Every two weeks host Will Francis explores with each guest all aspects of their own digital marketing expertise, as well as those soft skills like presentation and productivity techniques, that we could all learn from.
People are searching with their voices more and more.So brands need to start listening. In this episode Will Francis chats with digital marketing expert Clark Boyd about the different approaches being taken to voice search around the world and the opportunities marketers have yet to grasp.
To enable screen reader support, press ⌘+Option+Z To learn about keyboard shortcuts, press ⌘slash
Episode 8: What is Voice Search? [TRANSCRIPT]
DMI AHEAD OF THE GAME PODCAST
What is Voice Search? With Clark Boyd
Will: Welcome to "Ahead of the Game," a podcast brought to you by the Digital Marketing Institute. This episode is a big Q&A where we explore an area of marketing through a leading industry expert. I'm your host, Will Francis and today I'll be talking to Clark Boyd, a digital marketing expert and author based in London, all about voice search. Clark has worked around the world, helping brands like Adidas and American Express shape their digital strategy. He trains digital marketers, through a variety of institutions and programs, including Google, Cambridge University, Imperial College, London, and Colombia University. Clark, welcome to the podcast.
Clark: Thanks, Will. Good to be here.
Will: Where are you joining us from today?
Clark: I am in the same place I've been for the last five weeks, which is to say my spare room in my flat. I'm getting very comfortable in here.
Will: Wow. Is that in London?
Clark: In London. Yeah.
Will: So, first of all, can you just define what voice search is and how people are using it today?
Clark: Yes. So voice search is a search where the input is spoken by the user to a search engine or a digital assistant like Siri or Alexa. So it’s different to the traditional search where you would type a word in. We would be speaking. Now, the search engine then interprets the query and provides a response from its index of webpages. So pretty similar to traditional search there. The response could be spoken by the digital assistant, so we might speak to it and it speaks back, but we might speak to the assistant and it shows us results on a screen like a traditional search results page. Normally, the responses are given in a snippet of information, quite a short piece of information, just because people are on the go. They don't want to listen to a long answer coming through from a search engine.
Now, another point I would point out on this actually, is that a lot of people conflate voice actions with voice search. And I think it is more than just a pedantic point actually to demarcate the two. So when we ask Siri or the Google Assistant to play some music, or set a timer, or add something to our calendar, that's really just a voice action. So we're providing a spoken command and it carries out the task. A lot of studies that I've seen would count this as voice search, but for marketers, I don't think it's all that significant. There's not much that we can do there as brands to intervene or to shape any behavior. A voice search is when the user is asking for some information. So it's the same distinction as say, between setting a reminder in your calendar yourself and asking Google for the best restaurants nearby. You know, one is voice actions, the other is voice search.
Now, how people are using it. This is a tricky question because...it shouldn't be, but it actually is because we don't see the data split out, for example, in search console. There've been lots of rumors that this is coming and in beta, and we will see a filter so we can split queries by text or by voice. But as it stands, we're relying on a few different studies and that sort of thing, which are, you know, sometimes reliable, sometimes not. But the core benefit for users of voice search is convenience. It's when they're driving, when they're cooking, when they just need a brief update on the weather or news.
And there's a novelty factor, you know, which I think is fading. I think people have asked Google for enough jokes and things by this stage that they're getting a little bored of it, but to bring this through to what it means for marketers. So if people are searching by voice and we'll discuss in more detail, I'm sure, whether that is very much the case. They tend to speak in natural language. So it's not the blunt instrument of keywords when people speak to a search engine. And that, I think, is the really key point for marketers to understand. If they were saying the exact same thing that they were typing before, I wouldn't find that interesting. How would that alter my strategy? Not really. The customer wants the same thing at the other end. If they're speaking and they're revealing something more, then I can do more through my content. I can connect with them in a different way.
So, to use an example, if I were to search, say on Google for...look at what I'm wearing today, “gray, oxford shirt”. I might just type in “men's shirts”, go to a website and I would do the filtering myself. If I were speaking to Google, it would be very strange to just say "men's shirts" and then wait for the results to come up. I might say, "gray shirt, nothing fancy," to make clear that I'm not a fancy man. I don't want anything that crazy. That might seem like a persnickety point, but it's actually really revealing because the assistant can do more with that information. If you were a brand trying to connect with customers and you sold fancy shirts, well, you wouldn't want to bet on that keyword because you would know it's not your audience. There's more being revealed there. So the assistant can take me through to something more appropriate, but we've got the input. I'm saying “gray shirt, nothing fancy”. We've got a search engine in the middle. We need brands on the other side to create the right content there. How are they matching to those intent states? How are they creating content that can be served up for that result?
Will: But is it not the search platforms that are mapping that customer need to an existing content on the web? Because that would make the most sense for someone like Google, right?
