Using AI to Create Long-Game Scalable Content . . . FAST


Aki Balogh, Co-Founder and CEO of MarketMuse, a company that works with companies and agencies to transform content marketing into performance marketing, speaks on a topic that is relevant to both agencies and companies – content – and talks about the benefits of long-game strategy vs. the short-term transactional “push.”

Every company needs to create content to build product and brand awareness, drive demand and lead generation, and engage their audience. The problem with content marketing is that the research process can be tedious, the writing process unwieldy . . . and the results can be unpredictable, unscalable, ineffective, and inconsistent—everything is done by hand. Additionally, many companies write generic content, miss the opportunity to specifically answer people’s questions, and fail to describe the “sandbox” they play in.

MarketMuse utilizes a proprietary intelligence and strategy technology to compress the research phase by downloading a million relevant articles from the web and crunching the data to provide a metadata blueprint of how to cover the topic comprehensively, what questions to answer, what subtopics to develop, and what recommended resource links to include.

This work is done at scale, so every article the client writes generates a blueprint and a spec – a foundation that lets the client scale the company’s content channel as effectively as it can scale the ad tech channels and the paid acquisition channels.

MarketMuse’s focus, for the most part, is on long-form (1,000 to 2,000 words) top of the funnel and middle of the funnel informational or comparison content. Aki feels it is important to develop that top of the funnel content first – before talking about the specifics of your company and your product – and to use a branched hub and spoke model—one detailed long-form pillar content item with subsections and then links from those. Links provide content that is shorter but way more targeted. The objective is to provide logically-structured content for every topic (from 2 to 20.000) relevant to the client’s company for every stage of the buyer’s journey.

MarketMuse’s software is also able to do shorter form content at scale. Additionally, MarketMuse’s self-service SaaS platform provides Customer-side strategists at small clients the ability for to perform website content inventory and build content strategy. These smaller clients may be writing 5 to 10 articles a month . . . instead of 250.

How does MarketMuse convince potential clients to go for the long-term win?

In looking at a website on a first contact phone call, MarketMuse’s consultants can show initial or first value within minutes. Even in just looking at a site, MarketMuse experts can identify areas to improve rankings or traffic, then create a content brief, and send it to the potential client. The solution is so data and science-driven that MarketMuse is able to predict the ROI companies can expect to make if they implement recommended website changes. Aki says that their AI-supported process accelerates content analysis and planning very effectively . . . results typically show within the first 3 months.

Aki can be reached by email at Visit the company’s website at, to see the product, the free preview version of the brief and posted blog articles and case studies.

ROB: Welcome to the Marketing Agency Leadership Podcast. I’m your host, Rob Kischuk, and I am joined today by Aki Balogh, Co-Founder and CEO of MarketMuse headquartered in Boston, Massachusetts, and Aki is based in New York. Welcome to the podcast, Aki.

AKI: Thanks for having me.

ROB: Great to have you here. This is a little bit of a different treat for our audience, because rather than an agency, Aki, you have a lot of depth in a topic that is super relevant to both brands and agencies. Why don’t you start off by telling us about MarketMuse and what MarketMuse does?

AKI: Yeah, absolutely. The problem we’re trying to solve is we’re trying to bring performance marketing to content marketing, essentially. Every company needs to write content about their products and services. They use it to build awareness, to drive demand and lead gen, to show up on Google search, to engage their audience. Every company needs to write content, but how do you do that effectively and at scale?

The problem with content marketing is that it’s basically all done by hand. It’s basically brainstorming. If you’re going to write an article on marketing software tools, first you google it, you’ll read a couple of articles, and you’ll piece it together. You might interview product managers, you might interview experts in the field, but you have to piece it together. You’re not quite sure that your article is going to be successful. You’ve put a lot of time and work into it, but you don’t know if it’ll stand out in search or if it will be successful post-publishing.

What we do is give you that peace of mind by basically compressing that research phase. What our technology does is we’ll go to the web, we’ll download a million articles on that topic and compress it and give you a blueprint that shows you how to write to cover that topic comprehensively, what questions you should answer, what subtopics you should dig into, what pages you should link to that are semantically or topically relevant.

We basically can do that at scale, so for every single article that you write, you generate a blueprint and you have that spec, that content spec to work off of. That ends up letting you scale your content channel as effectively as you can scale your ad tech channels and also your paid acquisition channels.

