One goal of Impact Hub Boston is to be a space for growing companies. As our member companies become successful, some grow beyond the resources and space we can provide. Aki Balogh has been a member of Impact Hub Boston for over a year, and his company is now growing beyond our space. He founded MarketMuse, a machine learning company focused on content quality. We caught up with him to reflect on how he got started, and how Impact Hub Boston contributed to his business and his personal connections.
Aki, what’s your story?
I was a developer in my teens, and I really got interested in analytics and data science. I joined a venture fund in Boston, OpenView Venture Partners, where I was looking at Big Data investments. What’s really interesting about that is not necessarily the data itself, but the machine learning, finding patterns, and using that pattern recognition to solve business problems, problems we all face. So, I thought the most interesting part of Big Data was the actual machine learning.
A few years ago machine learning was just starting to grow, so I thought it would be really cool if I started a company that used machine learning to solve an actual problem that people have. I actually started in healthcare because I wanted to help people. But I quickly realized that getting access to data in healthcare is really hard and expensive, and it’s really hard to build my first company in. So, I pivoted to analyzing web content, because there’s a nearly limitless amounts of web content. You can just pull in data and analyze it at any level of scale. Plus, I was interested in content marketing in general, so I thought this would be a great first company to start—to use machine learning to help content marketers.
The problem that a lot of people have struggled with is: if you’re going to invest in content marketing, the only way to make it work is to actually sit down and create a bunch of content for your business. That’s a ton of work, and it’s really hard to do well. You’re trying to create not just a quantity of content items, but also high quality items, things that people actually want to read, and are relevant. So, I thought, I wonder if there’s a machine learning way to solve that problem? And that’s what we built at MarketMuse. We analyze a bunch of web content to help you figure out what to write about, and how to write about it, to create high quality content at scale.
So it sounds like if I’m starting a new company, I can contract with MarketMuse to develop content with less effort in less time. What are the inputs from me?
Let’s say you have a topic that you know you want to write about, for example, social entrepreneurship. We’ll then download tens of thousands of content items about social entrepreneurship, and build a model in our system called a “topic model” that learns what it means to be “about social entrepreneurship”—what are the related concepts, and the important pieces and topics that you want to mention. Those ideas are given as an outline, a guide for how to write about this topic, so we speed up the research on writing content.
If you already have content on, for example, social entrepreneurship, we can measure the quality by comparing the content that you have versus the “perfect” content outline (generated from our model). So, we can say “here are some of the gaps you have,” or “here’s something you should cover in this post.”
So your model is both a research lead generator, and a quality assessor?
Exactly. That’s for the common case where you know what topics you want to write about. There’s another case: for example, if I want to be the premiere social entrepreneurship content site on the web, what are the topics that I should cover? We can look at your whole site, and tell you where the topical gaps are on your site overall, where are you missing things, where do you have low-quality content. So, we can do a site-wide analysis, which is more complicated, but we can do that too.
What is it about Impact Hub that enables you to work more effectively?
I really like the community. I was working from home for a couple of years, both with this startup, and the previous startup that I worked at. Working from home is great for being efficient, but after a while, it just gets kind of boring and lonely, and you just need something different. So, I looked around, and I really thought the people at Impact Hub are nice, pleasant people with a lot of heart, and they’re just pleasant people to be around. When I got to know folks, I really had a lot of fun. Geoff Mamlet, (Founder/ Executive Chairman of IHB) has been really helpful, he’s connected me to a lot of folks. The biggest thing is the overall vibe. It’s just a really nice place to be working from.
I was expecting to find a workplace, and instead I found a place with a bunch of friends.
I need to focus on the work and make sure we’re successful, but as MarketMuse grows beyond Impact Hub Boston there are great people I’m going to keep in touch with.