Clark: Well, that's kind of the problem, is that they're matching it to existing content and the existing content is all written and belong to keywords. So they would take me to a men's shirts page when actually there's a better user experience there, which would take you through to, you know, what you're actually looking for. So if I were searching for a keyword that's really expensive for paid search like “business insurance”, that can mean so many things. Brands will pay north of £20 per click to get people through on something like “business insurance”. What do I mean? Am I looking to find my provider to make a claim? Do I need more information? Am I trying to compare? Do I want to take out a small business policy? We don't reveal those things through keywords. If we start revealing more through spoken keywords, it gives people an opportunity to create better content that maps to those intent states. It's not about the keyword matching, it's about mapping to the different intents that people might have, different styles, different needs, whatever they may be.
Will: That's interesting. So, yeah. You're saying that voice search queries, because they're in natural language, they just carry much richer information that allows us as brands to meet a more specific need more directly, more quickly, and really offer the convenience and directness that voice search potentially offers.
Clark: Yeah. I think if people are speaking – it sounds a bit trite and glib - but if they're speaking more, brands should be listening more. You know, they should be trying to capture that, turn it into something more useful.
Will: In your webinar, (you did a webinar about voice search for the Digital Marketing Institute which is very interesting, in the membership area), you talk about the differing adoption of that behavior around the world, which I thought was really interesting in that you said that we're slowly adapting, whereas some markets are just leapfrogging altogether. You know, they haven't been stuck on kind of, you know, old behaviors so much, because you just don't add much technology, it's all knocking around. They're moving straight to natural language-based interfaces. Tell me a bit more about that.
Clark: I'm very glad you asked. It's a passion topic of mine. I think that in Western markets, in particular, we have, because of technological constraints, we have learned different behaviors. So we're used to having internet-connected devices all around us, we're used to searching and we've learned to search in ways like searching for “business insurance”, searching for “men's shirts” because we're, you know, we're result-oriented. We're trying to get to the product at the end of the search query, but it's more natural for us to speak. You know, we're even seeing this in how, well, things like podcasts are taking off at the moment. People are listening more. They're speaking more to computers. There are voice blog platforms coming out now, so it's not about typing your blog posts. But this is all because computers have understood text much better. Now that they're starting to understand language, those who are just getting online, or getting their first device or, you know, getting vital services for the first time, they're able to go straight online through voice.
Now, there are...I find it staggering…but the UNESCO figures said there are 700 million adults worldwide who remain illiterate, which, you know, it's obviously pretty shameful, but it's a fact. And they're now getting devices from the likes of Google and Facebook with a big button in the middle where they can just press it and speak to the device and it will take them straight to what they need to do. So, for me, for example, or perhaps for you, Will, it's sometimes a little awkward to speak to a device. It just feels weird. It's not what we're used to. Speaking is a little more intimate than typing. For a lot of other people, that's not so much the case. And it's a cultural thing as well. You know, it's not just about those facts. My partner is Argentinian, so I hear the voice memos going all the time and the voice chats through WhatsApp all the time, which to me would feel very, very strange in spite of doing a lot of these things, I hate the sound of my own voice.
But there's a lot of people who are, you know, sending voice memos. And when I've spoken to them to say, why is this more useful for you? “Yes, they say, "It's faster, I don't need to use my hands," but they also say, and I think this is key, "It removes the ambiguity." When you text someone, you often use emojis, you know, to say... You know, that wasn't passive-aggressive by the way, I'm being genuine here. But when you speak, you give away a lot more. There's a huge amount in your tone, in your intonation, emotions, all of those things. So I think on a global scale and, you know, Google has said they expect India to be the first voice-driven market, obviously a very significant and large market. If we look ahead 5, 10 years, we have a lot of people growing up for whom it's no longer strange to speak to their devices and if they find it more useful, if they find it faster and if the results are better, why wouldn't they keep doing it?
Will: Absolutely. Yeah. And we've seen that happen with so many technologies previously in our lifetime.
Okay. So obviously our audience listening (to the podcast) are prominently marketers. What are the KPIs? What are the metrics that a marketer today can drive through voice search both in the short-term, so what are the kind of quick wins, but also what foundations should we be putting down for that kind of longer-term strategy?
Clark: It’s always interesting with voice search, because it has all of this hype around it, and, you know, there's real buzz status.
Will: It was everything that people were talking about. Every conference you went to in like 2016, 2017, there was just stats about voice search that always started… “in 2020”, or “by 2022”, you know, and yeah, there was a lot of hype and I think maybe some marketers have taken their eye off the ball because well, that didn't immediately deliver some results. But there clearly, from what you're saying, there clearly is going to be a real gain to taking notice of this. So yeah, just interested to know what we can be doing as marketers now.
