AI For Content Marketing With Jeff Coyle
When it comes to sales, marketing, and content creation, no one can beat AI. Imagine trying to write ten articles a day and only two get approved. You're research, planning, and prioritization will be all over the place. Jeff Coyle and his company MarketMuse have found a way to turn humans into superhumans (or at the very least, super content creators) with the help of AI. Jeff is the co-founder and Chief Strategy Officer of MarketMuse. He helps his clients with market insight and content planning with the use of advanced artificial intelligence. Chad Burmeister brings him to the show to talk about the use of AI in content marketing. Learn just how far the gap between a human and an AI can go and what happens when they work together.
Listen to the podcast here:
AI For Content Marketing With Jeff Coyle
I've got a special guest with me. Jeff Coyle is the Cofounder and prior customer in fact of a company called MarketMuse. When you talk about artificial intelligence being used in the sales and marketing motion, there are not a lot of companies out there that use it to the level that Jeff's company does. We are going to dig into that. Before we get started, Jeff, welcome to the show.
Thank you so much. I appreciate it. It's going to be a fun discussion.
I'm excited. We wrote a book called AI for Sales and interviewed 21 different leaders. It all starts with data, and then there are all different chapters on how different companies are using technology. Now, we are close to 100 shows in and I continually find and uncover new ways that people use AI in their motion.
I'm passionate about lead management, lead nurturing and lead qualification. I have built products around it. I have been in this space for many years, as scary as that sounds, whether it be a content strategy but also lead management platforms. Some people might not know that. We also use some technology in-house to manage our own processes. I have sold actively for MarketMuse. My mentor, manager and previous organizations were one of the visionaries of sales from a consultative sales perspective. We abide by that value-selling, consultative-selling approach. I appreciate everything that you are working on. It's the summary to say that. I'm the Cofounder and Chief Strategy Officer for MarketMuse. As I mentioned, I have been doing everything related to content strategy, lead management, search engine optimization and conversion rate optimization. I appreciate it and look forward to this discussion.
With that, our audience gets to know who you are and how you came about being in the role you are in. To go back to when you are younger, what was your passion? You are 6, 7, 10 or somewhere at a young age. What did you get up every day and say, "I love doing this?"
I had a 386 Step B1 taken apart on top of my brother's bed with fans shooting on it because it would overheat. I was doing that type of stuff as a young person. That was the middle school passion for me. When I went to school, I went to Georgia Tech for Computer Science. I knew that there were things, user interface, usability, and then the earliest forms of text search. I went to go work at a company called KnowledgeStorm, which was the first people who were selling leads to software companies before software companies even had content. We were convincing them to like, "Give us their brochures and we would syndicate them. Do you have any white papers? Let's get them online and we will market them, and then we will sell you those leads that we generate." We were convincing large businesses to write content, and then now it would be a joke to say.
On that team, I focused on everything from managing our taxonomy to content marketing and our own search engine paid strategy. Early on in my career in 1999, 2000, I identified that search engines are a thing. They were going to be important. I couldn't possibly imagine they would be where they are right now but that has been my passion to understand how one can exhibit one's own expertise or their business's expertise and how you use that to achieve a common business KPI. The biggest one I focused on for the large part of my career was lead generation, but then that has morphed into content as I worked with a company called TechTarget. It's one of the largest lead gen sources. It's a huge publishing network. KnowledgeStorm was acquired by them in 2007. I led their in-house team for almost eight years before leaving to ramp and launch MarketMuse.
During that experience at TechTarget, I’ve got my chops working with amazing editorial team members, writers, content strategists and subject-matter experts. We were coming at them with data early on. They were like, "Step back. We are the artists here. We know what we are talking about. We are the experts." In that time there, I tuned how to use data to my advantage and say, "You wrote the best article in the world about that concept. Here are some other things that if we covered, would do well. Do those make sense to you?" versus, "Go write this article on this keyword." That doesn't work. The other sentiment works.
What you perceive mentally as the cost of content isn't the true cost of content for you.
