Skip to main content



Published on

September 29, 2022


AI, content, content creation

Season 2, Episode 4: True Cost of Content

This week on Outsmart, your hosts Rex and Nicole dive into all things content marketing with Jeff Coyle, Co-Founder and Chief Strategy Officer at MarketMuse.

In the episode we discuss:

  • Viewing content costs through the buying team’s perspective
  • How AI can improve your content marketing workflows
  • Perception vs reality of the future of AI in marketing


Jeff: You know, it’s the equivalent of your batting average, if your content success rates are like 5%, right? Well, just writing 20 times the content that still punches 5% isn’t your most effective outcome.

Rex: Welcome back to another episode of the Outsmart Podcast with TEAM LEWIS. We are your hosts, Nicole Allen and Rex Petrill, and today we are joined by Jeff Coyle, co-founder and Chief Strategy Officer of MarketMuse. Jeff, it’s a pleasure to speak with you today. We can’t wait to dive into content AI and all the implications of what’s happening in the marketplace, now and moving forward. So, Jeff, welcome to the show. Why don’t you tell us a little bit about yourself?

Jeff: Oh, sure. Thanks Rex and hi Nicole, how are you? Um, I, like you mentioned, I’m the co-founder, Chief Strategy Officer for MarketMuse. MarketMuse is a content intelligence platform that sets the standard for content quality. So, what we’re focused on is giving teams insights about what content they need to create, how much content they need to create and update. Where can they have the biggest opportunity, what topics can they focus on, where do they have huge gaps? Whether just looking within their site or looking at competitors to understand where opportunities may be surfaced. And then, you know, we get into details. We build out content briefs at scale as well as we provide applications that connect the most common workflows for writers, editors, search engine optimization professionals and content strategists. I’ve been doing this for, scarily enough, 25 years from the standpoint of building content, building search engines, building products in various spaces. Lead generation, lead nurturing, all the way up to MarketMuse, which is, you know, the first content intelligence platform. So yeah, that’s my background.

Nicole: That’s awesome. And we are users and lovers and admirers of MarketMuse ourselves, our head of SEO, brought us onto the platform, so very grateful for that. Just curious, Jeff, when you were obviously co-founder, chief strategy officer, what was sort of the, the problem that you spotted and we’re looking to solve with MarketMuse?

Jeff: So, it goes along with that whole story, and I can do it in a short amount of time, but normally it would be six or seven hours and a lot of beer. Going from Georgia Tech, I was a computer science major at Georgia Tech, and I started with my first startup when I was a junior there, a company called Knowledge Storm. We were selling leads to software companies before software companies even had content. So, we were convincing Dell and IBM to actually use their brochures and have white papers and those types of things. So, we were selling leads at scale. We had a network of over a hundred websites and I was building both the products of how we show and get leads for people and, and how we distribute and, and everything in between. When we got purchased in 2007, we got purchased by a big publisher. That publisher, called Tech Target, they’re one of the best publishers in B2B technology. They have search security, search storage, computer weekly, I was part of that acquisition. Any type of topic that relates to B2B tech. What I saw there was, by the way, in Knowledge Storm, no content professionals, no writers. We were using other people’s content and we were syndicating it. Tech Target had a team of hundreds of writers, thousands of content contributors. And I realized how manual those processes were very, very quickly. I also realized where data that was often classified as being for the goal of having search engine performance wasn’t making it to editors and content strategists in the right way. It was almost being looked at as, as like a magic trick of some sorts. And so, I looked at all these manual workflows. Some of them took tens, hundreds of hours, as crazy as that sounds, but hundreds of hours. And I was looking to processize that and build out operations workflows, and I knew there was going to be better ways.

So, a couple of those things all then smushed together into one experience when I was looking for some automatic classification technology for taxonomies and I found my co-founder. I found some of the early work that he did in topic modeling, and I asked him, I said, “does it do this?” And he was like, “No, but it does that.” And what that was, was analyze a concept and try to understand what it’s about, try to understand what it means to be an expert on a topic. I had a process for doing that with research, but it was about 30 hours. He had knocked it down to three minutes. And I said, “whoa, this works.” So, I implemented a small pilot, a huge win, it was like five times the traffic to this one segment. And I didn’t give him an easy one either. It was super highly competitive, still by the way, still, winning in that space, and now it’s eight years ago. And then when I left, Tech Target to go work in private equity, he reached back out and said, “Hey, we’re going to really take this to market. You know all the workflows, you know all the pains, the pain of a content inventory, the pain of a content gap analysis, the pain of manually updating pages.” He said, “Will you join as a late co-founder?” I said, “What’s a late co-founder?” He is like, “You’re not going to get paid any money.” I’m like, “Awesome, let’s go!”