Clark: Yeah. Because that's what I've been thinking about since going to, well, maybe the same conferences and hearing the same lines and the same, well, lies in a lot of instances, you know. It's all self-serving. So many of those presentations will end with, "Oh, and here's the name of our marketing agency. If you don't want to be left behind, if you don't want to go completely bust, you'll get in touch with us and we'll fix it for you." But I kept scraping away at that and trying to figure out, is this just a complete nonsense or is there something behind it? And when I started looking at the companies that are spending billions of dollars on this, and if you look at where we are today, which as I mentioned earlier, we're looking at setting timers, asking short questions. People like Google and Amazon aren't bad business people. They don't want to spend billions of dollars just to set our timers. They don't care about whether we want to hear a song. That's not interesting. They want us to speak and they want to be able to give responses. And when I've been looking at it as, you know, what can you do today?
So you're thinking of voice search optimization. Well, it's tricky when you don't know which queries exactly are voice. So what you end up doing, and I don't think this is a bad thing, actually, is you're focusing more on mobile. You're focusing on more natural language content and you're thinking of that usage behavior. Whether it's expressed through text or through speech, isn't really all that important. The thing is that you, as a brand on the other side are connecting with people, you're providing the information that they need. So there are two key areas that I've been looking at, but I can't say I would put it down as voice search strategy and, you know, you can draw a line in the sand and if you do these things, you'll see 20% more voice search traffic in three months. Pretty much impossible to do that, but I do think you can look at two areas.
The first is how you structure your content and the second is the style of the content, how you say it. So when you're looking at structuring content, and I've seen, you know, whether people are coming through voice or not huge increases in traffic for clients and for myself, just by structuring content into these succinct fragments that respond to specific query categories. So we used to think of...we'd look at a landing page and think, "What's the purpose of this page? Why does it exist? Why is it on the website?"
Will: What queries in key words does that answer, that landing page?
Clark: Yeah. Exactly. And now we can look at the whole page and break it down further and say, "Okay, this section of content, what intent state are we responding to? How are we helping someone get something done here?" And, you know, I speak to some clients in that way and they think, "Oh, God, here we go. This is gonna be, you know, the whole page isn't going to read as one piece of content." But, of course, it can still flow. Now, the average piece of content from which Google pulls a snippet has over 2,000 words. Now, Google isn't pulling 2,000 words into the rich snippet, it's just pulling that succinct fragment that answers that query exactly. You know, you can't know how it will be phrased. As I mentioned, I might say, "Men's shirt, nothing fancy." I'd say, "Gray shirt with no stripes." That's potentially the same thing, actually. Now, linguistically, if you bring it down to its nub, you're saying something very similar there.
So I try and map that out within the pages and think, "What could there...?" So even under things like H1, H2, H3, what are we putting there that could be a response to something that would be helpful to a user? Putting yourself in the best position to be served for those responses. And you can research common questions, you know, you can use those within your content. There are numerous tools that will allow you to see the search volume for different questions. And you don't have to just create an FAQ page. I've seen that as essentially voice search strategy. They'll add an FAQ page and just answer everything there. But that's not the experience that Google is looking for.
In fact, if that question and answer is contained in a much longer piece of content, that shows more authority, that shows more trust, you know, that suggests that you are serious about answering it, and not just going for the rich snippet. So think about those points of friction, look at the customer journey. Along the way, what might be the things that they ask? Don't focus so much on whether they might speak or might type it. If they speak sure, sure, they might give more information and that's great, but it's still a customer looking to achieve something, looking to find something.
Will: Yeah. It's still on a written search results page on a screen. It’s still useful because that's how Google tries to serve the best answer to your question. And how important is it that that content is in some way tagged up, or is the main thing just that it's useful and it answers the question?
Clark: I think anywhere pretty much that you can be using structured data to mark up different aspects of your content. It's usually a good idea to do it. So those will help a search engine to navigate. If you can just conceptualize what is happening with the search engine, someone says something and it has to go and look for the exact piece of information within all the pages, within all the apps, everything that's out there and answer it very quickly. They could hedge their bets before and say, "Look, we don't know, here's 10 links, take your pick." But now it's having to say, "Not just that, here's the paragraph that answers that more clearly." But I think what tends to happen there is that people try and game that system. So every time something comes out like FAQ, schema, Q&A, people will go in and mark everything up. Google has already said, "We're not looking for that. These should be specific." Help pages: you see it working. You know, car insurance companies and the like have been doing very well out of it.
But I wouldn't...I say this because I've fallen foul of this kind of thing. I've found a little trick that will get me more traffic, and then I start projecting the next year, assuming that will stay the same. Google will close those loopholes. So I would think more about, you know, if it's images, if it's another company name using schema.org markup is really helpful there. When you're looking at things like Q&A pages and Q&A sections, think more about whether you're just trying to trick Google into giving you that snippet or whether it's valid. I haven't been using that format for my snippets that I have on lots of different queries. It's just been about answering a question very succinctly and directly, and hopefully with more authority than somebody else does.
