Interview with Martin Hack (founder of Skytree)
In San Jose (CA), we talked with entrepreneur Martin Hack about the business model of Skytree and how he started his company. Furthermore, Martin shares his learnings and advice for young entrepreneurs.
The transcription of the interview is uploaded below.
Martin: Friends of Entrepreneurial Insights, this time we are in San Jose. It’s very close to the German Oktoberfest and that’s why we are interviewing a German entrepreneur here in the Silicon Valley. Martin from Skytree, who are you and what do you do?
Martin (Skytree): Hi. Thanks, Martin. Thanks for having me. I’m Martin Hack, I’m the CEO and co-founder of Skytree. We’re a machine learning company in the big data space and we’re focusing on making predictions.
Martin: I mean you have a great weather here. So what would be your prediction for the next days?
Martin (Skytree): Since it’s California, it’s going to be nice because we really have nothing to worry about that. We have about 240 sunny days a year here, so in a way, nothing to worry about on that front.
Martin: Great! How did you come up with this idea of Skytree?
Martin (Skytree): Yes, so eventually this is my first company where I was the founder. I did a couple of other startups in the past, worked at big companies that would last for 25 years. I started out in Europe and then came to the US, 15 years ago.
About 3 or 4 years ago, I started to see that there’s a more and more need of a big data that wasn’t here before. People started talking about big data. But really, the insights, how do we get the insights from all these data that’s out there, it was pretty clear that there was still a missing link essentially.
A close friend of mine, Alex Gray, which I had known for 15 years, we stayed in touch over the years. He was a professor for machine learning at Georgia Tech. We got together about once a year. We started talking about 4-5 years ago, what would a company look like, what could be used. We thought all of the really important applications for that. Ultimately, we decided to take the plunge and started a company which was about 3.5 years ago.
Martin: Can you explain briefly what Skytree really does?
Martin (Skytree): Sure. So, we’re essentially an enterprise software company. We’re selling software in the cloud and on the premise, for being massive scale machine learning on big data. So that could be anything from making a prediction, making recommendations, finding outliers, finding patterns, those are usually the use cases where people use machinery for it. We provide the software in the services for a customer for that.
Martin: Do you also have third party applications that you are selling on your platform or everything is developed by yourself?
Martin (Skytree): Everything is by Skytree. We worked with a lot of the Hadoop partners out there, so we worked very closely with the 3big Hadoop vendors, they’re all partners and friends of ours. It’s very much big data ecosystem nowadays. It’s very hard to do it for one vendor alone. So we have a very good partner system out there and Hadoop vendors are very near dear to us.
Martin: Do you have a technology developed that you can apply to several problems or did you adjust those kind of technology and developed singular product?
Martin (Skytree): It’s essentially a platform, so there are multiple use cases for that, with what we call a vertical approach. Some of those areas are for example, in financial services, anything from fraud detection and fraud prevention, that’s a big use case. Another one would be around risk scoring, credit scoring. And another one would be around marketing and targeting, so who should we essentially market to. Those are the very common ones.
Other ones are, around what we call predictive maintenance, predicting when a third parts are about to fail. Think about cars, think about energy transformers, think about utility, sometimes multi million dollar units. If you can predict something before it fails, that’s a great asset to have at their disposal. So those are the things we’re working on with our customers.
Martin: When you started the company, how was your go to market? How did you try to acquire some customers? Did you start with one product and then just acquire a specific subset of the customer that you have currently or how did it work?
Martin (Skytree): So for us, we were kind of in a fortunate position because we were still in stealth mode. We had no website, no phone number, but we had customers.
Martin (Skytree): So how did that happen? There was such a demand for the technology that people through rather obscure channel find out about us and said, we want to work with you, can you get in and help us. So those verticals are the ones I already mentioned, financial services, retails, and insurance companies. The go to market was kind of already planned out for us without us even doing anything. We enforced that and grew the customers base in those verticals. But ultimately it wasn’t really a solution trying to find a market, the market was already there and asked, Okay, can you help us.
With the big data environment exploding in most of these customers, and these are all Global 2000 customers. So these are the biggest brands, the biggest companies out there who have essentially the need to do those kind of computation analytics. For us it was a perfect match so to speak, because we had something that they wanted and for us, we focus and tailor offering around those kind of applications.
Martin: Did you get to know this kind of first time customers before you started or did it just happen accidentally getting to know them?
Martin (Skytree): I think it was both eventually. I mean, some of them came through networking where we knew, I mean there’s only a limited number of customers that would buy that initially and we knew in Global 2000 list there are the 50 biggest banks, the 50 biggest investment banks, the 50 biggest retail banks and so on. So we knew who they were and it’s a rather closed community at that point. And then you would ask for introduction or we already knew somebody there. So usually you get the snowball effect or the net effect, and then ultimately if the number one in that industry is using your product, number 2 and 3 and 4 probably want to use very quickly thereafter because they realize they’re missing out in the market.
Martin: This B2B market segment, one of the major problems is really identifying the key decision maker. How did you identify those people?
Martin (Skytree): That’s ultimately the challenge of any kind of sales, environment enterprise sales is somewhat a combination of art and science to a certain degree. You just have to essentially figure out who is the economical buyer, who is the technical decision maker. And that might vary. There’s not one size fits all. It’s not always the same person who makes the decision. It could be literally across the board and some organizations, you might have 5 people who have to say “yes” before there’s a purchasing decision.