With marketing, it's taking all those things that were manual. My subjective perspective on content quality or what we should write or what we should update, automating those things and driving them with data, "Here's your advantage. Go update this page. I value your expertise. That's why I'm asking you to update this page. I need you to make it better. I need you to make it more focused on what it means to be an expert. Here's how to do it." That has always been my passion. It's to get somebody from subjective brainstorming as their norm or planning a calendar twelve months to do it confidently with data that they can use to level their own expertise up. The best people who use what we build are already subject-matter experts because they are becoming a superhero status. When I see that happening, that light bulb clicks. It's like, "I built something that I couldn't have done on my own and it's even better, even though I am the thought leader in this space."
Tell us where artificial intelligence fits into the process for a marketer because there's obviously, "What do you write about? How do you write about it?" It’s the length and the words. I have tried something where the AI writes the article. It was pretty darn good. It was better than what I would have done in a few cases. Tell us about where you see marketers leveraging AI? What should they be looking at?
In so many different regards but specifically for the content strategy side, it's about research, planning and prioritization. It's the first stage. That's the highest value action you can take. It's to say, "What should I do now? What should I prioritize?" I can come up with a terrible idea of something to write and then go through the process of writing it. It can be the best thing in the world. It can be the best article ever written about that but that wasn't the step I probably should have taken to start with. It may not be formed to expectations. I may need to write 25 articles about this. It was a good idea but I didn't understand the weight or investment that I needed to make to start to move the needle.
What happens so commonly is teams will go, "I need to write about this topic." They will go write one a couple of thousand-word article about it. They look at their watch and they were like, "That SEO told me to do that. Maybe they are not so smart but the reality was I needed to make a huge investment in that topic and build piles of content about it that tell the story that I'm not only smart generally but smart specifically. Our technology will give you the insight as to the level of investment, as well as how to take it to the last mile and execute." That's a big difference. That's where the real high-value actions get made. They get taken.
The other ones, not that they are any less value, there are just a lot more of them are the steps you take when you are writing. It's to say, "If I'm going to go cover content marketing strategy, I better include stuff about the target audience and buyer personas or else I'm not an expert." We naturally have biases in our brain based on our own subject-matter expertise. When we write, we will often have blind spots or biases. What artificial intelligence technology can do is a branch of topic modeling but it tells the story and says, "What does it mean to be an expert on this?"
That's our core technology. We have two patents on it. It goes out and reads everything that exists on a particular topic. It becomes the expert, and then it tells you what an expert would do effectively on that concept. We can do that at the page level but we are also able to do it at the site level. We can say, "How often do you typically cover that well? How often do you cover semantically related things? Is it natural that you wrote that wonderful article about this thing? Does it fit into this narrative that you have created? Was this a one-off random concept?" It was still good. It still should have some merit but you may need to build a foundation around it for it to truly shine.
We see this so often. I even had a customer who had two product lines and they weren't the same topic. They were like, "We want to grow each one of these business units by 30% next year in traffic and leads." I said, "To do it with this one, you need to write about 48 articles and update these 15. This other one, you’ve got to write about 250 of them. You’ve got a lot of work to do." They ended up pushing the other one off three years down the road and they doubled down on the small one. That happens all the time. It's like, "Product A gets ten articles this month and product B gets ten on this topic.” No one is updating in typical B2B technology. Nobody goes back and updates their stuff, which is the hugest mistake anyone could ever make, not even to link to your newer articles. It's generalized.
You asked about AI. At the page level, it's analyzing topics and telling you how well they cover that topic. At the site level, it's understanding your own authority where you have a competitive advantage. All are branches of natural language processing that can tell you what it means to be an expert. We also work in natural language generation. Our newest tech product and technology is called MarketMuse First Draft. It is built completely in-house. It is not using Open Source Microsoft or OpenAI technology. You can construct content drafts based on our other technologies. Think about what an advantage we have. We can build the outline. We can tell you what it means to be an expert? Our natural language generation platform will write like that expert. We can make it write like you. We can say, "Write it as Chad would." That's the special sauce right there.
Content Marketing AI: AI can make people be super-powered. It can make them feel good about their day. They'll be able to level up quickly and that process will be efficient. They will feel like they're experts, too.
