MarketMuse accelerates content planning, creation, and optimization through the use of artificial intelligence (AI). The platform helps you to create relevant, authoritative, and helpful content at scale that’s search optimized for high performance in Google. Taking the guesswork out of content development, MarketMuse primes your business for greater SEO and content marketing success.
This explains why the company has thousands of corporate customers.
Stratabeat had the privilege of interviewing Co-Founder and Chief Strategy Officer Jeff Coyle on the latest episode of B2B Marketing Tech Talks.
Grab your coffee, sit back, and hit play to meet Jeff and learn more about content scaling with MarketMuse.
Tom Shapiro: Welcome everyone to another episode of B2B Marketing Tech talks. I’m Tom Shapiro. I’m the host of the show, and I’m CEO of Stratabeat, which is a branding, web design, and marketing agency. Today, I’m very happy to have special guest Jeff Coyle with us, he is the Co-Founder and Chief Strategy Officer over at MarketMuse, which has a platform that helps content marketers to achieve massive success through the use of AI. I’m very happy to have you here, Jeff. I’m really excited about the topic. I think it’s going to be fascinating, and I’m eager to dig in. So, welcome!
Jeff Coyle: Thank you so much for having me. I really appreciate it.
TS: Absolutely. So let’s just start off with a little bit about yourself. Can you give us a little bit of background?
JC: Sure. So, as you mentioned, I’m the Co-Founder and Chief Strategy Officer for MarketMuse, and MarketMuse’s mission is to really set the standard for content quality on the internet.
My background, though, goes over 21 years in content strategy, search engine optimization, lead gen, paid, A/B multi-variant testing, community management—anything that relates to how traffic goes in and turns into something valuable for a site, I’ve probably done it.
I worked at a company called Knowledge Storm after doing some work when I was in school at Georgia Tech in usability theory, search engine design, and ad server design. And what Knowledge Storm did was, it was one of the first companies selling leads to B2B technology companies. We were trying to get companies like IBM to do content marketing. To actually get their content in the form of white papers in the form of case studies on the web so that we could syndicate it to generate leads for them.
This is when generating leads on one’s website wasn’t even really a thing. And we were acquired in 2007 by a major publisher who had a huge editorial team—over 200 editors, over 1,000 content contributors, and 100 plus websites. And the interesting part about my story is that was really the first time that I had interacted with really advanced subject matter experts who were also editors and writers. I mean, the big thing about the way that they work is that they’re very manual, but it’s very much related to their artistic ability, to their writing ability, to their subject matter expertise. And so, that’s a lot of manual labor. That’s a lot of brainstorming.
That’s not a lot of data-driven insights. And I was bringing that to the table and trying to integrate that into their manual workflows.
And I realized really quickly that what they don’t want to receive is a list of keywords, right? They want to see kind of why this type of analysis is going to give them insights that are going to amplify their expertise. It’s going to make them have a kind of a shared understanding of what success is. And that’s where building out those workflows and kind of working out how to do those things manually at the tail end of my experience at that publisher, I met my co-founder, my now co-founder Aki. And he had started to take some of these workflows and automate them. And one of them was the work of topic modeling. So he had already distilled basically a 30-hour manual labor process to build a topic model that wasn’t even very good into two minutes, and it blew my mind.
It blew my mind. Right?
And so when I left that publisher, I was working at a private equity firm. He came back and reached out and said, “Hey, Jeff, you’re the only one that knows all these manual processes and workflows and why some of them work and why some of them don’t. Will you join me as a late co-founder?” I said, what’s a late co-founder? And he said, “It means you’re not going to get paid for like 18 months.” And I was like, Oh great. So I can either go work over here at this equity firm, or jump off the bridge. And obviously I jumped off the bridge.
Now, we’re 50 people crushing it, thousands of customers, and I really have a solution that solves so many common reasons why SEO, content, marketing, demand gen, and writing teams don’t work well together.
That’s our goal is to bring those teams together so that nobody’s doing things subjectively, because there’s a lot of room for subjectivity. It’s a good thing. But when it’s backed by some data, it becomes extremely powerful. When a great writer, when a great subject matter expert has data behind their bullets, the bullets get bigger and they get more accurate. And when you can predict the success of content before you even start writing it, everyone is happy.