ROB: Very, very interesting. You started this company, you’ve grown it, it looks like you’ve raised some money now and you’ve got some more people on the team. Where did all of this start, though? Where and how did MarketMuse come to be?

AKI: My personal journey, I was – I don’t want to say a software engineer, but I was a computer science major, maybe more of a hacker, if you will, and went to a management consulting firm where we built decision support systems for Fortune 500s. We would look at massive amounts of data and different types of data and put them together in informative ways to let companies understand how to create new revenue channels or how to cut costs or just how to interpret their standing in their markets.

I really got excited about what data can provide, so I moved to Boston and I joined a venture fund, OpenView Venture Partners, where I was the first associate looking at big data and AI investments. I got really excited about what you can actually do with big data technologies, not just competing on analytics, but competing on analytics with big data.

I left OpenView to join a startup where one of my responsibilities was building out content marketing. We made a database product, so we wanted to write articles about how the database worked, how it was different from other databases, what kind of solutions it can provide and what problems it can solve, and customer stories and all that.

I found that we were writing content to do that, and I started thinking, gosh, it’s so hard writing this content. I wonder if we could just analyze articles on the web and build a decision support system for content. So, I sat down and started hacking the code together. It took a couple of years of building the RRR algorithmic platform, and I had a number of people’s help along the way, including my co-founder Jeff Coyle and earlier co-founders and employees.

But now we got to a point where we can be very prescriptive about “take a look at this site; here’s where they’re strong, here’s where they’re weak, here’s where there are areas of improvement, and if you actually make those changes, here’s the ROI that you can expect to see from doing that.” It’s all very scientific and data-driven now.

ROB: Very interesting. Writing content about databases seems like a surefire recipe for writer’s block. What were some things that you figured out eventually to write about that caught on with your audience that you wish MarketMuse could’ve told you to write about?

AKI: The basic problem was I used keyword research tools, which is the standard in the industry, and you go into one of these tools and you put in the word “database,” and the word database gets a million searches a month, let’s say. A tremendous number of searches for just the concept of a database. But that’s a very generic term, and it doesn’t really describe anything. I can’t really take any action on that.

On the other hand, if you look at the term “MPP columnar database,” which is a specific product that we sold, maybe it only gets 100 searches a month, but I want all of those leads. So, I try to write specific content that really digs in deep into the underpinnings and really describes and fleshes out in detail even the technical specifics of it.

That’s what I found. The factor here, one of the most important factors, is how specific a concept is versus how generic. What ends up happening is a lot of companies write generic content, and they miss the opportunity to specifically answer actual questions that people have. MarketMuse can now generate that landscape, but at the time we had to do it by hand.

ROB: You started down this path a little bit – I think people have written content for a lot of different reasons over time. Sometimes it’s been to rank, sometimes it’s been to trend in social. It can really change over time. But what are the core attributes of what makes good content in 2019?

AKI: I have to qualify it by, to your point, saying there are different types of content. We primarily focus on informational content or comparison content. So “What is X? What are the types of X? How do you compare A and B?”

A lot of it ends up being top of the funnel and middle of the funnel content, because we’ve found that companies usually generate quite a bit of bottom of the funnel and conversion content and offers and so on, but they miss the top of the funnel opportunity to describe that broader picture or that sandbox that they play in. I think that’s where we can be really helpful. We primarily focus on long form content. Typically, at least 1,000 words, although it can be a couple thousand words.

You can also do shorter form content at scale with our software, and there are very interesting use cases of that too. But we often find ourselves is we’ll go to a major pharma company, and let’s says there’s a pharma company, they have a drug to treat lupus, but they don’t have any meaningful content describing what lupus is or how to recognize it, what are the symptoms, etc.

Our guidance typically is you should write that top of the funnel content first before you start talking about how great your particular lupus drug is and why it works and all that.

ROB: Interesting. You’re talking about long form content. You’re also talking about playing the long game when it comes to content and not so much the transactional, “Here’s a landing page . . . here’s a squeeze page.” Not even a reductive “Here’s 3 tips on how to run your database better.” This is deeper, potentially thoughtful stuff.

AKI: Yes, absolutely. We’ve found that works. If you look at a lot of the content of startups and aggressive companies where they’re looking to capture a lot of market share, they might not have hundreds of millions of dollars of budget. You end up writing a lot of “The Ultimate Guide to X” or “Buyers’ Guide to Y” or things like that. That technique works really well because people have more sophisticated questions, and they will actually read it.