We are not putting it out there so that you don't have to update it. We are putting it out there so that you, as a subject-matter expert, will understand the context. It will take you, hopefully, a little bit less time to put something out there that's good. I wrote one, "I'm not a good writer." It's a fun punchline at MarketMuse. I had a draft that I worked on. It took me about 2.5 hours to do it but the punchline was, "I would have never done it. I did not pay the bill for it." There are a lot of people that are in that spot, too. They know that they’ve got to write twenty articles this month. They only ever knock out four. They should do twenty. We can get them to twenty.
It reminds me I'm working with a gentleman to write a kid's book. I didn't have a concept of a kid's book. We were at this executive retreat together and Timmy tells me, "That's what I do for a living." I was like, "Let's do it." He was giving me the outline, how to say it and how a six-year-old would want to read, how would my audience want to read it. You can inject your philosophy and ideas into it but it has to go through the Timmy Bauer filter for it to be relevant to a six-year-old.
That very much reminds me of what you can do because you give them the outline, and then you give them First Draft and now, "There's still human required because I do want your secret sauce on top, the header. I want quotes in there that might be relevant from a CEO at your company, product expert or customer." My big question here is thinking as I'm the Founder of a company that did $1 million last year and this year, maybe we will do $2.5 million or $3 million. "Can I afford a technology like this?" When I hear 24, 48, or 200 articles, I'm thinking, "There are a lot of people required on the back end of that." What do customers generally have in revenue to be able to afford the technology like this?
I will tell you what they typically have. They have what's called a content efficiency rate. If you are punching 100% now, you are good. If you are punching 10%, which is about what we typically will see, you’ve got work to do you. What you perceive mentally as the cost of content isn't the true cost of content for you. If I write ten articles and only one achieves its goals or KPIs, the true cost of content for me is 10X. I go to my favorite content network where I can get content for $0.10 a word, $100 but only 1 out of 10 does one. That's a $1,000 content item but it's much worse because of the opportunity cost of having not published something successful and the potential competitive risks that it creates.
When we are talking to people investing in content, what I'm looking for is, "What's your current efficiency rate?" There are a few tricks of the trade. If I can tell you a few things you can update and they are going to move the needle quickly, those are beautiful and then you are hooked. You can invest in technology here. You can invest $6,000 or $12,000. We have people spending $500,000 for those who are writing thousands of articles a month with our technology. It's going to be based on how much time you are going to put in, do it yourself or how much do you want to be done for you. If you are building more than a couple of pages a month, it's absolutely a fit to do this.
The cool thing is when you look at and you are like, "My situation is, there are so many things I want to write and know that I will win at. I don't have enough ability to get that stuff done." That's where a lot of people are. They've got this huge list of stuff they want. They don't know how to prioritize it, and then executing on it is very hard, "Who do I deal with? Do I go with a third-party outsourced writer that I'm not confident in? Do I wait until I get time to do it?" We are trying to solve your situation no matter what it is.
What I'm hearing is very interesting because I talked to a guy who created this website called TextChat.com. He ran call-center chat operations for years. He was frustrated because the reps wouldn't do it the way he would do it. What he did is he figured out a capability that when someone is on the website, it will text him as the founder and 2 or 3 other people and then he interacts with the chat. If he picks it up, now he's on the hot seat. He's like, "I’ve got five minutes. I'm going to handle the chat now." It puts the onus on the founder, which is the best tip of the spear entry conversation you can have. What this conversation is making me understand is that it's the same thing when it comes to content. I would rather have your best people, the CMO, the Head of Products who has been doing this for 30 years augmented by AI, rather than have a team of entry-level college grads that have no idea what it is they are talking about.
I want those people to be super-powered. I want their day to feel good. If they do get drafts from someone who's writing them, they will be able to level them up quickly and that process will be efficient. I want the writers to feel like they are experts, too. What's the worst thing you can ask an editor or a subject-matter expert to do? SEO. They don't want to do that. They are an artist. They spent their life learning about security technology. They spent their life learning how to write well. The last thing you want them to do is to know the nuance between external linking through semantically adjacent topics versus adjacent topics.
