Jeff Massa, Managing Founder of YellowBrix, speaks with GrowthCap about starting multiple businesses in the early days of the internet and where he sees opportunity as its evolution continues.
RJ: Jeff, I appreciate you taking the time. You’ve had a very impressive career starting in the White House with the Reagan administration. Perhaps a good way to kick off is with background on yourself, your career and what led you to where you are now.
Jeff: Sure. I’ll be brief. I actually started my career in the Air Force back in the late 1970s, early 80s. I trained primarily as a computer programmer. The Air Force spends a lot of time and money training people so I went to 1,300 or 1,400 hours of different programming schools for languages and backgrounds and internals on several different computer systems. I was invited to go to work at the White House where I was working with the White House Communications Agency providing communications and computer support to the President and his staff. The National Security Council asked me to join them and stay at the White House as the Deputy Director of Systems which operated all the classified systems for the White House Situation Room and the White House as well. I served as the Deputy Director for three different Presidents: Reagan, Bush (Senior) and Clinton. We were responsible for the infrastructure and classified information processing within the White House and distribution of that data wherever the President was in the world, including Air Force One.
An ex-boss of mine who would call me every six months or so and ask me if I was ready to move on to the private sector, and one time I actually said, “Yes, I think I would.” So he had me go meet a public company that needed to do a turnaround, and I found 1) I had no idea of what a public company was or anything about it, and 2) didn’t know what a turnaround was. But what I found out very quickly was that the skills I had developed running different government projects transferred very easily into the private sector, and I learned a lot about business very quickly. The company was burning cash, its parent had gone bankrupt and it needed to raise capital to get the company fixed. I joined as their Chief Technology Officer, I worked with the CEO to figure out what problems needed to be fixed first, then developed new technologies that we could implement to provide faster services, better services and to reduce the cost. We went from losing money to making a monthly profit within the first six months.
RJ: That’s a quick turnaround. I imagine that got you thinking more about using your tech background in business.
Jeff: Yes, not long after that, the CEO and I decided we’d like to start our own company and we did. We identified a number of different common problems that we believed we knew we could address. This is the early days of the internet, and AOL was somebody we had been dealing with and we realized one of their problems was they couldn’t index everything they received. We started a company that would actually extract the intelligent information out of an article and only index the keywords and entities that people would actually search for. The business was called Intell-X. We licensed that technology to any of the online services allowing them to index almost in real time. So an article that was 2,000 words long would have 50 to 60 different terms that would be indexed for extraction. And they were getting high precision and recall despite people searching so they could still get to the article. The company was acquired within about 18 months.
We then were asked to start a new business division for Infoseek, it was one of the large search engines back in the early days of the internet, and we created a business model to leverage search technology in corporations; this is 1993, so kind of “pre-FAST” (FAST Search and Technology, acquired by Microsoft in 2008) days. But we weren’t finding a lot of take from the CEOs and the CTOs we were speaking with about putting a search capability in their enterprise. What we did learn is that they were all acquiring content, we then looked at that and said, “Huh! Well, we already know how to classify content and how to extract keywords and entities and we know how to summarize information; we had many tools to manipulate information.” We created a technology that would organize information into categories and subcategories and would highlight interesting other things found in an article on a public company. We created something and called it “multi-linking”, which everybody now knows as hyperlinking, but it was an underscored component for every public company and every article that we published. You could then get stock prices, information about the company, who the executives were, etc., without having to manually search as it was all pulled together.
RJ: Did this technology then immediately get widely adopted?
Jeff: Classification and entity extraction are now used in many industries, there are many companies that provide this type of software (SmartLogic, Open Text, to name a few), and many different approaches, NLP, clustering, and/or vector analysis used for classification and primarily NLP for entity extraction. Hyperlinks became very popular and are used everywhere.
RJ: Going back to your days in the White House, did you help Al Gore create the Internet?
Jeff: (Laughing) No, but I was on President Clinton’s technology task force to modernize some of the technology that was being used in the rest the Executive Office to the President.
RJ: It sounds like you had the benefit of being front and center as the internet was coming about.
