Jeff Coyle: AI and the Future of Content Strategy – Episode 124

Podcast: Play in new window | Download

Subscribe: Apple Podcasts | Google Podcasts | Stitcher | TuneIn | RSS

Artificial intelligence is driving a wave of change in content strategy practice.

AI is here now and will soon touch virtually every aspect of content work, from ideation and creation to management and analytics.

Jeff Coyle can help you understand and navigate how these new AI-driven practices will affect your content career.

We talked about:

• his work at Marketmuse, a content intelligence platform
• some highlights of the recent Marketing Artificial Intelligence Institute, where he spoke about language and content effectiveness
• the prevalence of AI in the work he does
• what you need to understand about AI as a content practitioner
• how AI tools can help identify content opportunities that you might not have seen before and make better decisions
• some of the specific benefits of AI for content: automated inventories, better search, evaluating content quality, identifying plagiarism, governance, etc.
• new job opportunities that AI will create
• the importance of staying focused on quality content as you adopt AI
• how AI can help with content analytics
• the decoupling of content metrics from page-level analysis
• the importance of a long-term content perspective and being the storyteller in your space when applying AI to SEO
• his ongoing advocacy for making better content decisions

Jeff’s bio

Jeff Coyle is the Co-founder and Chief Strategy Officer for MarketMuse. Jeff is a data-driven search engine marketing executive with more than 21 years of experience in the search industry. He is focused on helping content marketers, search engine marketers, agencies, and e-commerce managers build topical authority, improve content quality and turn semantic research into actionable insights. His company is the recipient of multiple Red Herring North America awards, multiple US Search Awards Finalist, Global Search Awards Finalist, Interactive Marketing Awards shortlist, and several user-driven awards on G2, including High Performer, Momentum Leader and Best Meets Requirements.

Prior to starting MarketMuse in 2015, Jeff was a marketing consultant in Atlanta and led the Traffic, Search and Engagement team for seven years at TechTarget, a leader in B2B technology publishing and lead generation. He earned a Bachelors in Computer Science from Georgia Institute of Technology. Jeff frequently speaks at content marketing conferences including: ContentTECH, Marketing AI Conference, Content Marketing World, LavaCon, Content Marketing Conference and more. He has been featured on Search Engine Journal, Marketing AI Institute, State of Digital Publishing, SimilarWeb, Chartbeat, Content Science, Forbes and more.

Connect with Jeff online

• email: jeff at marketmuse dot com
• Twitter
• LinkedIn
• MarketMuse

Video

Here’s the video version of our conversation:

Podcast intro transcript

This is the Content Strategy Insights podcast, episode number 124. Creating and managing and measuring the effectiveness of truly intelligent content requires a good understanding of artificial intelligence. AI tools and practices create many new opportunities for content strategists, and they give content managers new ways to understand the performance of their content. Jeff Coyle is an AI and search-marketing expert who can help you account for the implications of artificial intelligence as you navigate your content career.

Interview transcript

Larry:
Hey, everyone. Welcome to episode number 124 of the Content Strategy Insights podcast. I’m really happy today to have with us Jeff Coyle. Jeff is the co-founder and Chief Strategy Officer at a company called MarketMuse, and welcome, Jeff. One thing I want to say, and you’re just fresh back from a really cool-sounding AI conference. Tell me about that and what you’ve been up to at MarketMuse.

Jeff:
Yeah, sure. So like you mentioned, I’m the co-founder and Chief Strategy Officer for MarketMuse, and MarketMuse is a content intelligence platform that sets the standard for content quality. And what I mean by that is we give insights as to what you should be creating, what you should be updating to tell the story of quality and expertise, and also expand the opportunity that you have. So we can tell at the site level, at the network level, or even at the page level where you have strengths and weaknesses. Gaps and opportunities and everything we do comes from a lens of quality and authority. So what are the next 10 pages I should write on this topic? Or what are the next 10 pages I should write, period, about getting quick wins. Anything in between an automated gap analysis platform and content auditing and inventorying platform, all these miserable processes. We built them into software that content strategy teams just painfully trudged through once a year, we can have that on-demand. We also automate content briefs and build out workflows for content teams and SVOs.