And we see that time and time again. It takes that lack of confidence and eliminates it from the editorial team, from the writing team, from the content marketing team, from the demand gen team. And they say, yeah, if we cover this topic comprehensively, if we cover this topic like we were an expert, we’re going to win. And that confidence just turns teams into superhumans.
TS: That’s awesome. And one of the things that you mentioned is the predictability of content marketing success, and that’s invaluable.
JC: Yeah, exactly. We see so many times people think that the data points they’re using to drive prioritization are predictable. But then when we unpack it, we look at their content efficiency—which is a metric of how much content we publish or how many motions of updating content we take against how many of them achieve our goals—those average about 10%, terrible. I mean, you literally, so every 10 articles someone publishes, one of them achieves its goals is our average that we see when we start working with you.
And we’re like, what if we took that up 30, 40, or 50%? What things would that change? And it’s not just the number, it’s everyone on the team feeling really good about the work they do. Nothing’s worse than writing an article or updating a page and optimizing it or making it better or expanding it, and nothing happens. I mean, that has the morale impact, the sidewinder spiral that that has is not easily quantifiable, but we just see it turn really mediocre teams into magic.
TS: That’s awesome. So I want to segue into something that I saw on your website. So you claim that you help brands to consistently craft the best content on the web at scale. And it’s a bold statement I love it. But I’m wondering if you can speak to it, how exactly do you help brands win at scale?
JC: So, I like to break down the content lifecycle into pieces. I’ve got research, planning, briefing, writing, editing, publishing, and then optimization or improvements. And there’s also reporting and promotion. Boom. Get into that. What I like to do is inject novel approaches of using artificial intelligence into each part of that cycle. So that no matter where you fit in the puzzle or on the team, there is a solution that’s going to amplify your expertise. So in the research side and the prioritization side, it’s not just pulling out of a database and sorting by a field, which is unfortunately what about 90% of search engine optimization professionals do—they still cling to Google AdWords, keyword planner data, or search volume data for all of their work. It doesn’t work. It’s not predictable, and it’s at best a finger-in-the-air guess.
And the way that you can unpack that is through content efficiency math, or by looking at a case where you use those data points to predict how accurate are your predictions. And when that number is less than 30, 40%, it’s actually not predictive. You might as well flip a coin at that point to guess what you should do.
And so, what we do is we try to bring novel data points. We set the standard for content quality and we have a way to approach the evaluation of content comprehensiveness and quality. So we can actually tell you whether an article exhibits the signals that a subject matter expert would. And we’ve quantified that as a score for content, as you can see, whether it matches up to what we believe an expert would write if they were covering the same topic.
So, if you look at that at the page level, that’s very powerful. So now we roll that up at the site level as well. And we look at proprietary metrics that we’ve built for topic authority. So we can say how much of an authority you are on a concept. How powerful are your pages? And not just links. How powerful do your pages cover the topics that they’re focused on? How much breadth of coverage do you have on these topics? How much depth? How frequently do you cover them like an expert would? How much momentum do you have with your clusters of content? Okay.
The reason why I answered the question that way is because if you have those two sets of data points, you can make sure that you are both picking the right content to create on the right topics or update to expand your existing pages and each to actually execute and get you to the last mile with the content, to actually craft the outline properly, or ensure that it meets a standard of quality. So what should I do? Actually execute it. And when you got those two pieces, you can consistently craft the best content on the web at scale. There’s no reason to ever write content that isn’t equal to or better than all of your competitors every time, and MarketMuse is the only solution that allows you to do that.
TS: That’s amazing. And I think you kind of answered the question that I was going to ask you next, which is all around differentiation. The martech space, especially around content marketing and SEO tools, is so crowded. Just crazy crowded. How does MarketMuse differentiate itself? You hear some buzz around GPT-3 and solutions like that, but if you could speak to that, that’d be great.