So that was basically the real story of that last part. But yeah, it’s all the pains that content strategists, editor, editors in chief, um, and search optimization teams. It’s what makes them fight and it’s the stuff they do manually is all solved by MarketMuse.

Rex: So, you just described quite an evolution there of going from brochures and sales enablement materials that Dell and IBM are like, why would we ever give this to the public? Or what is the value here? So, we’ve gone from that to now integrating AI into the process and cutting down all that manual work and making content work harder and better for your clients, for our clients, et cetera. What, but taking kind of that step back and that overarching, we’re talking about content at the end of the day, so from your perspective, and, and we have this conversation all the time, why is it so important for brands to produce original content?

Jeff: It illustrates that they, you know, the old saying of show them, you know them. Right. You know, it illustrates that you truly know your audience, whether they’re reading, the magazine, we work with tons of publishers. Whether it’s an information type of focus or they’re in a purchasing funnel. Whether the goal is that they’re a customer and you’re providing additional insights, implementation, or troubleshooting. Or something along the funnel of, I don’t even know what I want, all the way to, I’m the biggest champion of your firm there ever has been, this is content that I want to read. Exhibiting that you understand that entire journey of a reader or a buyer journey is extremely important for every type of brand. Because the market’s going to change, your situation’s going to change. People need to know that you are the person that’s going to provide them with that insight.

So, with that, when they move into that funnel or they move along that funnel, they go back to you, or they understand that you have that expertise and that’s going to be where their trust lies. They’re going to read that article; they’re going to move along to other stages. The other side is that, is that exhibition of expertise and that exhibition of authoritativeness is also rewarded with organic search success. So, when we understand the entire journey and we show it with our content, we have the ability to perform in organic search. We have the ability to rank well in organic search and without that it’s a lot more variable, a lot more likely to be fluxing with every algorithm update, a lot more of a house of cards that can fall if we don’t have that infrastructure, if we don’t have that foundation of high-quality content.

And that’s what you see every minute, every day, with big businesses, you see market shift. Businesses that had great content infrastructure had the ability to maintain. The ones that have never explored content, it’s as if they don’t even exist online. They’re relegated to have to buy tons of traffic, spend a lot of money, those types of things whenever things change. And that’s really the story. In 2019, the companies who were out there saying I would never do content, I would never do paid, everything I do is 100% outbound and ABM and then 2020 happened and they’re like, “we got to write content.” So, there’s a lot of different ways that will get you to believe in the fact that you should be there when people are looking. And it’s not just, you know, buying Yellow Pages.

Nicole: And it’s a longer tail play too, right? And we see that a lot of times with brands that we’ll talk to about what their sort of big picture, long term content strategy is. And considering, and again, most of our clients in our conversations are playing in the B2B tech space where there’s like six to 10 different folks making up their decision maker consortium and there’s like seven to 12 touchpoints. And so, when you think about the actual sheer volume of that content, I mean it takes a lot for brands to invest that time and invest the understanding in the audiences and their different pain points and everyone that kind of makes up that buyer’s committee. So, I guess that’s another, that leads me to another question. How can brands really figure out kind of the true cost of content or the return on that investment? Right? Because that is often a big conversation that we have is, companies’ kind of look at it and they go, “Wow, we’re biting a lot off and what are we really getting in return?” Right. Versus just to your point, like throwing some money at the problem and we’re trying to kind of coach folks to be a little more strategic about it.

Jeff: Well, the way that you set up that question actually makes the response the most intelligent response required. Because most people are just looking for what’s the cost of content and what’s the effective cost of content? And that’s a whole answer. But when you talk about it from buying team attribution perspective, it goes into like 3D matrix. There’s a great image if someone searches the Gartner buyer journey, and it looks like this crazy map of lines. He also did one a few years back that looked like a subway map. And the reason why that’s tremendously relevant to this discussion is because yes, you can’t predict the way that the buying team will learn. You can’t predict where in the information journey or buying journey each of those team members are nor what they consume. So, your perspective of this like funnel path, that early-stage awareness, they will read this guide, middle of funnel, they will read the comparison of features. It’s just not real, right?