Will: Well, that answers the question. I mean, because I suppose just for anyone listening who's not sure what we're talking about in terms of markup. Yes, there are ways you can find out about schema.org, that you can basically tag and demarcate your content as being certain types of content to make it easier for search engines to know whether it's an FAQ page, or a recipe, or what have you, right? But yeah, I agree. I think what's clearly happening is Google's smarter than that. You know, we can't underestimate how clever Google is. Google, it's been their business for decades now to find the best answer to your question and it will do that, whether that's marked up or what, you know, it will find the paragraph that meets your need, your informational need, won't it? But it still hedges bets with a few other questions below that, you know, and did you kind of, you know, did you mean this? But clearly, for marketers today, the goal is to be as useful as possible in terms of meeting the informational needs of the audience.
But it's really interesting what you say about that granularity, going beyond landing pages and actually thinking about each paragraph and saying, "What question does this answer?" But then obviously, you know, because like you say, Google seems to favor longer content and it's less about... You know, clients sometimes talk to me about, "We need some SEO content," which I hate that phrase because it's like, well you just want to pump out thousand-word articles that you think are keyword-rich. Google doesn't really want articles. Google wants resources, Google wants guides. Doesn't it? It wants several thousand words of authorative content on the topic that it can pull a paragraph out of, but as a whole, is the best resource for that informational need on the internet. And so, voice search, just kind of cherry-picks from that, clearly. It's really interesting, hearing you talk about that. But I suppose what a lot of brands...when we talk to brands about that, what they struggle with is how do I go beyond serving up cold informational searches and get into, you know, stuff that's more brand-led? How does a big automotive brand or, you know, a big fizzy drinks brand play into that because they can't answer these, you know, “how far is the moon from the sun”? And “what's the weather going to be like tomorrow” type questions? Do you know what I mean?
Clark: Yeah. Yeah. Exactly. And that's what I've been thinking a lot about because at first, when clients were asking me, and this is back in, you know, agency days, clients will say, “Oh, voice search is going to be big. I've seen this stat about how everyone's going to be, you know, searching by voice”. And they would be, say, an automotive brand or something like that. And I think, well, if someone searches for something that's relevant to you and you are the answer that is spoken by Google, you're not getting the click. You're not getting really the traffic through. You're not even seeing who these people are. You might see that you've been served up as a snippet, but that's about the best you're going to get. And at the end of the day, you want to sell cars. And we as marketers are in the industry of shaping behaviors, of understanding people and of persuading people. Are we driving towards that goal? No pun intended. By answering these really flippant little informational queries and often the answer to that is no, not really. So I avoided doing that altogether.
I think it's much more important in those instances to think...and this is...I mentioned, you know, two ways I'm looking at it, which is the structure of the information, but then the style of it. So if you're going to be read out, how are you going to make some sort of impact on the end-user? Are you going to sound like a Wikipedia page? You know, you've got the keywords in there, so you get spoken by Google. Fine. Will people even remember that that was you? Is it really worth putting all your effort into this? Just for really, you know, vanity metrics. And people win awards for this kind of thing, voice search innovation.
And I hate when people reduce that down and just say, "Yeah. But what did that mean for revenues?" But what did it even mean for the next stage? You know, did that person come back? Did they download a brochure at some stage? And that's where we're seeing some brands experiment, and it's not implemented yet, but we're seeing them start to experiment with more kind of dynamic ways of responding to those queries. So, for example, if someone were to search, "Should I get Geico Car Insurance?" That's a brilliant opportunity if you're Geico to say something. Now, at the moment, and for those...not initially a Geico, like the biggest ad spender in the world, but, you know, massive insurer in the U.S. and that would be a great opportunity for them. If you search that now by voice, you get, you know, just an FAQ page, something really, really dry.
Now we're expecting those of us in the industry of keeping up the speed for these things, which is a sad hobby, but one that I have. We're expecting...this sounds like we're a little club or something, but we're expecting to see this year that the Google assistant will become much more deeply integrated in the Google Chrome browser. So at the moment, it's very difficult with those queries because you wouldn't just say, you know, "Buy me a BMW" and Google goes and does it, you know? Or maybe it would, I don't know. But... Yeah. Those of us in the industry of keeping up to speed with voice search aren't making that kind of dough. But you might say something like, you know, ask for specs or something like that. And that could be fine. That's okay. But what if you were able to integrate your brand voice into the response? What if you had a skill or an action that was about essentially like a friendly car salesman guy? Now, I don't know if they exist, but you could create one and have a bit of a brand voice.
I mentioned Geico and I did purposefully, because for those of you that might've seen it, they are a huge ad spender and they have a little gecko guy who is ubiquitous. Now, he's on baseball fields, he's all over the website, except when you go to the specific policy pages, and then it is the most... But I can pick on Geico. They're big enough to be picked on in these instances, but everybody's the same. They are no worse or better than anyone else. It's all SEO content. Just what you mentioned, Will, 1,000 words of SEO content. And this is a huge, huge, huge company. So they're not even providing the opportunity for something, a little more stylistic that has their brand voice. And I think as a brand, you know, if you're thinking, "Well, okay, these things might be happening, but maybe they're not, what can I do today?" Well, at least consider what your content sounds like when it's read out. Would you be proud to have that read out by Google? Would you want people to hear it or is it just SEO content that you put on the page?