Jeff:
So yeah, this event was fantastic. It’s put on by the Marketing Artificial Intelligence Institute, which is a team led by Paul Roetzer, who more recently is focused on this, but prior to that, was the founder of PR 2020. He wrote the marketing, the agency blueprint, which brought the point system to agency. He’s just this really amazing visionary. This event had folks from Ernst & Young’s AI group, Google AI, ‪Vedant Misra, who I had the opportunity to introduce in the closing keynote from Google’s team, he’s worked on Minerva. Previous to that, he was at OpenAI. So he worked on GPT-3 and number and codex, which is the code writing component and many other things. And there was great panels on computer vision, a lot of discussion of image generation, a lot of discussion of natural language processing, prediction, generation. And I was lucky enough to speak about language on the initial keynote panel, and also talk about content effectiveness and efficiency and how you can really get predictive if you implement AI with what you write. You don’t need to just write and hope anymore.

Larry:
Yeah. And all that stuff you were talking about, all the things you do. I mean, is it true to say that AI is the technical underpinning, the power in the background that drives all of that?

Jeff:
Yeah, absolutely. Whether it be natural language processing as a branch of artificial intelligence, components of machine learning that work into that, or even elements of generation. But also just better ways of doing graph theory or processing large amounts of data, building language models. All of those things are branches of artificial intelligence that we work in, but being at that show, it’s an illustration that the branching is not even close to being finished. We’re talking about now, new job types. We’re talking about the types of businesses that we’re going to deal with are going to be either AI native, AI emergent, or obsolete. And there’s not going to be a question about a company not embracing artificial intelligence regardless of where they’re working, what they’re working in. Over time, completely obsolete. And there’s really not a question of whether. It’s a question of when, now.

Larry:
One thing, I know I have some CEOs and C-suite and executive and founder-level folks who listen to the podcast, but I think most of my listeners are more practitioners. Content strategists and content designers down in the weeds actually doing the work. And then that trio you just mentioned of native to emergent to obsolete. I think everybody wants to avoid that last category. Do you have any thoughts about how content practitioners can get… Well, I guess, and let me start that by saying, let’s define some of these terms about AI, because I think you don’t have to get too far in. I’m by no means an expert, but I know what NLP – natural language processing – and machine learning, and then that things like GPT-3 are based on large learning models and there’s a whole bunch of lingo and practices going on in the background. What do you think is most relevant for people to understand as content practitioners?

Jeff:
I think a lot of it is that a lot of these technologies and concepts, especially in the field of linguistics and language, have existed for a really long time, but they’ve been advanced with artificial intelligence technology. They’ve been advanced with cloud computing capabilities and just the improved capabilities, other capabilities and innovations, so much that those subfields of content, linguistics, language, science have just changed such that while those might be the guiding principles, the implementation of them becomes so accessible and so easy to implement and so interesting for every team, not just the big teams, not just the ones that already have implementations. The accessibility really changes it. So, natural language processing, it’s branch of computer science. It’s been around for a long time. You’re parsing text, you’re learning from it in some way, but the expansion of the capabilities makes it so that what you can do with it.

Jeff:
Anytime texts exists, effectively, anytime audio exists that can be converted into text. Anytime now images exist or vice versa, they can turn into trainable sets. They can become parts of predictions. They can become parts of other tasks. There isn’t a situation now where if text items exist, there isn’t something that you can do with them that you couldn’t do with them four years ago, three years ago, two years ago. And then the implementation, the use cases for a content strategist then become very exciting, because you can take text data from anywhere and turn it into your elements of your content strategy, which you didn’t have access to in the past. And that’s what really excites me. It’s being able to break down the silos that have existed in all of our computer systems, all of our business systems, which we don’t even think about as being part of our content, and learning from them to potentially either make those things part of our content strategy, or know what that is telling us we should be writing about.

Jeff:
And when you do that, that’s when everybody starts opening up their brains and saying, “Whoa, wait a second.” We have 50 people in our organization who are focused on product marketing materials, or post-sale docs, or customer content, and none of that goes into the content that goes on our website. Wait a second. None of that’s being used in any way. We’re not learning from how people read those documents. We’re not learning from how people try to access them, what they’re looking at in order to educate the rest of our materials for our own in-house materials or the things that we use for internal communications. There’s just not a situation where if text exists, it should and could be used across the organization.

Jeff:
Another obvious use case that I always like to talk about is phone information, calls, whether it’s a call with your team, whether it’s a call with a prospect, whether it’s a call with a customer or a support rep. The intelligence one can glean from that is, again, it’s the accessibility and natural language processing gives us the ability to learn from all of those things and take all that into account. And I’m not talking, like rudimentary stuff, we’ve always had this. Now, it just becomes accessible for everyone to really want to take in all the content in all the text to make better decisions. And I’ve seen some amazing platforms that can take 50 disparate data sets from every organization and start to do fun, interesting things like predict outcomes for, is this lead going to be a good one? Before they even interact with us.