The prescription is not necessarily long form content in every single case, but the model that we’ve implemented within our solutions is a branching – it’s a hub and spoke model. You have one pillar content item that is a long form, meaty item, and it has a lot of subsections and then links within those. You basically dig into the subsection that’s relevant, you click into those links, and you follow to supporting content items which are shorter but much more targeted.

Think of all the topics relevant to your business – and it could be two, five, 20,000, it could be any number of topics – but across all of those topics, you want to have content for every stage of the buyer’s journey and you want to construct it using this hub and spoke model.

ROB: Right. As a company, you work with brands and you work with agencies, and you work with agencies that work with brands. When you are talking to a brand and someone is maybe skeptical or just don’t get it yet, how do you get them to think about investing for the long term in content, and just believing and trusting the process that the results they want are going to come out the other side, but they’re going to have to be patient and they’re going to have to play the long game? I’m sure many of the recommendations content-wise make sense, but there’s still a matter of trust and time that can be hard in a “pay for an ad and get more eyeballs” world.

AKI: It’s a very good point. We are able to show initial value or first value very quickly. If you get on a sales call with us and we’ve looked at your site, we can show you first value within the first few minutes.

We don’t even need to get on the call, actually. I can just look at your site and identify a couple of areas where you can improve your rankings or improve your traffic, build a content brief on it, send it to you. You can see what specific recommendations we have on that article and then just, in your head, multiply across the articles on your site and all the articles in your editorial calendar and estimate the value.

So, we can do that, building that business case very efficiently. But when it comes to implementation, basically you’ll see results within the first 3 months. Google crawls sites at least once every 2 months. If your site is very small, it might not get crawled as often, but if your site is large it might get crawled daily or multiple times a day, of course. When you make changes to your content, Google detects that the content quality has improved, and you basically start ranking higher for existing topics and you start ranking for more topics, and you can see that dynamic play out within the first 2-3 months.

So, we have always been able to show value and a proof of concept if one is needed, but you can also run that yourself. If you go to, you can put in an article, put a topic to optimize it around, we’ll give you some initial insights. You can make those changes in the content and start to see gains. It’s not quite as instantaneous as Facebook ads, but considering where content is, it takes months, potentially, to do a content audit and to build a content strategy, and then you need writers. We certainly collapse that time to value to no more than 2 months, I would say.

ROB: Got it. I’ve spent some time working on product; you think about product, you think about how to meet the needs of a variety of customers. Sometimes I think it would be helpful for people who are listening to think about some specifics.

Is there may be a specific example and specific recommendation of what it might look like to build better content for the long term for a particular client? What type of product or service is it? Just an example. And then what were some of the immediate opportunities and what were some of the long-term opportunities, if we’re thinking about how this all works together?

AKI: Absolutely. Just to play with that, I’ll continue with that pharma example that I brought up earlier. That’s a real client. They have 12 drugs that we’ve contracted to work with them over the next year, and every month we would build content inventories and build content strategies for different drugs. So, every month we focus on a different drug and we look at what content they have across their properties, where they have strengths and weaknesses and opportunities, and we build them a content plan.

Then for each of those priorities, they turn into essentially a set of new and optimized priorities. Create an optimized priority, so new content to create and existing content to optimize or repurpose. We deliver the plan, and it will basically say “here are the 250 articles you need to write this month in order to build up your standing around this drug.” We give them the briefs, and then their writing team executes it.

That is one real scenario, but all of our larger engagements look like that. The smaller engagements tend to basically be the same except we have a self-service SaaS platform that does the content inventory, and the end user, the content strategist on the customer side, uses the software to pick out and build a content strategy. They might write 5 or 10 articles a month, not 250.

ROB: Let’s say it’s something like a blood pressure medication what are some of the things that you would expect, or not expect, to be suggested as adjacent topics and things to build value over time?

AKI: It could be what are the symptoms if you have low blood pressure? Maybe it’s dizziness or shortness of breath. How can you tell if you need blood pressure medication versus something else? What are the side effects of XYZ? What types of blood pressure medication are there? How does it act? Maybe one influences a particular pathway and the other a different pathway. What are the differences? How do I decide?

It’s describing both the high level of how to even know if I have a blood pressure problem to the prescriptive side of how I choose, and then you can think of creating that content for the patients or the people researching this issue, creating content for the healthcare providers, etc.