Will: It's a good point, isn't it? Because with the way it's structured on your website, you might think, "Well, no one visits these pages and if they do visit these pages, they understand it's like the corporate kind of backend of the website that you might accidentally wander into." But search engines don't see websites like that. They just look across the whole website and go, "Where's the answer to this question." And that might be on your, yeah, your policy page or some really dry page that doesn't sound like your brand at all. And they've encountered all this marketing and yeah, Mascots at baseball games and funny TV ads. And then they get in this kind of, you know, lawyer read into the... So, yeah, it's a really, really good point though to makes sure that even your, you know, your terms and conditions, like every bit of content on your website reads like your brand because, yeah, whilst at the moment we can't actually have our own voice, literally, that will come in future. It's the words and the way things are phrased that convey who we are, you know.
Clark: If you look at, you know, a retailer and they sell...they're selling trainers, look at actually somebody like an Adidas. If you go through a traditional journey as a user, which is you would go to the homepage, you see all of these great graphics and all the great stuff about the brand and then you move through to categories and you look at different products, you have seen a lot of other information. So if the landing page you arrive at, you go onto the trainers page, if it has a couple of 100 words at the bottom stuffed with keywords, it doesn't ruin your experience. You can ignore that. We're kind of trained to know that no matter the brand, whether, you know, it could be a high luxury fashion brand, and they will have that piece of content. And why is it there? Well, because it has worked, you know, it has actually had a real tangible benefit. You know, I've written some of those descriptions. I've been party to this. I can't wipe my hands of it altogether. That's how search engines have worked. They've needed something to cling to. Can't read the images and it's quite difficult for them.
But now we need to think what if that piece of content was the only thing that the user saw on the whole website? What if they searched for trainers and Google searched up that little bit of content? Is it even worth ranking for that? If all you're going to say is “trainers online”, “trainers UK”, “get your trainers from Adidas today.” I'm not sure it's really, really that worth it anymore. And it's not as effective as it was. It's as we've been discussing, exactly what Google is trying to get beyond. That's the point of all these expensive natural language processing algorithms, is to, you know, understand the meaning behind language and not just read those keywords. So it's absolutely, you know, terms and conditions, FAQs, but the product pages as well are real focal points. And some people are starting to notice this now, just how absurd it is. If you remove SEO content from its context and look at it on its own, as I've done, and many a spreadsheet. But yeah, think about how that actually sounds. And you can play these things out. You know, there are plenty of tools online. Put your content in and it'll read it to you and think, "What if someone did search for something high purchase intent and all they heard was keywords back from our brand?"
Will: So to sort of summarize, anyone listening who really wants to make a first step into optimizing for voice search really needs to think about: does all the written content on their site accurately convey who they are? Is it useful? Does it answer the questions that people are likely to have about that product sector, about that industry, about the topic that they operate in? They’re kind of quite short-term tactical things. Is there a bigger picture? Is there a longer strategy? I mean, to lay the kind of, you know, paving stones for the next five years? I mean, what should I be doing now based on where voice searches are going?
Clark: I think the most practical way that we can think about the next sort of 5, 10 years, and it's never really altogether practical because, of course, you know, gazing into the future, being very speculative. But if we follow the trend lines and look at where they're probably going to go, I would be really thinking about that. You know, people talk about a brand voice, when people are planning that and they're sitting having their annual meetings about... We've all seen these presentations, this is what our brand says. This is what we don't say. This is what we look like. You know, this is...if we were an aftershave, we'd be this and know all those kinds of presentations. That should not be including SEO, we should be thinking, "What does our voice sound like on our website if we had our own voice?" And that's where I think this could be headed, where a brand not only could have Google read their content back, but that we could have audio clips within our pages that would be played through Google. That it would just say, here's an answer from whoever, Geico, Adidas, whoever. We’d be able to play that strip through.
So if it's able to understand audio much better, and we can see that Google is indexing podcasts. So the technology is there and it's even helping people navigate to the point within a podcast that answers their query, it's a huge shift in how that works. Now, brands aren't really working with that yet, but if you did have audio within your landing pages and you had your own brand voice and you were answering some of these queries, why wouldn't Google serve you up? Wouldn't that be a better experience for the user to hear directly from the most authoritative brand in that space? So I think that could be a big part of it.