Jeff:
Or is this a churn risk? Is this a customer that’s likely not going to be around? We can look at all this data that we have and start to predict it. Or what should we be writing about? Because we are missing content from an entire stage of the buyer journey. So any kind of use case has become more accessible, and that’s the big takeaway that I have. There’s going to be major job changes. There’s going to be major organizational changes as a result of this, but where we are right now, there isn’t a use case that can’t be influenced by AI.

Larry:
It sounds like from what you said there’ll be a lot of change, but you’ll still have a job. It’s not like a robots coming to take your job scenario. And a lot of what you were just saying about making better decisions by accessing the content you already have and looking across, like if you’re in marketing, looking over it. I’m thinking of just a lot of the work that’s done. I remember I had Scott Abel on the podcast a couple years ago and he was talking about technical content, and they had measured in that world that something like 45% of tech writers’ time was just trying to find stuff and figure out what they have already. It sounds like AI can help with that as well.

Jeff:
Yeah, absolutely. Automated content and inventorying solutions, or better search technology, internal search technology, is all completely available. Searching by relevance, use case, automatic tagging of content is far more sophisticated, but also you can evaluate the quality of content. You can evaluate it based on metrics that had never existed. Being able to say whether this exhibits signals of expertise, whether it is potentially wrong, whether it’s plagiarism. You can get into whatever your use case is, whether it is within the terms of governance for your organization is another really interesting field for content governance. And that gets more advanced year, over year, over year. All of those things really get into it. When I said the team changes, the robot is not going to take your job, but the robots are going to certainly put pressure on low-skilled, low-quality, heavy. If you have organizational debt or technical debt where you can’t execute as fast as a fast or nimble team could, the robots are going to eat you at some level.

Jeff:
And if you are inefficient, an efficient robot or an efficient AI is going to be better than you. If you are at the top of your intelligence, your creativity to guide these things becomes a hugely valuable asset. The creativity, human creativity, isn’t something that’s going to be able to be easily replicated or imitated, as we kind of say. There’s imitation in language that will be accessible. It already is. Creation of content and generation or in response, prompt-response, whether it be an answer, chat bot, or those types of things, can pass the test of this could be written by a person. It may be weird. It may be a little off. They may be lying, which I would love to get into, as that is the popular misconception about generation and sentience versus imitation. But the human’s ability to use that information is one component.

Jeff:
The other one, and this was a really cool concept, is there’s going to be jobs that are prompt engineers in the future. Prompt engineers. And that’s a person whose job is to send the right information into the artificial intelligence and be able to craft . . .

Larry:
Is that the human in the loop?

Jeff:
Yeah. It’s a person who’s, popular DALL-E-2 for vision. You’re really smart and you’re really good at presenting the prompts to the AI that gets you back the images that are the right images for the project that you’re working on. And maybe there’s a tweak needed. Same thing for language. Same thing for parametric. There’s language models, one I know that I built, that are out there where you can send in parameters and infrastructure and structure so that the output is more likely to be aligned and to take less time for then you to have to edit or improve.

Jeff:
So those types of roles using these technologies becomes a hugely in-demand practice and designing the why that you would use them. And that’s where this really becomes scary from the lens of, where’s it going to go? It’s not really a jobs are going to go away, it’s that if you’re making a living writing low-quality content that doesn’t perform, that’s going to go, because it probably shouldn’t exist anyway. But that’s really the head of this, is to say that that might go away. And anyone probably listening to this show is going to be better for it, because we’re going to have so many new capabilities that we couldn’t have imagined at our disposal.

Larry:
And you mentioned a couple things. I’m kind of an individual contributor. I love practice. I’ve done a lot of management and I’m interested in it, but you’d mentioned a few times decision making and guidance, things that sound like management and leadership kind of stuff. But you also mentioned this new of individual contributor role of the prompt engineer. It sounds like this is going to affect everybody up and down, to the extent that hierarchy still exists. However jobs are organized, this is going to suffuse everything, it sounds like.