ROB: Got it. You started this business – this business has now grown to many more people than yourself, and I think that passion for the company that you have built, as well as that journey of learning how to lead more and differently, is something that you share in common with many of the agency owners that you may talk to and who have found this podcast and listen to it.

What are some things you’ve learned from your experience building MarketMuse that you might do differently if you were starting over from scratch?

AKI: Before I started MarketMuse, I did some research. When I was at this venture fund, I talked to over 700 CEOs in 9 months in order to really get what’s called pattern recognition, get some understanding of what works, what doesn’t work, and what behaviors tend to correlate with success. I found that, for example, open-mindedness in a manager or a leader tends to work quite well, at least in the startup realm.

So, when I built MarketMuse, I tend to tell people I was the first person at MarketMuse working on this problem, but I’m not the only person, and maybe not even necessarily the best person. But I brought in a co-founder, Jeff Coyle, who is an excellent product strategist. He joined 3 years ago, and he’s been here through most of the ups and downs. The first year and a half it was me basically just coding in my apartment quietly, but when Jeff came on it turned from a science project into a business.

Jeff brought a lot in around understanding customer workflows, empathy for the customer, the customer pain points, and that customer focus was very helpful. I give him full credit for bringing that into our company, and that’s just something I learned along this journey.

We brought in a sales leader a few months ago, Chuck Frydenborg, who is excellent. He used to be the head of professional sales at Gartner. He’s installed a lot of discipline and methodical process to our sales efforts and our customer success efforts. That type of energy has been very creative.

What I try to do is essentially learn. For example, at the moment we’re looking to figure out our channel and agency partnership model because we have worked with agencies in the past and we have made them extremely successful, but we have not been able to scale those interactions. We really need to have a model in place that can work where we become a profit center for agencies and not a cost center.

So, we’re building out that model and we’re trying to learn from agencies. I very much am looking to partner with agencies to figure out how to make that go-to-market work where it’s lucrative on both ends. It has been very much a learning journey for me. Every time there’s data that proves one of my assumptions wrong or somebody’s assumption wrong, I really appreciate that, because we’re all just trying to figure out this market together.

ROB: Interesting. It’s interesting, even if you go way back in your background – you mentioned you were at OpenView Venture Partners – OpenView, even to this day, is quite well known for the quality and range of content that they put out into the world. How do you think your time at OpenView shaped your disposition towards raising money?

AKI: That’s a great question. Prior to OpenView, I actually worked at a midmarket private equity fund. It was an internship, but in 5 months we looked at 70 micro industries and we found some great acquisition targets. I have to say, because it’s so funny – my old boss from that time ended up buying a dating site for farmers called Farmers Only, and he grew that into a very large company. It was pretty cool. But he did not use traditional venture financing for that.

When I went to OpenView, I saw companies that had raised a lot of money and companies that had raised a little money, and basically my dial got set to, “Don’t raise money, or push it off as late as possible.” What we want to do is figure out all the hard stuff, and then when we’re ready to scale it and pour more fuel on the fire, only then do we raise money. We don’t want to raise before we have figured out some of the hard stuff, because money doesn’t solve problems. People solve problems, and money can paper over the problems and make you think that you’ve solved something when you really don’t know what you’re doing. So, we try to minimize that.

For the first couple of years, we did not raise anything. Then we raised some small amount of angel, and then last year we raised our first venture round, and now we’re still very cautious about raising new venture. From an economics perspective, we don’t really need to raise any more in terms of cash in, cash out. But there’s a broader game to be played, so there will be future fundraises.

But what it’s essentially taught me is just to be very, very careful in raising, because everybody – and I’ve fallen into this trap many times – everybody thinks they’ve figured it out when they really haven’t. when you’ve really figured it out, it just screams at you. When you have product-market fit, it just screams at you.

There were so many times we thought “Oh yeah, now we have product-market fit,” but we didn’t, and then X months later it’s just yelling at us, like “What are you, crazy? There’s a long line of people out the door waiting to buy your product and you can’t service them. You’d be a fool not to raise at this point.” That’s the point that we raised at.

ROB: That’s pretty interesting. I think quite often on this podcast, we will host agency owners, and they really view it I think as a point of pride that they have had to get through and solve many problems without raising money. For most marketing agencies, raising money is not an option because the exit multiples are multiples of revenue.