Now, if I were to look very broadly, and this is where I probably have most of my, my interest in these things, about what these companies are up to, you know, why they're spending so much money. And it's everyone. You know, this isn't just Google going after this. It's Amazon, Apple have just spent over 100 million on a company called Voysis to improve Siri. Facebook are in voice. Obviously, we've got the likes of Alibaba, Xiaomi. Huawei is developing its own assistant now that it's, you know, kind of out of the US-based Android ecosystem for the moment. Why are they doing this? It isn't even just to answer those queries, we're thinking about it, or at least I'm thinking about it. I'll spare the innocent. It's my feeling, thinking about this as looking in Google search console and I can filter by voice, by text. And I would filter by voice and I would see all of these queries written down and that would be incredibly useful.
But if we're looking further into the future and, you know, I'm getting a bit more sci-fi about it. Every time that we speak, we're giving something more away. So Google has obviously just bought Fitbit for, you know, a lot of money. It's very interested in the health space. And I saw just yesterday, there's been a study from like the American Heart Association, that finds that they can predict heart defects through people's voice, kind of their voice fingerprint, the way that you speak reveals. The same nerves that work with your voice work with your heart. So there are anomalies that come through in the way that you speak that can pinpoint that in 5, 10 years, you're going to have issues with different heart conditions.
That's the kind of thing that can be very profitable. If you are the people that are being spoken to, you're capturing and pinpointing those issues and you have all of their health data. So you're wearing a Fitbit. You're in the Google ecosystem. You're surrounded by voice-enabled devices that are constantly listening and it can start to pinpoint those things, that can be a huge, huge business. We all know that the healthcare industry has struggled a bit to kind of transform digitally in the way that others have for a lot of very good reasons, very complex, very difficult to do. But now the likes of Google and Amazon are getting into that. And I saw in an Amazon patent where...anyone that's seen any of my previous stuff. Yeah? Both people, well, surely have seen me use this, because I love this one where Amazon has had a patent approved where Alexa would be able to listen to you and understand, for example, if you cough...different significance at the moment. But if you were to say “I'm hungry” and cough and sniffle: now the search query there is “I'm hungry”. You're saying to it. "Yeah. Show me some recipes." But Alexa responds by saying, "Would you like a recipe for chicken soup?" So it's gone beyond what is said, listened to the other cues that are around it, and thought, "Okay. This person's sick." Then the first says, "No, I prefer something else." And Amazon is then able to order them cough drops.
So, as marketers, what are the queries here? What's really happening? If I just saw, "I'm hungry" coming through at my search console and I'm still thinking in my 2010 way, you're not really seeing that much. But what if Google and Amazon and the likes were able to group together those kinds of revealing factors? What if they could sell those as intent signals as in-market segments? So at the moment, they're looking at things like, you know, I looked on What Car websites, so maybe I want to buy a car sometime soon. You know, that's really how a lot of it works. But if they were able to pick up on how I ask certain questions that I'm interested in certain things or even the way that I ask about business insurance. (laughs) I was going to say maybe I sound excited. I can't imagine anyone ever does, searching for business insuranc. But if they're able to pick up on predictive cues like that and group them, then you could bet on those intent states. Rather than betting on keywords, you bet on people who are properly in market for the product.
Will: I would be amazed if that's not the case within the 2020s at some point. I mean, you know, it's like the inferred demographic data that you can target in Facebook, for instance. So, you know, some of it's inferred and so it's not maybe data that people have proactively gone and entered into the system. And likewise, there'll be a dropdown for, you know, certain, yes, sicknesses that are inferred or certain, like, say intent states or and life stage and stuff like that. I mean, God, you think about how much, you know, Amazon really wants to know when you've had a baby, for instance, which, you know, it's happened to me in the past. And then they're all over you for like telling you what you need to buy all the different stages and what have you. But this takes it a whole step further, the workout, what kind of baby you have and what they're into, and what they react to, and what times in the night they're up and all sorts of stuff, right? That's the goal for them, isn't it? To have this... We were already been fingerprinted by our text searches and things like that and, you know, remarked in data. But this is like a whole other dimension of granularity detail that we can be targeted with. So I suppose... I don't know what's the action there for a marketer, it's to be...I mean, it's to be there it's to be present and be available online, it's to definitely be in the right places and have your products there, available to buy, well indexed. What else could we do as a market to prepare for that future?
Clark: Yeah. I think when we're looking at a lot of these kind of trends, it's tricky because there are different things coming out pretty much every day. You know, and the way that companies move is a bit of a zigzag, you know, they go in one direction and then they go back. And if as a brand you're following those updates, it's hard to know if you're coming or going. And if you're a marketer working in an agency, in particular, it's hard. You know, you provide this advice to the client and then a week later... You know, like the meta descriptions, they made them longer. And we thought that was permanent. Changed them back. And you've got a lot of people then go, "Actually, can we just call that back?" And it takes ages. You know, we're working quite slowly.