They’re going to bounce to other sites, they’re going to bounce to other things. So, attribution has to happen in an almost like a blob. And I’m moving my hands, making a blob graphic, but it’s all the content that you have about a particular topic that exhibits expertise no matter where on the journey all works together to allow you to be there when Joey gets added to the buying team and Joey’s like, I have no idea what applicant tracking software is. And then has to go in and then go to Coursera and then go read. You know, they’re going through the process, and you might not even ever know that they did, but you had the ability to possibly be there, right? So, when you think about that, what does that mean? It means that you have content. It means you have content, and you don’t truly know what its goal is. For 100% of buying team attribution, every page you create needs to have a goal to either be generating initial awareness, it’s generating initial entrances to our site, or it’s support content for a particular purpose. Maybe it’s a way finding thing, and it’s never going to be a great landing page, but if someone lands on your guide and then they’re in this industry, you have the industry version of that guide. Well, they’re going to click on that, and that’s really valuable to them. The punchline here is every page needs to have a particular goal, and it all needs to be associated with a collection of content that all works together for a unified purpose.

Now, the cost of content, then you ask somebody a question, how much does content cost? And they’re like, I don’t know, $500. And you’re like, okay, how much of your content is successful? Right? And they’re like, I have no idea. Then you actually do the math. First of all, they didn’t know why the content existed in the first place. Okay, but you get them over that hump. Now, how much of its actually successful? 10%. That’s about the average, right. And generating recurring search engine traffic, it’s less than that on average. That means you create 10 articles for one to be successful generally. Okay. Now you just said $500. Let’s just say that was real, likely not in any sort of business. Then that becomes $5,000 for an effective page because you’re at 10x. All that stuff. So, you’ve got to understand what the page is for, what purpose is it? How much fully loaded was that page? Not just the writer, even if you outsourced it right? You had to send them the request, you had to receive the request, you had to potentially do an editing round. There’s a production round. How much are you going to have to maintain that? You’ve got to bake in the fully loaded cost. If you looked at the page, your salary needs to be accounted for in part to this, right? Internally. So, counting the total cost of content, those numbers get in the thousands no matter what. If it’s a page that goes on your site, for most businesses, it’s over a thousand. If you think it’s a hundred or 200, take a big step back because you got a lot of work to do because you’re not there. And what that’s doing is it’s also debasing the value of content internally if you’re in B2B tech especially, right? It’s degrading what people think content costs and it’s degrading your budget.

So, by saying, hey, I got this article, I got this great writing resource,, and I’m getting that for $150. Like, no, you’re not, first of all. And then plus the odds of that being successful if it’s kind of a bargain is quite low. And then as we’ll get into I’m sure with some of the AI topics, it also leaves a lot of room for it to not be legitimate.

And so, you’ve got a lot of work to do, but content efficiency and effectiveness is how much content do I create or update, and how frequently does that hit its goals? And if you can’t speak that you don’t know your true cost content, and if you haven’t done your fully loaded costs of content estimates which include time, salary, refinements, edits, publishing, production, development, sourcing of imagery, all that fun stuff. Then you really don’t know how much your content costs. And thus, the trust internally, like nobody trusts you internally to actually do what you say you’re going to do. Sorry, that’s like doom and gloom but it does basically resonate. You’ve got to have it all laid out for someone to trust that you can predict it’s going to be success.

Rex: Yeah, and it may sound daunting, but really, I think the key that we focus on there is purpose and goals, and that everything is mapping to your overarching strategy, that everything is working together, that you’re not writing for the sake of writing. We say that all the time across marketing channels. You don’t have to be on every single platform. Do the platforms that you do, do well, right? Do them to the best of your abilities and do them to the purpose of serving your overarching strategy, but then also serving the audience.

So, looking at those pages, you’d rather improve your conversion on success for fewer pieces of content, then just be like, I’m going to play a volume game and hope that 10% work. Cause like you said, your volume game is going to get degraded. Your quality of the content’s going to go down, you’re going to get repetitive, and all that stuff. And your individual value, you know, it’s hard enough to get someone to read a blog page. Why not make it the best that you can and be very strategic with it and have a purpose and a goal for every piece that you write.