Jeff: Yes, I was able to watch and live the lifecycle of mainframes to minicomputers to distributed networks to the internet.
RJ: And thinking along those lines, where do you think technology is evolving towards? Obviously mobile, among others, has been the big buzzword over the past few years. What are the areas that you’re particularly excited about?
Jeff: Well, I think there are probably three that I would say I see as big areas. 1) I think mobile apps are going to be huge, and 2) part of that is cloud infrastructure, which will allow data to reside wherever you need it and you can access it via app. And then the last is just big data, I think that’s a fun area to play in and to look at and to figure out how you take this massive quantity of information and extract bits of it that make sense for your business.
RJ: Are you looking at opportunities in these three areas?
Jeff: I am. I’m working with a few companies right now to help them in either solidifying the strategy that they’re trying to raise capital for or looking at the markets that they go into or helping them with some of the products that they’re working on. At the same time I’m also looking at some potential companies that I might want to approach either as a potential investor or to do something else with them. I keep my focus primarily on the three areas that I’ve talked about – mobile apps, cloud infrastructure and big data.
RJ: So switching gears here, you’ve been involved with raising capital from investors, can you tell us about the benefits you may have experienced working with certain types of capital providers?
Jeff: Sure. You know, because we raised capital in the day early days of the internet, I don’t think we received a lot of the network benefits that companies can get today when they’re going through a round of investment, whether it’s institutional or venture capital. We did get introductions into many different companies through both our VCs and institutional investors and those were very valuable, but I actually think that in today’s market you get more benefit than we did back then.
RJ: Did you have private equity investment in YellowBrix?
Jeff: Sure, the first investor that we had was George Soros, and they stayed in for every round that we subsequently raised.
RJ: A large part of our readership is CEOs of emerging growth companies. Having been integrally involved in building successful companies, can you kind of share with us your insights into what made you successful?
Jeff: For me, in every business that I’ve been involved in even when I was just programming, I was always customer centric. You’ve got to understand what the customer wants and needs and how it’s going to be used. Then you can build tools that make that easier, make it quicker, and make it simpler. I think computers are great for lots and lots of things, and one of the best things they’re great at is extracting relevant timely information that people can take action on. So you don’t need to use your brain to figure out how to get what you want. You’ve just got to figure out once you have what you want, what to do with it. And that’s the way we’ve always structured everything we’ve done. A lot of conversation with the customer, and a lot of use analysis, which, is the same thing I did as a programmer analyst back when I was 20 years old.
RJ: For those who are building a product or expanding into other product areas, it sounds like you would be asking the prospective customer and surveying them simultaneously or even ahead of building a product?
Jeff: Always. Always, not only by speaking with the customer, but because of the technologies we developed we could track how the customer used and interacted with the products we provided and in many cases those were the things that clarified what they were actually doing. This allowed us to understand what they wanted and to develop tools that would shorten that cycle. For example, we implemented a summary technology on all of the news that we provided. YellowBrix had licensed content relationships with six to ten thousand different publishers at any given time during our 15 year run, and what we found is that organizing the information into a specific category was fairly important. But then you still have several articles written and people want to know which one they want to read. Relevance you can do with a computer, but importance is very difficult to do with a computer. We would take summaries of the articles and when you could put your cursor on the headline, three bullet points would appear, summarizing the article, and people could scroll down and then click on the article they wanted to read just because they got a gist of what that article was about.
We learned that we needed something (summarization) to help guide people to the right article by watching them jump in and out of articles very quickly. They would click on it and then they’d jump back. If you didn’t spend more than a second on the article it was the wrong article, but if you spent 30-40 seconds, a minute, two minutes then that was something you wanted to read. We had to try and figure this out, not because somebody said, “Hey, if you create these summaries we could go faster.” We just watched how they used it and said, “Ah, even though it’s relevant they’re not being able to get to the importance factor first.” So it’s a combination of observing and talking to customers to see how they use information.
RJ: Jeff, this has been really informative and I think our readers will have a lot to take away from this interview. Thanks again for taking the time.
Jeff: Thanks for having me.