I think what may not be seen, and you helped us dive into a little bit, is raising money is often perceived as a solution to a problem rather than just a tool to accomplish what you’re trying to do. It sounds like you’re stopping right at that point. It’s usually around the second venture round when you become very heavily committed to certain trajectories and paths and also lose a measure of control that you’re not going to get back – which is fine if all goes well, and not so fine if it doesn’t.

AKI: Absolutely. There are so many examples of those. I would agree with the agency owners where they have built it up the hard way, just by selling products that people want and listening to customers and keeping it simple. I completely agree. I think that’s a great approach, and I wish we could’ve done that.

I was not able to do that because I really wanted to build an AI product, so I just spent a lot of time writing code. Even so, it took a while for the initial prototype to come together. But that dynamic made it hard to – we would not have been able to sell enough consulting to get by. But we definitely sold a lot of consulting in the early days. That was one of the ways that we bootstrapped, partially.

I think it’s great. Look, at the end of the day, the economics too, if you can own 100% of your business, there are some very attractive economics. You see startup founders time and again where they believe their business is worth hundreds of millions of dollars, they raise $50 million over multiple rounds, and it turns out their business is only worth $100 million – which is amazing, but their personal outcome is much less and they have worked really hard to get there. They have worked the same amount of time, but they have a much smaller personal outcome.

ROB: Right. One of our former guests on the podcast is an email-focused agency based here in Atlanta called BrightWave, and they just sold. When agencies sell, you don’t get to see the published price and whatnot, but you can look at some Inc. 5000 rankings and you can look at some typical multiples for agencies, and odds are that they sold for somewhere between $20 and $40 million. They owned 100% of the company, and they probably were also able to take money out of the company all along the way because it’s much easier to take revenue to your personal bottom line from a services business versus a startup.

Then you compare that to something like – you probably are aware in your industry of TrendKite. TrendKite sold for a lot of money, about $200 million, but if you do the math all the way down to the founders, I think the personal outcomes may not have been so different.

AKI: Yes, I agree. For example, OpenView, one of their biggest or most successful early investments was ExactTarget, which sold to Salesforce for over a billion. But from what I understand and what I’ve read online, the founder of ExactTarget had 5% or something, a single digit percent of the company at that time.

Also, when you raise from VC, you have common stock but they have preferred stock, potentially with preference or layers of preference. So they got their money back first, second, third, fourth, and fifth, and then sixth is all the common stock. You will get the money. If you’re in certain industries and so on, if you have the pitch, you will get the money. But it is an extremely dangerous and expensive tool if used the wrong way.

That said, I will say because we held out and we really, really roughed it for years – and I mean really, truly roughed it – there’s even a story online of how I used to eat rice because it’s cheap and I would eat rice every day, which is a little bit of an embellishment. [laughs] But really, I paid $600 rent for a long time. We did everything to keep the costs low.

But when we did raise, we really did it – we were lucky and I think successful because we got a group of investors around the table that were actually helpful and take a long-term view themselves. They’re not quite that typical Silicon Valley, hawkish, “raise, raise, raise” kind of stereotype that you also get. We got great business partners on our board to help us think through thorny issues and help with learning and evolving.

But yeah, if we could’ve built this business without raising a dollar, certainly it would have been more amazing. But I’m very thankful for, of course, what we have.

ROB: Congrats on everything so far, Aki. When people want to find you and MarketMuse, where should they go to find you, learn more, and maybe talk to you about your agency program?

AKI: Yes, I’d love to. My email is, and our website is just, where you can actually see the product, the preview version of the brief, which is free, and you can read articles on the blog and case studies and so on. I’d love to talk.

I will say that it’s something that we have been working with agencies for over a year and a half now, maybe over 2 years, and I’d really love to have some models where we can essentially help agencies develop a content strategy offering that they can provide using our software. That’s the net end of it, and I feel passionate about that, so I’d love to hear from anyone.

ROB: Perfect. Thank you so much for joining me today, Aki, and sharing your journey.

AKI: Thank you, Rob.

ROB: Thank you for listening. The Marketing Agency Leadership Podcast is presented by Converge. Converge helps digital marketing agencies and brands automate their reporting so they can be more profitable, accurate, and responsive. To learn more about how Converge can automate your marketing reporting, email, or visit us on the web at


Leave a Reply

Your email address will not be published. Required fields are marked *