Jeff: I was just going to add I agree with that completely.

And what I hear a lot is if you’ve ever heard the bullets before Cannon balls, right? If you ever heard they, you know, shoot, and then shoot the fire, it doesn’t work with content. It might work with content sites. If you’re going to publish 20 content sites and put out a particular amount to all of them and a couple of them will work, then invest more in those. Right? You might be able to get away with that there. That’s why a lot of this advice, that type of advice, comes from the affiliate world. Cause if they publish 20 sites and 18 of them don’t work, okay, cool. Go get another site or go buy another domain, right? You can’t do that if you work at BMW, right?

Or, you know, a particular B2B technology firm. You can’t just pick up your toys and go buy a brand-new domain. These are brands that have significant value, to do anything that puts you at risk, or to burn a lot of money with volume. If you have a pocket of your site, this low-quality content, it impacts the rest of the site. By the way, that’s not like hearsay. That’s literally how it works in organic search. Your topic site section or your topic site authority combinations really matter. So doing anything that could jeopardize those is a tremendous liability. The opportunity cost is one side, but the risk and liability is dramatic.

So, I didn’t mean to cut you off, but as you were saying, I’m like, yeah, you can’t just throw out hundreds of content items and hope some of them work. It doesn’t work that way. Over time the loss that occurs from that is just too dramatic.

Rex: So, we talk about looking through the various processes of going through the identification of goals and purpose, but then also the, the work that goes into creating those pieces of content. And I think, Jeff, you had talked earlier about, you know, 30 hours or whatever it is to just the research aspect alone. How has AI and introduction of technology to the process really streamlined things? And where do you see that value? Where do you see it kind of moving forward?

Jeff: Yeah, no, that’s a great question. Where I see it going is a little bit different than what I think the market perception of where it is going. So, I’ll start with the market perception of where it is going. The market perception of where it is going is, is closing one’s eyes and going, I want a car. And then you open your eyes and there’s a car there, right? But in the form of content. So, the perception that we’ve got it all figured. And that the output of AI is content that’s suitable for framing is really what I think the, you know, the main street of the world believes AI will do for content. Right.

Rex: Like I’m going to feed keywords into a computer and it’s going to write a perfectly polished piece.

Jeff: Right, right. Where I see teams most successful with artificial intelligence is understanding all the use cases and the entire state, all the stages of the content life cycle, and already having a culture of content prior to that and seeing each of those phases get improved by being more predictive.

And so, there’s, there’s some really great maturity models for AI out there. Marketing Artificial Intelligence Institute has a few. There’s a number if you search AI and marketing maturity, you’ll probably find some curves. But one of those things is when you transition from rules to predictions.

So, what you want to figure out is where in your workflow is there manual labor? You also want to figure out where in your workflow your kind of like, you think the reality is based on ifs and thens and rules. When there is actually very little reality that works like perfectly on ifs and thens.

So being able to predict outcomes and then taking advantage of getting manual labor out of the system at each of those stages goes into, you know, making better successful things. Right? So, with content, it’s like the equivalent of your batting average, if your content success rates are like 5%, right? Well just writing 20 times the content that still punches 5% isn’t your most effective outcome, right? So, I like to look at each of those stages, the research stage, planning and prioritization, briefing. So, building out briefs as your sources of truth to keep accountability in place depending on your resourcing, the actual writing process and editing processes, publishing, promotion, as well as post publish updates and improvements in enhancements. And then even getting into reporting. A lot of teams do manual reporting, and don’t have anything predictive in these systems.

So, by going through that process and understanding where you think it’s a black box and not predictive, and what you’re doing manually, you can find the best way to integrate artificial intelligence on your content team. All right, so that’s part one. The other part is if you rush at any of those stages, you’re going to spend a lot of time with checks and balances after the fact with whatever that payload is, right?