Jeff:
Yeah, I think it’s going to separate the “get it done” versus “get it done better” mentalities. Where people trying to cut corners with AI, in every industry, in every space, the cut corners professionals, the ICs, individual contributors, who take technology as a way to just fast-track without a lens for quality, if you look at the historical cycle of the world, of every technology, they often have a boom, but then they have a bust, and we’re in the early phase of writers and editors booming and then busting. They haven’t yet fully busted, when they’ve taken the desire to use this to cut corners or to fast track without a lens of quality. And I think we’re in this booming lens of generation to generate lots of crappy content, just to be blunt, and get it out there and see if it turns into magic and money and promise. That isn’t where it will land.

Jeff:
But I think what it does is it says, okay, well, if I’m a writer, how should I be thinking about this? If I am a content creator, how should I be thinking about this? What we shouldn’t be doing is sniffing and going nah, bad or yay, good. The human mind reads content and typically goes good or bad. But what we should be thinking about is where in the stages of the content life cycle am I doing manual tasks? Am I doing repetitive tasks? Am I doing tasks subjectively that aren’t using data? There’s no why. All the way down to is the output hitting expectations for performance as well as quality?

Jeff:
So analyzing your own process is the superpower of today. Going through it and really, really digging in. Being honest with oneself. How much time do I have writer’s block if I’m a writer? If I’m an editor, how many developmental edits do I do that created cogs of loss, time loss, where the writer then struggled to respond to that developmental edit? Or how much time did I spend copy editing? How much time did I spend? And then understanding that fully-loaded cost, not just of how much money that writer charged you, but time, how much costs were connected to how much upkeep is needed on this piece, this article, and then connecting that to performance. That’s your job now.

Larry:
That gets into whole metrics world. And you just mentioned from the creation side, I love that you can, and I can of picture how you could infer writer’s block and editing snags from this. But I’m wondering, are there other levels at which this can help? For example, one of the things I see a lot, there’s a lot of people working with decoupled content and trying to measure content performance independent of page level, Google analytics kind of metrics. Can it help with that kind of stuff?

Jeff:
Yeah, absolutely. I mean, for sure. And it is certainly dependent on, every page or every collection of pages need purpose. So the purpose of a page, and this is why I struggle a lot with the certain types of content as SEO audits. I’ve been in the search and content space for now 24 years, as scary as that sounds. But if you only evaluate content based on performance metrics, you miss the whole point of content. And it’s to say that some content is support content. Some content, their main goal is wayfinding from another power source who might be more charged with entrances or might be a marketing conduit. An example of that is, I have a long form guide on my site that gets a lot of early stage awareness traffic. But when they land there, I have content that appeals to many industry-specific areas.

Jeff:
They’re not going to get a lot of entrance traffic to those, but I want somebody who lands on that basic guide who is in that industry to know I have a home for them on my site. If I go through and I sort descend my pages for Google search traffic, those industry-specific wayfinding assets are going to be near the bottom. It would be foolish for me to delete those for that reason. But that is how a lot of people profess doing audits right now. It’s bananas. And so the decoupling of stats from pages, it goes along with my perspective of content working together as a blob. It’s a blob of expertise that tells the story that we understand you, customer. We understand you, prospect or reader.

Jeff:
We have informational, pre-funnel, we’ve got awareness, we’ve got consideration, we’ve got purchase. Or if you’re at IBM, care, consider, choose, whatever you want to call it. We also have post-purchase, implementation, troubleshooting. We’ve got content for our customers. And by the way, all that I just said is on one topic. So the world of one keyword to one page, it should have never been in existence, but it certainly has been dead the last decade, and if anyone’s telling you otherwise, let them come see me.

Larry:
I got to say, that’s something that me nuts. I got out of SEO about 10 or 15 years ago, and people are still doing it the same way. And even back then, Google was clear that they were more about satisfying user intent. That’s why I’ve identified as a UX practitioner for the last 15 years, because I thought that that would be better for my SEO. I could serve people better by doing that. But am I just being cranky or am I correct in seeing things that way?

Jeff:
No, you’re totally correct. I mean, for longevity with a brand, there’s still ways to cut corners. There’s still ways to go aggro and be super aggressive and get dividends. There will always be. There’s always going to be a fringe, but that fringe acting in a pedantic way in order to say this, “I made this work one time with this crazy aggressive exception” helps no one in the industry. What it shows is that yeah, cool. You can do that. But if that website that you did that on all of a sudden shriveled up tomorrow, what would you do? You’d go buy another website. But you can’t do that if you’re a brand. We took that gamble and it didn’t work out, so let’s just flush it and go start our new…

Jeff:
If you were working on superhomeroboticvacuums.com and that crashes, you go by myhomeroboticvacuums.org and try it again. That world of aggro or affiliate, in a lot of cases, the affiliate world. The lessons from that don’t apply realistically into small business all the way up to large enterprise for that type of longevity. If you want to run with razor blades with those businesses, everybody in the business better know about it. Your CEO better know about it, too. Because those decisions, those mistakes, just are painful. I know I’m getting a little bit off of a tangent, but you tell I stand on the same soapbox you do. The cranky old man going, “Don’t do that!”

