Friday, March 8, 2013
How Frost Venture Partners Is Incubating The Big Data Future
Story by Benjamin F. Kuo
Stuart Frost is head of Frost Venture Partners (www.frostvp.com), and the founder of DatAllegro, which he started in Orange County and eventually sold to Microsoft for $275M in 2008. DatAllegro was one of the first companies in what is now the "big data" industry, and helped to pioneer the entire segment. Now, Frost has set up shop buildling out a whole series of companies in the industry, and we caught up with him to learn more about the companies (Predixion Software, Cirro, Lineage, Kumo, etc.) he's been spinning out of his incubator.
What's the idea behind the incubator?
Stuart Frost: Post Datallegro, I did a two year stint at Microsoft. I enjoyed my time there, and it was part of the deal, but it wasn't a long term career for me. I've been a serial entrepreneur for about 20-odd years, and I didn't want to sit at a big company. When I left there, I thought long and hard about what I wanted to do. I decided to become a parallel entrepreneur, and started a whole series of companies in big data. DataAllegro was one of the early players in what has now become a big data revolution, in some ways. So, I set out a little bit in an ad-hoc way starting companies as ideas came to me. Now, we're turning into a formal incubator. However, we're very different than most incubators, in that we only do big data. Second, we only incubate companies we ourselves start. We are the founders of these companies, and the ideas come from our own team. We've got a very strong team of general partners, and increasingly we're getting ideas from our major partners. IBM has shared ideas with us, EMC, Microsoft, and a whole raft of other major companies are working closely with us now. On the customer side, we have companies like Samsung, GE, Kaiser Permanente, Accenture. It's been great. We now have seven companies, and have our own venture fund as well, and we'll be doing four or five for the next few years.
Why is Big Data taking off, and why the interest in this area?
Stuart Frost: I think it's a combination of things. Clearly, globally, data is coming out of the explosion of cloud, mobile, social media, and healthcare areas. It's just exploding. There's a volume of data, plenty of it, including information coming from sensor networks, machine-generated stuff, and lots of unstructured data such as MRIs, X-rays, tweets, and everything else, which everyone is trying to make sense of. The other problem, is though there's a huge amount of data, the value is pretty sparse. It's not like a structured database, where you can edit a few numbers and figure out the value. Instead, it's like looking for a needle in a haystack. Each of those dimensions really overwhelms existing platforms, whether that's a relational database or business intelligence tools. It's created this perfect storm, where there's all kinds of new technology and challenges, which are slowly being fixed.
Now, I think we're seeing major companies who see that this is great, are doing pilot projects, and see the value they are getting, and realize what they need to do is go mainstream, and broaden access to all of that data to a much larger part of their organiation--and do that, ideally, without hiring hundreds of data scientists. There's not many people on the market with those skills.
We are really set up to solve those problems, and deliver the hard core, enterprise class tools and applications that people need, in order for big data to go mainstream. There's lots of breadth and opportunity here, and some of the things we're doing are horizontal platform plays. For example, Cirro is a data federation engine, which links big data with corporate data in a very effective eway. We're also doing analytics around healthcare, as a vertical play. There's lots of data coming from healthcare and devices, and figuring out how we can get value from it, improve patient care, and reduce hospitalization for heart patients. There's lots of different plays, and lots of different ideas.
Has it been difficult to run multiple companies, and what kind of issues have you faced there?
Stuart Frost: First of all, it's not just me. I've got a great team around me, and I'm not running any of the companies directly. I think, to really explain, I should walk through the process. Let's say we saw an idea coming from a major partner, a customer like Samsung or someone like that. If they have a requirement in their organization, they'll share that with us, because they can't find a solution from either major vendors or startups which solves the problem. They'll hand it over to us, tell us here's a need we have, and would you like to start a company around it. What we'll then do, is we'll take that idea, and do lots of market validation. It's a very lean-startup oriented process. Without spending a lot of money on it, our small team here does market validation on the new idea. If that validation process results in a "yes", and we believe there's something here, and that it's a broader opportunity than just the partner that shared the idea, then we decide to start a company.
We have our own small seed fund, which we trigger to fund that company and hire a team, a CEO and a CTO. What they then immediately start to do, is prototype the product. It's lean startup oriented, where we're constantly validating with our customers and partners. Luckily, we have a fabulous network of visionaries in all of these major companies we can test our ideas with. Once we get our initial proof point and prototypes, we raise more money, both from internal and external venture capital. Locally, people like Miramar Venture Partners has been very supportive, and Okapi has been in a couple of deals. Also, we have a lot of corporate investors. EMC's a big investor in one company. That really accelerates the process. One of the things we're doing, is keeping the amount of capital required for these companies pretty small, initially. That's problem driven by the fact that most acquisition get done, commonly, in the $50M to $100M range, and if a company just has too much capital deployed, when the opportunity comes up, they can't take it. Or, they might have taken funding from a major VC, and will need to swing for the fences, which means they are going to have to pass up on an acquisition and build a bigger company. We're trying to avoid those dynamics, and maximize the chance of a very high success rate of exits across our portfolio. Of course, those acquisition opportunities don't always come up at the right time, so we're also building some very strategic, long terms plays, so that if we need to get more capital, we can do that. Predixion is a great example of that. They're three years old, and raised a Series B, and are about to close a Series C and have a term sheet. They're off to the races in terms of a revenue stream and great growth. They're going to swing for the fences now, which will be fun.
What's the secret of getting traction in the enterprise space--it seems like that's one of the hardest areas for software startups?
Stuart Frost: It's really part of that validation process I talked about. If you're right from day one, and working on real customer problems that are repeatable, and you've painted the vision of the way you want to take a solution to the customer, and there are multiple customers, you're in much better shape than the typical technology startup. If you think about how they do that, they tend to go stealth mode for a year, build something they think looks good for the market, and then launch it. But, if the market says--we don't want it--they go away and go through several iterations to figure out what the market does want, if they're lucky. Maybe two years later, they have a product that starts to sell. We're right up front, from day one, engaging with customers, partners, and engaging with potential acquirers. That's a big part of the lean startup methodology.
We're get early, constant validation, and have real conversations with customers who need these solutions, which sets things up really well. One of our new customers we started in December, and got our first customer contract in March, which was pretty quick. Not all of them are that fast, but that's not completely atypical to what we're doing here.
How has your model of multiple companies scaled so far, and has it been difficult to build these software startups here in Southern California?
Stuart Frost: I worried that it wouldn't have scaled, but the reality has turned out perfectly for me. I'm not a detailed micromanager, and never was, so it's working. I can set up teams, help them, and nurture them. Plus, we haven't had to many problems finding a team here, there's lots of world class people here.