So, if you rush on a keyword research plan, you get a big garbage bucket full of keywords, and like they have data points you don’t trust, you don’t believe they’re predictive, well, there’s going to then be a manual process of evaluation or pruning, or cleaning involved. Same thing goes with content. So, if you’re going to generate content, you’re going to treat it like it’s an outsourced writing professional you’ve never met. And by the way, who won’t take your feedback, or can’t take your feedback in this scenario. So, you’re going to be editing, improving, proving, doing developmental edits in the dark, right? So, if you do move too far along, expect that there’s going to be blow back then manual processes because those processes for evaluation and, and refinement can’t be automated when they come from bad payload, right? So that’s kind of like the way that the world works. So, you’d much rather be prepped and clean all the way through. Then using generation can be a great accelerator, using that as inspiration, using that to fight writer’s block, but not using it to do any sort of tricks or magic.

Those are the teams that are having the most success. And by the way, this is coming from someone who built a generation platform. So it works, it just doesn’t work the way people think it works. But it does, I mean, for someone who’s not a great writer, and I’m not, which is always a fun, ironic paradox type thing. My contest strategist is probably laughing, but he normally he’s recording me speak and then helping me to get those things into a beautiful narrative. Solutions that give me outlines that recommend approaches for writing are heaven sent for me, but would I ever take that and put it on a page and press go? No. So this brings in, you know, two potential risk factors for most teams. One, if you do that and you rush, there’s going to be a lot of cleanups, or you’re not going to clean it up and it’s going to cause problems. And that’s what’s happening a lot right now with a lot of writing outsourcing is, it’s actually generated, spun in some way, derivative, plagiarized, and people can’t tell. And it’s causing tremendous problems at all, small businesses all the way up to the biggest right now. Because approaches for detection are not there. So, you’ve really got to do checks and balances across that whole.

Rex: You’re giving hope to all of us marketing professionals who are not going to be put out of jobs by the computers quite yet. So that’s good news for us.

Jeff: It’s not happening, at all. Generation is extremely important to this world. It’s the fastest moving, exponential moving content technology. But I’ve been saying this now for, you know, since I was first inspired to build the generation platform. When I first saw Washington Post’s Heliograph in action from the AI scientist, it was sitting in a conference in Las Vegas and I was watching him build out content for the Olympics and I was like, “oh my gosh, this is so great. I can do it better.” So, what people think it’s going to do isn’t what it’s actually going to do in most cases in natural language processing and generations illustrating that right now as well. There are small use cases for whenever high-volume permutations and testing is needed. You know, like ad text is a great use case, social ad text, you know, those are great use cases for high volume with tests. But on its own, content at scale for a site that requires longevity, it’s not the best use case. So, there’s going to be refinements and improvements that are needed to make it work for that specific workflow.

That stated data, templated, conditional templating using data to build narratives, and then taking a storyteller’s eye to that, can be an extremely deadly weapon. So just think about where your building blocks are, and you are an editor transitioning from being a writer to an editor. Some people aren’t good at editing, right?

You may need to take an interview from a real expert, combine it with this data story that you got, combine that with generated text to build something magical, and then that may improve your hit rate by two x, it may speed you up four hours. It may also slow you down. You know, it might slow you down the first 15 times you do it, but times 18 to 50 it speeds you up.

So just like with any technology you got to figure out all of those components. And what’s happening in the market is nobody’s doing that second chunk of work, right? They just think magic trick and publish. They’ve got to go back and realize that this is just yet another source of content. It’s potentially, you know, just an outsource writing professional.

I might want to pair it with data stories and do things like the macro-SEO industry does. I might want to pair it with real editorial work and real journalism. And that’s what people don’t want to do, because the goal, they really just don’t want to do work at all. And what this is doing is, is actually potentially creating more work.

We’re making it so you have to build a completely new infrastructure. But we’re all laughing because we already know this. Because if you’ve managed a content team, or worked with an agency with 20 writers, you’ve had to do this anyway. It’s just another really fast outsourced writing professional for you to deal with who you know, some days has really bad days.

Nicole: It’s helping us all work a little bit smarter, right? It’s like Clippy 5.0.

Jeff: Yes, Clippy. Oh, the original Clippy, by the way, fun fact, the original Intel, it’s called Intelligent Agent. If you’re not familiar with the term, the original intelligent agent was actually really smart. Clippy was actually really smart, the original one, but it was so big. It was like the biggest application ever, like massive. I forget the number. It was hundreds of megabytes during the time period when that was a lot. Right? That had to be dumbed down. Go look it up. That’s a real story. Yeah. I love intelligent agents. I think they’re awesome, but yeah, they just never got implemented right.