So the best way to approach something like voice is to have one eye on the customer, the journey that they're going through, and how they're expressing that. So what are the ways that they're able to get that message across and then have another eye on, what is the technology actually capable of doing at this moment in time? And you can start to phase that because you can get too far ahead of these things. I've certainly trialed things that search engines just haven't been able to do, things like actually embedding audio clips and trying to use structured data to get those played through searches. And you know, they're not expensive, they're just my little trials, but if they're not ready for that sort of thing. So there's no point doing it, you know. For example, you could remove all of your, you know, SEO content today and replace it with, you know, interpretive dance because it really conveys your message and gets across to the intent state and it might be the most beautiful thing in the world, but if a search engine can't understand it you're not going to get there.
So the closer you are to the customer, what they're doing today, what they're interested in, and, you know, those things don't change as quickly as the expression of them. So if you're thinking of, you know, if you are an automotive brand, well, what are some of those trigger points that you know exist in people's lives? How are you creating content that caters to those? How are you there? How are you findable? And then you're always thinking about the most important thing. So just keeping close to what the customer's trying to do and trying to filter that through a search engine. How can I use a search engine to connect what we do, what we have to say with that person at the other side? That tends to put you in a much better position than focusing on it, looking at the customer through the lens of a search engine. Because the search engine is always changing, I think you're better starting with what the customer's doing and then...you always have to look at that bit. Like I said, you won't get any traffic if you're not thinking about how the search engines function.
Will: I think this and all the developments that are happening at the moment and changing consumer behaviors are pushing all marketers to focus on creating the best quality content that serves the customer need. That's not for the brand, it's for the customer. You know, it's for people out there, it's for the audience, primarily. And there's still a lot of marketers are yet to make that switch, you know, that rewiring in their heads of realizing they've got to create the best content for people out there, not just branded content or brand content. So you're talking about the various assistants and Amazon users Bing. So we're all very, obviously, Google-focused and have been for years and kind of forgotten that any of the search engine exists. Do we need to start getting our heads around Bing and pay more attention there?
Clark: Yes. So if people start using Amazon a lot more for their searches, then they're going to be getting a lot more Bing results coming through. So I think a lot of marketers have cottoned on to this. It's more popular in the U.S. than it is elsewhere, but, you know, Bing has started to release some interesting features through its knowledge graph and with the comparisons and taking slightly different approaches to Google. So it's definitely worth looking into your traffic to see if that's significant. For some companies, it can be, you know, 5%, 10%, which is definitely worth looking at, you know. It can be a sizable number of visits for some companies. So it's often... I think you're right to mention, it's often missed that Amazon will default to Bing to get its answers. So I would definitely recommend at least having a look at that data to see if it's worthwhile.
Of course, it's not, you know, chalk and cheese, you're still trying to optimize content and a lot of the same structure data formats will apply equally across them. The main thing would be looking at the search results page features that show up in Bing versus in Google. So, of course, you'll have classic links. Of course, you might have images, you might have videos, but videos will be pulled from different places. There won't be the same YouTube dominance that we're seeing. Youtube has been the biggest winner in Google search results this year, which is a surprise to absolutely no one and how it answers those rich snippets. Now, do you have to structure content a little bit differently to get in with Bing rather than Google? They're small differences, but I think they can be quite significant.
Will: And you talked in your webinar about...I think you touched on how marketing, in general, was becoming more conversational where the consumer, you used a Spotify example. That's what I took from it anyway, that marketing is...I mean, of course, with the advent of social media, we were pushed to become more conversational with people, and voice and audio had just kind of taken that to the next level. So talk to me a bit about that.
Clark: Yeah. I like the example of platforms like Spotify and I think some more interesting ad formats from Pandora in the U.S. is a very similar platform to Spotify, (not the charm bracelet jewelry people, the audio Pandora, it's tough to make that distinction). Now, what I like about those is that they're not taking conversational marketing and assuming that you want to be best pals with BMW. I barely want to hang out with my best friends, nevermind with a brand in my living room, you know. It's specific to me, but yeah, I wouldn't be looking at it that way and that we can, you know, be almost chummy with people. The Spotify example and Pandora as well have ad formats that will play and the listener can say, "Okay. Send me more information." Or, "Oh, I'd like to know more about that." or, "Skip." It puts them in control. It gives them the opportunity to get, you know, more information on their own terms. And that seems much more in line with today's customer. You know, the big shift, of course, has been having these smartphones where we can always access what we need on the go. It's not up to brands to decide what we should be listening to or what we should see. It has to fit in alongside that. So in audio, there are really good ways to do that. At the moment, it's, you know, quite succinct. It's not about asking you, you know, how's your day going and, you know, would you like a test drive? It's literally just playing the ad and you can skip it, which means the brand doesn't pay for it, which is good for them. And you can have more information sent of say, "Send more to my email address or send more to," wherever. "Text me more details."