Larry:
No, you mentioned governance a while ago, and all of a sudden you have me thinking about brands contending with little mushroomy popup micro sites, doing stuff like you’re just talking about.

Jeff:
Yeah, well a lot of times the big brands struggle with this yin and yang. What they aren’t seeing, though, is that other people are controlling their message. And as a content strategist, that to me is the biggest story for small enterprise to enterprise, whether you’re in direct consumer, fast moving goods, B2B technology. If you have a product that a lot of people love or buy and know that it exists and you’re marketing it into the journey, letting other people control your journey when you have the opportunity to communicate against that journey becomes the content strategist’s meaning of life internally, in my opinion. So if I work for a shoe company, but all the benefits and features of said shoe company, all the reasons why people should buy it, all the people that love it, those are all being talked about on publisher sites that aren’t even affiliated with me.

Jeff:
And I’m not taking the opportunity to be a publisher and tell those same stories. Not only is it the massive missed opportunity, but it goes against the world of why one would want to be a content strategist. Someone else is controlling my message. People are researching. You look at the people are going to B2B tech. They’re going to G2. Great. G2 is great. They’re going to every other site. They’re going to Yuk Yuk blog, who said they had a bad experience with your site. They’re going to Stack Exchange. You’re not taking the opportunity to talk about implementation wins and troubleshooting, or you’re just not taking care of that section of your site if it’s on a support forum. It could manifest in any way, but the desiloing of content and the empowerment of content strategists for a medium to large brand is, “control our message so that we have longevity. ”

Jeff:
And that’s what, coming back to AI. AI can tell us where we don’t have any footprint. And a lot of times, it’s in the middle of the funnel and it’s in the funnel that’s being covered by publishers. So if you are working for a middle, mid-market to large enterprise, it’s where are people finding out about our content or about our offerings? And if the answer is from our competitors, from publishers, or from aggregators and not us, you got to reload the mirror. That’s a problem.

Larry:
And I can see how AI could help with that. I had May Habib on a couple of episodes ago and she talked about that that’s one of things that people are finding with her product. She used the analogy of it’s like you’ve been working with the same agency for 20 years and I have this stable, centralized content strategy, that AI can give you that kind of overview of what you’re up to. And like you said, the implications of that for branded CEO are just off the charts, I’m sure

Jeff:
Oh, yeah. I mean, being able to be the storyteller in your space. I mean, just look at the brands, even in eCommerce who are dominating here and doing well is key. So

Larry:
Yeah. Hey, I can’t believe it. We’re coming up on time already, Jeff. Anything last, has anything come up in the conversation that you want to make sure we get to or just that’s on your mind these days?

Jeff:
Gosh, really, no. I mean, shoot me a note, jeff@marketmuse.com, if you want to talk about content strategy and AI. I, humorously enough, will usually engage with those types of things as long as they are those types of topics. Jeffrey_coyle on Twitter, jeff@marketmuse.com or LinkedIn, always easy. For me what I’m looking to identify really is the true value of making content decisions properly. I’m trying to advocate and evangelize for that. I want everyone to want to know what their batting average is with content, that if they publish 10 articles and only one of them works, whatever works means, that over time, that’s not a business. It’s not a practice. It’s not a meaningful organization. It’s just something where at some point, people are going to give up or they already have. And so my job, I feel, in this industry is not just content strategy, content intelligence, but it’s making us all predictable as an industry, as a group of content strategists, gives us longevity.

Jeff:
It makes content strategy relevant. If we predict more winners more of the time, because we don’t look like we’re playing a guessing game, or we’re just standing on some sort of creative brief, or we’re not standing on governance. We are part of the wins and we’re part of the wins all the time, not just a small fraction of the time. And so that’s the role I’ve taken in this industry, is to tell people and educate folks that if they’re batting 10%, they have to bat 30% to stay relevant inside the organization, or else they’re going to become dinosaurs. Even more dinosaurs than I am.

Larry:
Or me. Yeah. Well, thanks so much, Jeff. I really enjoyed the conversation.

Jeff:
Been a blast. Thanks again, Larry.