When you're writing, always start by building a foundation.
To your point, going back to the prior posts where that can be resurfaced by your application, I can see it now. It just gives you a notification and there we go, "Here are our three recommendations for this article. You need to backlink it to the newest one, push the button and execute."
The feeling you get if you don't have it now if you are using MarketMuse is painful urgency to do stuff you can't do. For example, on MarketMuse, there is a post that we have to write. It's sitting out there. It's a true thing. I have some sites that I also manage on the side and stuff like that. I always know the next article I should write and why. I always know where I have had unsuccessful content and I know why. It's because I went for the shot and didn't build the foundation around it or I've got this topic, crushed it on this one thing and there is a variant now that I can go right to.
An example, we own and build content briefs. If you go type in content brief into Google, it's marketing. We haven't written an article that covers all of the different types of content briefs and templates and gives examples of templates. It’s because of marketing, I know that's the article we should write next. That's special. That's like building out a strategy for me. It tells me, "These are my outside competitive advantages. These are places where I have a chance to win." You mentioned, "What should a team do if they are under-sourced?" The team that's under-sourced should make sure their mirror is clear so that when they decide what to create and update, it's not rolling dice. It's going to be successful. That's how you, as a small team at 1.5, 3, 5 or whatever can use your money and time wisely.
There are surface questions that get served that get 1,000 a day and there are some that get three. I would choose the audience with three unless they are spending $10 million each time for the three.
It's not just about volume, though. It's about, if you were to answer all those questions, it would show an exhibit that you are the expert, and then all your boats rise. It's not just about direct gratification. You’ve got to write them all. That's the thing. There's an article called Keyword Research and The Search Volume Illusion. It gets into so many different aspects of this because some of these searches, too, we call them Know-simple Searches. It's a simple answer versus a Know-intent. Know-simple means, "What's the capital of North Dakota? Bismarck." There's only one intent. Do I need a long-form guide to North Dakota to answer that question? No, but then when I searched for North Dakota, I may get a good cornucopia of possible articles because that's called fractured intent. It doesn't know what I want.
If they type in, "What's the capital of North Dakota?" You better believe that is going to be an almost zero-click search. The key there, zero-click search. Very few people are clicking on results. They are getting information and leaving before they do. That's a new phenomenon. In your case, "Can I write these answers to these common questions? Also, maybe have a guide to the connective tissue of those questions, what weaves them together, and then be the expert on all the things so that I put myself out there and for as many of those searches and concepts so I'm going to get as many clicks that exist?" This one question might be, "What's the hub size for a '68 Pinto Hatchback? Eighteen." I just made that up. I may be part of the Classic Ford section of my site.
We have been talking a lot about the use of AI in marketing and content. If we pivot about sales, you mentioned you are on the sales side at times in the business. We are all selling something every day. Have you seen any AI technologies in sales that are piquing your attention? I think you've got an interesting appreciation for the use of AI. You might see something and go, "That scales in an interesting way."
There are so many. I always like to say, "We are stacked so forward for the maturity of our company because we are all software people and problem people." We see the drought and the good in stuff. We are probably stacked for our company two years in our senior with technology. I'm passionate about lead management, my dislike of historical lead qualification, the "best practices." How Machine Learning, Data Science and artificial intelligence, I feel one of the biggest things it can do is take user data from your site, product and customer success platforms. It can take all those demos and psychographic information. It can piece them together to say, whether something truly is a predictive good fit for your business or not.
Content Marketing AI: The team that's under-sourced should make sure their mirror is clear so that when they make a decision about what to create and update, it's not rolling dice.
In automated lead scoring or predictive lead scoring, that's my favorite field for sales with AI. There's this company, one of my favorites is called MadKudu. If I can buy stock in MadKudu, back up the truck. Go get them. MadKudu is the coolest thing. Things like that excite me where it's like, "It's not something your brain can do." That's what key. That's my differentiator. My brain can't do it. My brain wants to associate if-then statements, "If the person has a revenue of $100 million or more, then they are a good fit." It's wrong in almost every case.
