devices and computers and things in a way that adults are not able to. It’s happening even faster that kids are naturally learning to speak to devices and are going to grow up with the expectation that they will be able to speak and get results back rather than have to type something into a phone.
Rob: I’ve got a similar story because I have a nine-year-old son. If he wants to spell something, he asks, “How do you spell environment?” and it then provides back the answer. There’s some practical applications where voice search is definitely adding value.
I think as a support for things like homework etc., it doesn’t feel like cheating because it’s like asking a parent. You’re asking somebody with knowledge who can provide you with an answer and then you still have to translate what you’ve heard, write it down, and apply it to your homework. I think it adds quite a nice experience.
Now, with devices like Alexa and Google Home, there is no screen. There is no choice but to provide back an answer in the format that a four-year-old will understand because it’s talking back to them. When that becomes ubiquitous in everyone’s home and you’re just constantly having that dialogue, brands will be forced to support it and to provide information in the right format that can be read and provided back as a voice result.
Rebecca: Do you have any examples of how companies are using voice search today and some of the ones on the forefront? There’s the classic story about how Domino’s now has something like 20 different ways to order a pizza and one of those is using Alexa to just say, “Alexa, order me a pizza.”
I’m just curious if you have heard of any companies who are playing around with it today and have had success or even a failure that they’ve learned from.
Rob: We’re playing around with, for example, recipes and cooking. I think there’s a lot of opportunity around things where you’re using your hands and they might be dirty. I’ve heard a lot around recipes and cooking portals and platforms where you can ask for a recipe and be in dialogue with a voice assistant whilst you’re cooking. Then, it can also help with timing, etc., so it will start telling you when to do things. This is a good example that I’ve seen with a practical application.
Other than that, I haven’t really. But it’s certainly an area where we at Jellyfish are trying to come up with some innovative ways to work with brands to start leveraging voice search. I think there’s a first-mover advantage and if the historic methods of SEO are anything to go by, getting there first and establishing yourself as the number one result can only be a benefit for a brand. It’s kind of like a race for most of the search terms to mark your information up so that you can be that first result.
Rebecca: What are your thoughts on the negatives of this? Now that there’s an always-on device that you’re not just activating when needed, might it pick up conversations and be able to understand what you’re saying without you explicitly asking it?
Rob: Yes, so I think it’s an interesting question. I think there’s no way that the devices and the search platforms, etc., are going to exploit that feature. I think it’s purely coincidental. I’ve seen on loads of occasions where on Facebook, for example, you’re looking at your feed and all of a sudden an ad just comes up. You think, coincidentally, I was talking about that only yesterday. It’s almost like Facebook’s listening to me, and that’s not connected in any shape or form.
I think this happens because of the way we’re able to stitch together data. We can look at people’s behavior and create lookalike audiences. You don’t even have to be directly behaving a certain way; you just have to be segmented into a customer profile that has similar behaviors and then you’ll be linked together. I just think it’s less noticeable when it comes to banner advertising or search results, etc. All of those are highly personalized and it’s much more relevant to you as an individual.
Particularly with so much emphasis on GDPR and personal, identifiable information, I think it will just be the death of voice search before it even starts if the technology was actually listening in and using that data to then provide results to you. I’m pretty sure it’s coincidental and it’s all about the fact that the data and the footprint and the breadcrumb trail that you leave online is helping brands make things more relevant to you. I think we should embrace it and enjoy the fact that you don’t have to sift through and listen to all of these results or ads or anything that is not applicable to you or not relevant. Our experience is just much better. That’s my view on that side.
RampUp What thoughts do you have for brands and also the search engines for determining who’s too young to be marketed to? Especially as kids get older, they may be asking Siri and Alexa more serious, personal questions. In this age of consumer skepticism and consumer trust with personal data at an all-time high, what do brands need to be cognizant of as a generation of people grow up knowing about voice search and using it regularly?
Rob: When you’re connecting to Google Home or Alexa, you’ve got an app. There’s a certain responsibility of the parents to ensure that they stop results or they don’t recognize searches that are inappropriate for that child’s age. I think that the algorithms are so sophisticated that it can actually work out who’s using the Alexa or Google Home on a regular basis and can figure out the age just by the voice search query and maybe the time of day, the habits, etc.
The device could actually just say that it’s not sure. It could just say, “Please can you provide me with the safe word.” There must be numerous tactics to avoid inappropriate content and advertising to minors. I don’t see that in the future being a constraint or hurdle for the proliferation of voice search.
Rebecca: I would just add I do completely agree that a lot of it should fall on the parents. The only thing that I would ask for is more granularity in terms of parental control settings from those companies, because I want to definitely encourage my children to use technology. Obviously they have to grow up in a world where they’re comfortable with it. I subscribe to the belief that they should be allowed to use it but number one, in moderation, and number two, with some controls.
It’s very difficult right now to control in a very specific way things that they’re allowed to access and not access. That would be one of the things where I think companies could go much, much further.
RampUp: Right. It sounds like there’s lots of opportunity but also lots of responsibilities that have yet to be ironed out, so definitely watch this space. Any other parting thoughts, either Rob or Rebecca?
Rebecca: Rob, I think it would be great to hear a bit of an explanation about what Jellyfish is and why you guys are interested in voice search.
Rob: Sure. Jellyfish are a full-service, digital agency. We are very interested in the full user journey, so right from awareness through to advocacy. We believe that everybody should take a customer-centric view where most consumers are device and platform agnostic and they expect brands to know where they are on their buying journey at any one point. We’ve been working on ensuring that we’ve got that technology layer which is the robust platform in which we can build upon.
We want to make sure that we can build every single asset and provide the right experience at the right moment. We make sure that we’ve got a full analytics and insight team that can look at the results, lay the tracks, make sure that we’re measuring all the key micro-moments that matter, and analyze them. We test, analyze, and refine so we’re ensuring that we can provide that full, digital solution for brands.
I think the days of one message to many, the traditional advertising where you put out one message and you deploy it in every media channel possible is going away. It needs to be much more personalized and much more relevant for each individual. We’re basically building a business to support brands to do that and obviously voice search is a part of that. You’ve got to be there if somebody’s using that particular channel and that method to seek information; you’ve got to be there providing the right results to cater for that moment.
RampUp: Great. Well thank you both so much for your time. I appreciate it.
Rebecca: Thanks, Audrey.
Rob: Thanks so much.