Rex: You just can’t have word taking 20 minutes to open every time because of Clippy.

Jeff: It really did. The original intelligent agent was like this poorly monstrous software product that never really worked well. It really worked well, but it never really worked fast enough to be into production. Total tangent.

Nicole: No, I liked it. Anything else that we want to hit on?

Jeff: So, one thing I often see with people who are integrating artificial intelligence to their content as part of like a content auditing or inventorying process is that they will look at the content and they will try to compare it only to competitors to see how can they be more like their competitors or how can they do the things that the people who are winning in organic search are doing.

So, they’ll copy who is the number one. By the way, terrible idea, especially if number one is like a huge enterprise that does things their own way and kind of gets away with their version of murder in the organic search results. What’s really important is that you do kind of a two-part play with data analysis of understanding what it means to be an expert about the topic, understanding that entire journey that I talked about before. So, information journey, buyer journey, don’t forget the post-sale, and all the concepts that need to be included in order to exhibit expertise for the goal of what you’re doing. Right. And so, what that also can do is it, you got all that information, you have the topic model, you have all the concepts, you understand the cluster that you’ve got to cover. Now you can look at the competitors with a lot more effectiveness, because you can see okay, if you’re writing an article about, you know, content marketing strategy and you don’t talk about content marketing, well that’s the table stake stuff, right? But what you’re also going to see are angles and concepts that experts would cover, but nobody is. Right. So, if you’re doing this the right way, you see topics that would differentiate you and exhibit expertise that the top five articles aren’t even thinking about. For one, maybe they’re two keywords focused, maybe they’ve just got these long form meandering guides that somebody told them they should write, right?

But when you can really dive in and do research on these topics, you can find concepts where you can be differentiated. So that you get more opportunities to exhibit expertise, it might give you an idea for differentiating this piece and writing another one that’s a support piece for this page. So that’s the type of way that you can approach using natural language processing and topic modeling in a much more effective way than just what you often see by the hawks in the industry saying, “copy your top three competitors, smoosh them all together, do what they do.” Totally bad idea, by the way. When I see people doing that, it’s wonderful. It’s like chopping down the tree becomes real easy for my clients. But what you really want to be thinking about is how do I ensure that I’ve got both the table stakes stuff, but I’m also differentiated. I also have angles that nobody thought about so far. If I’ve got those, I’ve got a lot more longevity. So that’s the approach to AI from a content, page by page perspective that really differentiates teams. And by the way, the best publishers in the world, they’ve turned this into an art.

Nicole: Yeah. That’s the stuff that gets me excited because, I mean, I feel like every day clients are asking us, what’s the white space? Like okay, I know where my competitors are, where are they not like right? And what are those conversations that are happening in the places where they’re not? And trying to kind of eek out those kinds of topics and conversations and turn them into a content strategy or a media relations or thought leadership strategy or whatever the case may be, to really to fill that white space.

But it’s the ability to identify those gaps. That, to me, that’s where I see applications of AI that get me excited because that was kind of my background in being more manual and kind of like creating, you know, going through Google and different site maps and just kind of creating my own little, like, beautiful mind of where everybody’s living. And having the ability to have that kind of technology cut that time down and make it less of a manual process, um, is really exciting. I think.

Jeff: Yeah, and I have examples of that every day, and it’s, it’s really like the old manual process of, of blue ocean, red ocean, right? So, it’s, you know, I can go write the same article everybody’s written in that bloody red ocean where I can figure out what do I know that sets me apart? What information does my business have that nobody could replicate? Those are my favorite articles are the ones that, like I’m using data from MarketMuse, for example, that nobody has. Or I’m using statistics that I know I’m using expertise that only I know about. You might not be able to do that with every article, but what it can do is it can make your red ocean content do even better, because it’s being propped up by some uniquely differentiated pages.

So that’s just something to always think about if you’ve got clusters of content on your site that really provide no additional value, you’ve got problems. And one of the ways to get out of that is to really start thinking about what you can do to be differentiated.


Nicole: Well, Jeff, we’ve stolen enough of your time today. I know that you’re in prep mode for a conference, so really appreciate you taking the time to chat with us. I think we’ve both had a lot of fun in the conversation and just thank you for joining us on Outsmart.

Jeff: It’s been a pleasure. Thanks Rex, thanks Nicole. Anytime.

Get in touch