Now, the other thing that people then look out with conversational kind of marketing is, of course, chatbots. So that seems like a very natural place. The clue's in the name. You can have a lot more kind of conversational interactions with a company there. And that seems more likely to be merged into things like WhatsApp and Messenger than within the search results themselves. Because the big challenge in this, and I've tried to think of ways around it, and I'm sure somebody out there has come up with a better solution than I have where if we...you know, coming back to this idea of trying to be, you know, practical, what does this mean? If you're being more conversational today, if there's a chance to interact with people, what does that mean? Because we all know that Google is crawling and indexing our content. It has to be in a fixed place, it has to, you know, exist somewhere on the web. So you can't just write every single possible conversation into your webpages and then hope Google serves it. So a lot of it exists in databases that are pulled through chatbots and that's the best way to do that at the moment. I would imagine that search itself will become more conversational. I mentioned the Google assistant really coming to the fore in Google Chrome and really trying to post that as more of a way of interacting.
So that might lead to a more conversational search journey. At the moment, just because of how fixed content has to be, it's tricky, unless you were to write transcripts and have the worst experience ever. For most people, the search engine can't really do that. So audio ads, definitely, you can speak to them, you can ask more information, you can skip them. Say you're not interested. Chatbots are getting much better often used for customer service, but could be an assistant, you know, they could help you go through a website. Say I'm looking for my gray shirt, I can't find anything that's plain enough for my taste. Say to the chatbot and that can help you. For search, I think the Google assistant is going to be really key there.
Will: Is it a realistic action for marketers at the moment to create Google actions? Have you seen any brands do that well?
Will: And I'll just clarify what that is. A Google action or Alexa skill, it's like an app or applet, a mini-app created by third parties on Google's or Amazon's platform and certain queries trigger these and then you enter into them and then you're in the applets and it's got its own audio, its own voice. The one I use the most currently is one that makes fart sounds because my little boy finds it hilarious. But have you seen any brands doing these well?
Clark: Yeah. I mean, you could get good ones for...I think there's a tequila brand, I think might be Patrón that'll help you make cocktails and so, you can say, “Okay, Google talk to Patrón” and you can have it, you know, guide you through. Of course, it's gonna nudge you towards tequila-based stuff all the time. Tide, the detergent people, there are over 200 different types of stains. So you can speak to them and say, “I've just spilled red wine on my shirt. How do I get that out?” And it can help you out without that sort of thing. So there are, you know, there are useful things you can do. The interesting bit though is so many companies, and I'm sure some people out there will be thinking, "We did this. You know, we made an Alexa Skill. We all went away on a course and learned how to do it." And, you know, Amazon puts on these day-long workshops to get everybody on board and then nobody ever used it because you've got another point of friction. You know, you have to enable it. And there's a whole store, which is another SEO challenge. Now how do you understand how to get to the top of the App Store within this or the Skill store, as they call it, and make sure people are actually using it?
And for a lot of marketers who will be acquisition-focused, it often seems like more of a brand play, doesn't it? I mean, you're focusing on things like the customer journey. I mean, you know, the Domino's one is really popular because it has a nice little voice and the guy talks back to you and lets you know when your pizza's coming. But if you're thinking about it from the position of selling more pizzas directly, it might help later on. Maybe you come back. But as a paid search marketer, for example, you're not really too focused on that. So the ones I've seen that work well are even just little things like Nutella sending out free samples to people when they make a certain voice query. They were sending out free samples to people. And then you've got the physical product, so, you know, you might go and buy it sometime later. But most of it is novelty that I've seen at the moment, and the clue's in the name, you know, it's Google, it's about actions, about just getting things done. More core brand marketing, how do you stay top of mind? How do you link yourself to the name of a product or, you know, an action?
So it's more subtle, but I don't think it's taken off in the way that they would have expected. You know, there are tens of thousands of skills and actions out there, but it's a bit like apps on a phone, you know, how many do you really use when it comes down to it? Every company has one, but how many are you really using? The bigger way that this will work is integrating all of that into the core search journey. You know, any point of friction is a real problem. People have enough to do. They have enough technology, how can you fit seamlessly in, and often I don't find that actions tend to do that too well.
Will: That's really interesting. Good. That's a good answer now. Thank you. Well, our time is up. Thanks so much, Clark. It's been such a fascinating view on voice search from someone who's clearly thought about it an awful lot. Tell listeners where they can find you on the web.
Clark: Linkedin is always great for me. Just search Clark Boyd and there's only one of me as far as I know, and I have my little newsletter - hi, tech - which has a surprising amount of subscribers.
Will: It's very good. I get it every week and it's a great addition to my inbox.
Clark: Excellent. Yeah. Well, thank you very much, Will. I've really enjoyed that.
Will: Thanks very much for your time and chat to you soon.
Clark: Thanks, Will. Bye.
Will: If you enjoyed this episode, subscribe wherever you get your podcasts, and for more information about transforming your marketing career through certified online training, head to digitalmarketinginstitute.com. Thanks for listening.