JC: Yeah. I mean, I think there’s three classes of differentiation. What you mentioned and GPT-3 is in natural language generation, about which I will definitely get into some details. We have a competitive product to GPT-3 in market. It’s called MarketMuse First Draft, and we’ve built it all ourselves. We’re not reliant on any third-party data for that solution. It is a rival solution with Microsoft and Open AI. So yeah, that’s a fun battle to fight. Right? But we are. We are the engine that can, and I say can, because I don’t like to say could. But the other piece is, really, I like to say it’s AI, not API. I’m actually writing a post about it. A lot of the solutions that are out there, they’re not building them themselves.
They’re using third-party API, they’re using publicly available libraries. And it’s basically like their software is just the fact that they can write interface code that looks at IBM Watson’s API. And then they might do a creative thing to display it on a screen, but they really don’t know how it works or the thing that they are building works. And they’re beholdened to that relationship, that API relationship. One thing that really differentiates MarketMuse is we built all of this ourselves. We own it all. And we have one of the largest databases of topic data, our data warehouse is massive. Our topic modeling technology. We’ve been building topic models for over six years. We really have correlative associations with success. And what we’re able to do is, this isn’t a trick. We’re not telling you that it’s a quick win all the time.
We’re telling you an accurate level setting, expectations setting data set. And what will you see in the market unfortunately, is a lot of people who are saying things that aren’t true and a lot of opinions, quick kinds of hacks and tricks. And a lot of them steer people the wrong way. One thing that really differentiates MarketMuse from the other content optimization or search solutions is we’re thinking about concepts and topics, expertise, quality, and comprehensiveness.
We’re not trying to game a system. Because when you do that, the pain is coming. It may feel good. And you may have an adrenaline rush because you game the system one time. But if you are working with a business, you can’t afford that to go away. These affiliate pouts, I like to call them. They say, ah, here’s the trick to get your affiliate site to rank.
But then when it crashes, what do they do? They just throw it away and go build another one. You can’t do that. If it’s your agency, you can’t do that. If it’s your client, you can’t do that. If you’re a major business, what are you going to say? Oh, well, it’s dead. Let’s go change our brand. That’s not how it works. And so this brazen, ridiculous, SEO theory that is not backed by anything or the perspective that they have, we’ve tested all of this and it’s super simple. I mean, what differentiates MarketMuse is, we’re talking about real expectation setting, proven processes, and proven workflows. And we’re not trying to trick anybody. On the GPT-3 side, natural language generation is a Wild West horizon right now.
And what you’re seeing is a lot of the software that’s coming out, they also didn’t build anything. They have what’s called prompts, prompt relationships with GPT-3 and GPT-3 is a language model that you can access. And you can set up a prompt saying, Hey, here’s how I’m going to use this access. And then you can hit it like an API almost. And then you’ll show text that it generated. But that’s the extent of the 30 or so software companies that have launched this year, basically just having an API relationship with either Microsoft as a reseller agreement where a direct relationship with open AI and they’re paying basically a server fee and they’re doing creative, really cool things. I’m not debasing them in that respect. But one thing that we do is we build it all ourselves and we’re building solutions for content marketers that truly get down into the weeds of what it means to be an expert.
And we’re improving our natural language generation solution every day as well, it’s not done. And the wacky thing with GPT-3 and natural language generation is that market changes every four months. And it’s like exponential, exponential growth. If you wrote an article about NLG last year, it’s so completely out of date that it would be a joke. And Google’s new net language model is I think five times GPT-3, there will be a GPT-4 I’m sure very, very soon, and it’s only going to get better.
And we’ve seen that with our own work. Four years ago, we bought the first NLG for ourselves. It wasn’t good enough two years ago. We decided to pause only with our most recent iteration, where did we say, wow, we’re going to be able to bring this into the author ring experience. We’re going to be able to bring generation in with guidelines. So you can say, I want an article that’s about this long, covers these topics, fits with this outline, stays within context. Doesn’t lose track, has a long-term memory. And those are things that GPT-3 cannot do, and we can. And so that really differentiates us.
TS: Yeah, that’s incredible. And I think that you can see that play out in some of your case studies. I was reading about how one of your clients went from 4,000 in monthly traffic to 400,000 in monthly traffic in under a year, which is obviously very impressive. So, could you speak to that case study or others? How are you able to achieve such massive results?