It's funny because we had our forecast call. I went through one of the reps' pipelines and you sort top-down because the biggest deals can move the needle the most. There were three $25,000 sitting there. I said, "Let's go look them up on Crunchbase. Are they funded?" "I don't know." "They have been around since '92. They are not funded and they are flat to negative growth. I think they can stroke a check for $25,000." He was like, "Good question. I didn't even know if they were funded." "How many sales calls have you done?" "Three." He wasted three hours of company time on a prospect who could never buy and that's just one input.
When you start putting in what's their title and product-market fit, there are so much that AI can do. I remember Kevin Haddad was a Manager of mine at WebEx. His thing was, "If I could just see the title of the person in my sort in Pivotal CRM back in the day." I was like, "You can." I added that as a field and he was like, "You just turned the lights on in the darkroom." Again, one data point. Now, you multiply that out by 100 data points. You don't need more than just a handful.
How about all the opportunities in your CRM, their speed and velocity to close and conversion rate for renewal, the account manager time spent on those accounts? Now, connect that back to their demos. If someone is on my website, I know, whether they are likely to be a fit for X product versus Y product or they are likely to be a fit for a premium offering. Which sales reps should get that lead based on their historical performance? That's what AI does for sales. It can distill all that in a number. I remember building my own models for qual, qualifications. I hate BANT. I'm all about that. I read a great article and it was anti-BANT.
It was probably by Keenan. He's Mr. Anti-BANT.
They are value-selling those folks, "I'm a Bob Apollo." If I can buy a shirt with Bob Apollo on it, I would have it. He's from Inflexion-Point. The characteristics of somebody who can buy too, it's my role. It's also familiar buyers and unfamiliar buyers. Have they ever bought a software solution for this amount? That's a painful thing to have to overcome. I'm selling to SEOs to content marketers. Some SEOs have never bought a product that's more than a couple of hundred dollars. You’ve got to get ready for that journey because that's a hard one to overcome. How do I prep my team so that, "This may be the first vehicle they are to buy?" That's hard work versus, "I am a CMO. I have a MarTech Stack. I own a part of that. I have a stack of these things. I have migrated from one stack to another. I know how much our Salesforce implementation costs and it keeps going up every month. I'm comfortable that if I needed to invest in something that costs $30,000, that I could."
There's technology out there similar to MadKudu that you mentioned. I'm going to look into it. Maybe I need to say, "I would like to put a $100,000 investment into you. Meet me and we are both going to do it."
I am all about predictive lead scoring because it has made me believe it can be done. I always thought it could be done. In all of my versions, I would always have a better model internally than the software could create. I always had the, "Jeff's lead score," and I'm going to call out names but all the other platforms. Mine was always more predictive. For the third time, I’ve got one of those better. I was like, "This is the answer because now I don't have to do anything and it's right here. I love that." Nothing kills me more than the first one call qual sequences where somebody is, "Do you have an active project?" "Okay, cool. Click DQ." Do you know what that means? That means you are buying your leads because you are judging.
The best people in the field who use AI are already subject-matter experts because they're becoming superheroes.
The human mind naturally judges something they buy in a package. If you get a thing and it's a whole package, you say good or bad. It's just the way that you think. If you are buying leads, you tend to judge them. That is a super-duper problem with leads, content and everything because you buy it from a writer and you are like, "Is this good or bad?" What you should be asking yourself was, "Did this give me any value that I can use to my advantage? How much value did it use?" I can get this lead. My qual person calls them. It’s not an inactive project. I can either not think that somebody at Boeing wants my software or I can think, "Someone was researching content strategy at Boeing. There's probably a way in."
Others are probably doing the same research. There's another technology that I would recommend you guys to look at that might be an interesting piece of generating the content, Codebreaker Technologies. They have a thing called a BANKCODE. It will read LinkedIn. In under a few seconds, it will spit out the order of who they are as a person. It's almost DISC and those kinds of things. This is a little bit more new-age than that. B is Blueprint, A is for Action, N is for Nurture and K is for Knowledge. My son is at the School of Mines as an Engineer. He's not going to be activated first. He's going to be blueprint first, "I need to do A, then B, then C." If you could serve up the content to the specific buyers based on their BANKCODE or DISC code, then it turns out you can.