That part is essentially of the engagement, you have to essentially figure those things out with the customers. The most important thing is that you have a sponsor upfront that actually says, “Yes, I believe in this technology, this is going to get us to the next phase” where the customers want to be. And it has to be value. If there’s no value, nothing is going to happen.
Martin: Martin, in terms of corporate strategy, what is the competitive advantage of Skytree?
Martin (Skytree): I think the technology part which is ultimately driven by the people. The people who we have in the company, how we started out, what we have right now, are the ultimate differentiation for us. In our engineering department, we have about 92% PhD. A lot of the guys that created and invented those algorithm worked for us, or we work very closely with them. So we already kind in a way, the folks who did a lot of the machine learning that’s out there today. So we came up with it.
In other words, one of the big factors for us early on was essentially the people that we’ve evolved with. Not just on the hiring side, but also the partnerships, we have a massive partnerships with the university networks out there. In a way Skytree, we are kind of curators between academia and the industry and we put a huge focus on hiring essentially the best in machine learning. That’s what we have done.
I would say, if it’s only one thing that differentiate for us, it’s the people inside the company and the people that we work with, that allows us to build an environment and a system that’s pretty hard to beat in the market place.
Martin: What have been you learnings from working with those university institutions?
Martin (Skytree): I think it started out with us, because we have a lot of academics that started the company. So we were kind of thrown into that already. So it was a natural fit for us and I personally was very positively surprised because essentially whatever you work with, you get a very high level right off the bat. When you work with universities out there, with very clear goals, very clear targets around the objectives they want to accomplish. And they’re usually the folk leader in that space. So they are, that university or that professor that came up with something, we work with them. That makes it a very nice partnership for us and at the same time, many times we recruit directly from those universities. Getting their master or PhD programs.
Martin: In terms of market development, can you give us a broad overview of how you perceive the current market development and predictive analytics?
Martin (Skytree): I think it’s not just so much about predictive analytics around big data and predictive analytics. So I would say big data certainly is at a certain hype level that is hard to beat at this point. It’s definitely out there. People are using it. But people are now looking to hear what’s the next thing after big data? If SaaS and Cloud was in maybe 4 -5 years ago, big data certainly happening now and what’s the next thing.
We believe machine learning, specifically predictive analytics is going to be huge part of it because that’s the thing that allow us to gain insights in what we call actionable insights from the data. So you might have a data lake, a data hub, whatever it is, now how do we get to the next level, which is the insights.
So predictive analytics and machinery is going to play a major role. To our point is where we’ll live in the next 5 years every Fortune 500 is going to have a machine learning system at their disposal. But it’s in their cloud or in premise, but they are going to use it because it’s the next logical thing after BI, Business Intelligence. Business Intelligence has been around for 25-30 years. People want to go, what’s the next thing. How do we go from looking at yesterday’s data at predicting what’s going to happen to a business tomorrow. And that’s really machine learning and that’s where the journey is really going.
Martin: What would be your forecast? Would it be more each and every company develops their own predictive analytics tools or is there so much economies of scale when some sell product like you also, develop a tool and sells it to all the other companies?
Martin (Skytree): It will always be companies out there who want to do their own thing. There’s nothing wrong with that and open source is a great example of that. If you look at Linux, it’s everywhere, people are using it and it’s a great success story.
However there’s also going to be a maturity of the market that just want to use the system. A great analog to that would be, a relational database 25-30 years ago. People asked why they need a database. Today nobody would ask that.
While you can still build your own machine learning system yourself, potentially most customers just say, “Hey, can I have a system that can do the same thing what these guys are doing?” We think that’s to your point, economic of scale, it’s going to ultimately probably succeeded in the market because it’s just much easier to deploy, to manage, to use, and to support ultimately. That where we see the market is going right now.
ADVICE TO ENTREPRENEURS
Martin: What have been your major learnings and maybe dos and don’ts that you have seen over the last years?
Martin (Skytree): There are a lot of things you learned by failing and maybe those are the most painful lessons but the most important ones. But a lot of the times it’s actually working with the right people and basically surrounds yourself with the smartest people you could potentially work with, or have as advisers or mentors.
That was essentially from day one, our mantra basically to be out there working with the people that we bring in and them essentially be at the certain level and at the same time have advisers and a network of people that can support you.
Ultimately, surround yourself only with positive people. That’s something I’ve learned over the years. You don’t want to be with people that dragged you down, you want to be with people that lift you up potentially. Whether that’s in life or in business. It’s ultimately the people that are positive are probably going to emerge as victors.
Martin: Okay. Great! Are there any specific learnings to machine learning or building big data companies?
Martin (Skytree): Yes, I think there’s certain things in hindsight, you could always say, we should have done this, this and that. But if I would have to do it again, I would say, I wouldn’t change that much. I would change potentially the makeup of the product in a certain industry, the way we go after. But those are small details that you can basically learn while you’re doing it. There’s nothing that’s a major, oh wow, this was like a major screw up.
But the small things sometimes they do have big impacts. One other things that we’ve seen early on and in hindsight we could have done is essentially be very specific and even more focus on certain industry and verticals and application essentially. So that would ultimately accelerate time to market and will get you a better product market fit.
So those are the things in hindsight – yes, we are fixing them. We are doing a better job now but if you do those things earlier on, you probably are in a better position going forward. But nothing that you couldn’t fix.
Martin: Great! Martin, thank you very much for your time and your insights. Now we should get a beer because the Oktoberfest is coming. Thank you very much for watching us.