JC: I mean, that’s a great example, but I will say, every fun case study, it makes me so happy. I mean, MarketMuse is it’s not telling you it’s going to be easy. It’s telling you how much work you have to do. And there’s a really big nuance and it’s because we actually care. I care about everyone. I mean, probably to a fault. I mean, I’ll get a Facebook message at like 11:00 PM. And sometimes I’ll answer it. Because I really care about these sites doing well. And in content, the most simplistic view is how many items am I creating?
How many items am I updating and expanding? And how were those things woven together? And so the example you’re using, a big piece of that was updating and expanding existing content and how you prioritize that. And in that case, we evaluated an extremely significant creation motion as well, that for the gaps that they had, really, what would the cluster of content be that would tell the story that they’re an expert on those topics? That’s the answer that we’re giving every time we’re working with somebody and what do you need to update that you currently have? What do you need to write that you don’t?
When you have that entire kind of payload or collection of clusters and publish them, it puts you in the best possible place you could be today, right? And then all of the off-page factors, all the link factors, all the authoritativeness calculation then does its work. And we’re giving you the best chance you can with the on-page [factors]. And in this particular case, and in hundreds like it really, it was just a skyrocket for them. I mean, it was situations where we found low-hanging fruit, quick updates that did lead to wins. We found large clusters that needed to be built and we’re able to predict very, very well in that dynamic. I think we’ve expected 600,000. Just to be blunt. And 400,000 was pretty good.
TS: Pretty damn good.
JC: And there’s a concept in there that people often will mistake, and it’s called the—I referenced it as a term pool multiplier—and that sounds really esoteric. But it’s really the reason why search volume, which is the number of monthly searches that Google reports via their AdWords platform for broad match plus is not a great North star. Because if you look at the predictability of the actual traffic you generate based on that data, it’s just completely unpredictable. And when you have authority on your site and you write a great article targeting a term that has significant volume, you also start to rank for things you can’t predict. So these unique queries that have never been written before, there’s a big pool of those. So what we look to do is measure based on your authority, the likelihood of you capturing that entire pool of specialized unique non-replicable queries with your pages. And that’s what writing great semantically comprehensive content does. And so in this explicit example, and like I mentioned, in hundreds like it, that’s where a lot of the wins happen in the margins, too.
TS: Cool, that’s very impressive. And so I’d like to flip the conversation away from your customers, and let’s look at MarketMuse in your own marketing for yourself. Can you tell us what have you found to be highly effective in marketing for your business?
JC: Now? So that’s a great question. So prior to transitioning to Chief Strategy Officer, I was Chief Product Officer, but I’ve also managed the marketing group. So I was like lead the marketing team, the product team, the data science and the engineering teams, and one thing we have always had is, we eat our own dog food. I hate that reference, but we do. We love using MarketMuse to be successful with our content. We are all content strategists on that side of the business. My lead content strategist has been Stephen Jeske. I’ve sent him so much stuff that we should write, and it’s his job to prioritize it. And one of the ways that he does that is with MarketMuse. I mean, it’s been very, very successful for us.
Our content marketing driven leads are a dramatic thing. We focused on our credibility and our communities—we have a Slack community of content strategists that’s over 1500 people.
We are consistent with the work that we do with our own webinars series with partnerships. And we actually have, we call it, we have 13 channels we manage. With a very, very small team of marketers of inbound leads. The biggest challenge that we’ve had… And I’m very honest sometimes to a fault. The biggest challenge we’ve had is for the longest time. So we were getting over 3,000 top of funnel leads every month. And we’re only converting a very small number of those because our product was only for mid-market to small enterprise to enterprise. Our entry point was an average deal size over $24,000 a year for some time. And it took us a long time to figure out that there was a great market for some lower priced offerings for smaller teams.
And, so one thing we did was we started to build to self-service and not just for sales-led growth, with focus on transitioning to having both sales-led motions, as well as product-led motions. So the best thing we’ve been able to do has been to consistently generate the buzz and the interest of the top line, and then adapt based on funnel conversion rates, based on our ability to service clients, and then get better at self service led onboarding so that we can offer a $150 a month offering.