Automatic predictive lead score intent and buyer journey mapping is the other field that I believe is the spot. If I can say, "You are at this stage in the buying cycle and you see the things that get you to the thing," it's not there yet. The only hard part about it is no one will understand it. The people building that SaaS platform will be like, "This is the best thing."
For years, we have tried to talk to salespeople about how to have a conversation with a certain type of buyer and it's very hard. Maybe I have a little bit of that in my unconscious competence to speak slower, ask questions or bullets, but I doubt it. I'm probably who I am regardless.
I hate this qualifying like, "It's okay." I want the person to have a journey and I want to facilitate that journey, whether it is no-touch navigate through this content discovery so that I am ready when I am appropriate or it is, "This is a person with a million-dollar check in their hand from Boeing." There should be a place for them and that's your son at the College of Mines. Congratulations, if that's a true story, by the way.
He made the Dean's List in his very first semester. That was a true story, too.
That is a place. That is impressive. I love the BANK, the Blueprint, Action, Nurture and Knowledge are cool. Whether it's care, consider, choose and you are rocking an IBM or it's awareness consideration post-purchase troubleshooting, adherence post-docs or content marketing team, know after the deal gets sold, they are not involved with any of the content. Your post-purchase adherence content, post-purchase troubleshooting content and post-purchase champion development content are as important as your pre-purchase content. Nobody is doing that.
When I look at where most of our customers come from, it's word of mouth. If I could nurture that in an automated fashion, look out world, you could 10X your company just by focusing on that.
It's a problem we have tremendously because we are short-staffed. Everybody is. One thing we found out, which was cool was when a person who works at a mid-market company leaves that company, they will typically bring MarketMuse with them. We had a gentlemen come from one sales enablement software company and go to a major credit card company. He immediately brought that team with us. “Where did that lead come from?” “Champion referral.” It was actually Champion's write a lot.
Content Marketing AI: When people write, they'll often have blind spots or biases. What artificial intelligence technology can do is a branch of topic modeling. It tells the story and says, "What does it mean to be an expert on this?"
There's a program called Job Hopper Program. We would monitor it, so then when I joined RingCentral, we brought in the Job Hopper Program there and said, "How many people that are CIOs or VPs of IT with call center people?" They leave one and go to the next. If we could be their first with a call and an email, we are miles ahead of what anybody else can put them in.
It has happened so many times. I can't even count it. We haven't begun the lightning in the bottle to figure out exactly how to turn that into always happening. You can't ignore it. Those are instant wins for us and all kinds of different challenges. We've got problems.
Jeff, I appreciate you coming to the show. Thanks, everyone, for reading. We are talking about AI in the marketing realm but also in lead qual and sales. It's here. If you are not looking into the two-letter word called AI, you are probably missing the boat. I'm finding it’s built into products. When I ask someone, "Who uses AI?" in an audience of 300, five people raised their hand. Maybe it's twenty by now. If you say, "Who's using ZoomInfo for data?" 50 people raised their hands. It's like, "You are using AI."
My new phrase is, "AI versus API." I'm trying to trademark that one. Is the company you are working with building its own analogy? Are they just using someone else's technology to be an API? There's a big nuance there because of longevity.
I outsource my heavy lifting to the same company that everyone else is using. How's that a competitive differentiator?
We build everything in-house. Natural-language generation is quite a David versus Goliath situation because it's us versus OpenAI and Microsoft but we are going for it. It's fun.
Thanks for joining, everybody. Jeff, I appreciate your time. This has been a great conversation. We will catch you on the next episode.
Thanks so much. Cheers.
Important Links:
Keyword Research and The Search Volume Illusion – Article from MarketMuse
Kevin Haddad - LinkedIn
About Jeff Coyle
Jeff Coyle is a co-founder at MarketMuse. Previously, Jeff was VP of Search at TechTarget (NASDAQ: TTGT) where he built a 30-person content optimization team. He joined TechTarget through their acquisition of KnowledgeStorm. Jeff is a sought-after speaker at conferences and has also advised Accel-KKR portfolio companies on lead generation.