And that has been the journey for us. It has not been easy. I mean, getting to the point where somebody can walk into the door, buy something for $149 a month and have a successful experience when it’s as complicated as this stuff that we’re talking about is super hard. And that’s a journey we get better at every day.
TS: Excellent. Thanks for that. So, I’d like to ask you to look into your crystal ball. Where do you see content marketing technology and SEO technology going in the coming years?
JC: On the SEO side, I’ll approach that last. I think [content] generation is going to become a way that true subject matter experts become unbreakable, incredible, Hulk-level monsters. I mean, a great team of editors. The value that they have and will have over the next few years has gone up 100 X. And there’s no joke there, no BS, this isn’t taking away their jobs, but what it is, is it’s collapsing the market on low quality content.
So, writing farms, low expertise content, no expertise content, onesie-twosie penny content per word goes away. And because you can generate better content than that right now. So why would there ever be a market for those people to write bad content?
One of the jokes I always say is, we want to set the standard for content quality. That’s our tagline. But the other side is we want to rid the world of bad content. And so I’m excited. I’m really excited about the fact that the world of bad content is starting to disappear. And my crystal ball shows that in a few years, there’s no market for it. And you’re already seeing that happen. The large writing networks are scurrying around trying to create managed services, trying to integrate technology, trying to bypass. And so that’s already in motion.
The other side is accessibility to amazing solutions. There’s $3 billion companies now in SEO, there never was before. All the enterprise SEO solutions of the 2000s have kind of hit their cap. They were structured wrong.
They’re built for enterprise, old-style sales led growth, but then everyone’s realizing that these other types of solutions can become, you may be not want to call them unicorns, but they can become wonderful companies. And so VC investment into search and VC investment into content is going to come back when the likes of SimilarWeb, SEMrush, Ahrefs and their stories play out. Because those are all great companies. Two of them will be public this year. One of them is just a bad-ass. Hopefully I can say bad-ass. And that will happen really abruptly and people who weren’t watching this pulse from a VC perspective, from an investment perspective, are going to go, “Oh, wow, I didn’t even know there was $2 billion public companies in this space. I’ve been maligning it for years.” And so that’s going to change.
That’s going to change the market. And then the last thing I’ll mention is sadly, the touts are out. They are still there. They are on YouTube. They are saying it’s easy. The snake oil exists still. And it’s a thing that we just have to keep that beating down with large sticks until it’s hopefully quiet. Because if you think this is a quick win and you walk into this world and you buy solutions that are cut rate, low quality, and you do it, you have success for a few months and then you crash. And you’re like, what am I supposed to do? Or you throw your toys in the air and give up. And that’s really where I see the market going.
I still see some major mistake diagnosis, major mistake communications happening. And sadly, that’s where I see part of the market going to, it’s that free or $39.99 thing that it looks and it feels good, it feels real good. To the untrained eye it looks the same. So then you use it and you’re like, “Hey, I did this thing”. And then it doesn’t work over time. And so if you want longevity, the technology marketing, search technology marketing content, they’re powerful vendors with backing of real technology. And then there’s stuff that is backed by affiliates.
TS: Jeff, if someone wants to learn more about MarketMuse, where should they go? What should they do?
JC: They can email me email@example.com. You can go to the site. We have a trial experience. It’s not full function, but it is a pretty great representation of what we do. I do two content strategy webinars typically per month. And I have an archive of over 50 of those. You go type in Jeff Coyle webinars, you see a pile of other ones and connections there. I’m on Twitter, I’m on Clubhouse a lot. Hit me up on Clubhouse. If you want an invite, I’ve got like 65 of them. And also on LinkedIn, any way to get in touch with me is fine. I’ll give you my cell phone, if I could. I actually care that much.
TS: Awesome. Well, Jeff, thanks so much for joining us today. MarketMuse is fantastic, and I really do believe that it is the future of where content marketing is going. And it’s been awesome having you on the show. So, thank you.
JC: Thanks so much, Tom. I appreciate it. And I’m really excited to see what’s in store for Stratabeat as well. I’m really pumped about the things that you were talking to me right before the show. And I absolutely recommend everyone give a look at the types of things that you’re doing. It’s really amazing, novel, innovative all in one.
TS: Awesome. Thank you.