IBM’s startup guy: We’re cracking new markets

Drew Clark has one of the best jobs in tech. As director of strategy for IBM’s Venture Capital Group, his job is to run around Silicon Valley, seeking out startups for IBM to partner with — a great occupation for a guy with breathless enthusiasm for new technology ideas. When I first met Clark nearly a decade ago, he was all fired up over SaaS; these days, he’s more likely to get overexercised about inventive vertical analytics apps.

A physicist by training, Clark joined IBM 30 years ago as a programmer, just as IBM began to develop its software business in earnest. He eventually moved to marketing and brand management, serving as IBM’s “search czar,” and at the height of the dot-com boom, co-founded IBM’s Venture Capital Group — which, interestingly, has no fund. Instead, IBM works with VCs to take promising startups to market and increase the odds of their success.

I caught up with Clark in San Francisco at the IBM SmartCamp Global Finals for entrepreneurs, an event aimed at identifying early-stage startups that align with IBM’s Smarter Planet vision of an interconnected world. Along with examining some of the most promising ventures, we talked about today’s relationship between startups and conventional IT, and how IBM plans to play a pivotal role in a new ecosystem of cloud services. The following is an edited version of a wide-ranging discussion.

Eric Knorr: How did it come to be that IBM’s Venture Capital Group invests no capital?

Drew Clark: In 2000 … we felt like we were observing what everybody observed: We saw in excess of $100 billion invested into venture capital on a global basis. And we said — holy cow! What if we could do a better job of leveraging that spend so that some of it came our way? We don’t want to be the only ones footing the bill for R&D around software, hardware, and services. And what if we could kind of use this as a leverage point for IT investment?

We were going to form a group on the West Coast, and I was one of the founders of it, and we were either going to go raise our own fund like a VC or we were going to partner with the VCs in some strategic way and kind of marry our skill set, which is about markets and customers and kind of deep technology knowledge. Those are kind of DNA elements of IBM. The DNA of the VCs, of course, is about identifying great entrepreneurs, promising companies, and being able to look at five companies that do roughly the same thing and figure out which is the best one to invest in — not really our strength.

We tossed around the idea of — should we raise a fund? Should we not? We went up and down Sand Hill Road, literally, and we talked to VCs. Uniformly, they told us — thanks very much, IBM, we don’t think you should raise a fund. Where we could use you, IBM, is after we put the money in, help us take the company to market. Help us understand the market. Help introduce [our startups] to your customers, get them into your direct channels, get them into your partner channels.

Because you remember, the background of IBM is we’re plumbers, right? We build the infrastructure, so in many cases, we need the apps and services that the entrepreneurs are bringing forward to kind of complete the solution picture. It’s complementary.

Knorr: It seems IBM is more and more interested in startups outside mainstream IT. You’re even going as far as clean tech.

Clark: Oh yeah, absolutely. You bet. In fact, we have a company here today, SecureWaters, in the clean tech space. They have a water quality monitoring system. What they do is they sell an outcome, which is safe water. They do real-time sampling of the water. Suppose we’re talking about a reservoir somewhere, like Hetch Hetchy, somewhere in their system. They have a sensor that collects water. It either collects it as a sample or it continuously monitors. You can do it either way. Their company is two pieces. It’s a device, and it’s analytical software, and the two work together.

Knorr: So it’s still IT related — sort of like bioinformatics.

Clark: It’s incredibly complicated software. You’re going to see the IT piece in a second. The sensor — this is the wildest part and people can’t even believe this — actually observes the behavior of the bacteria that’s in the water, and apparently the bacteria behaves differently in the presence of certain toxins, certain chemicals, certain toxins and toxic substances, and it’s repeatable.

They introduce these compounds that they’re trying to find, and they observe the behavior, and they make a signature of that. And they plug that into the algorithm. When they’re monitoring and they see that behavior they say — oh, this is a heavy metal in the water. It’s very clever the way they do it. They have a video camera that can look at the algae and the bacteria … which in the presence of some heavy metal alkaloid or something jump up and down, or in the presence of other things swim left to right.

Knorr: Pattern recognition.

Clark: Yeah, and that’s analytics. They’re doing analytics in a very different way from what you tend to think about, yet it’s a very powerful application of analytics. It really is a software business, a very analytical business, it just happens to have a collection point. A lot of Smarter Planet businesses are like this. I got to tell you about one more because these are so cool.

Knorr: I don’t think I could stop you if I wanted to.

Clark: They have a smartphone app and the sensor is the camera. All these companies are kind of outcome driven. Their outcome is a long life. They want you to live longer. They help detect skin cancer.

Knorr: Wow.

Clark: So you take a picture with your camera of a suspicious mole or lesion or whatever, they [SkinScan] send it back to their server, and here’s the cool part — they run fractal geometry against the image. I mean, these are two kids. They came up with this I don’t know where. Fractal geometry, you have to see this to believe it. And they take the photograph of a very innocuous-looking thing. It’s just a brown blob.

Then when you see it on their server, it has all of those topological kind of rings and different colors and certain streaks of like red and green showing off unusual behavior. Then just like SecureWaters, they run that through a taxonomy of known issues, so they can make a judgment about whether they think it’s maybe precancerous or cancerous and maybe you ought to go see a dermatologist or a doctor. They don’t claim to be a medical device, they don’t claim this is 100 percent accurate, but at least it gets you moving toward the doctor. And it’s a $2 iPhone app.

Knorr: That’s amazing. Well, it seems like your interests have broadened from the core IT stuff.

Clark: Our customers have broadened what they mean by IT. We still have a warm fuzzy spot for the data center, of course. But increasingly our customers are moving toward the edge, and their needs are moving into all kinds of previously what were known as vertical spaces.

Knorr: As you say, the common thread here is analytics, right?

Clark: That’s what ties it all together. You and I have been talking about startups for a long time. There was a lot of talk about 10 years ago about sensors. A lot of startups got started building smart appliances and smart sensor devices, but a lot of that got commoditized pretty quickly. Then they moved up the food chain to data and said — well, the real value is in the data. It never really was in the sensors.

What we’re seeing now is that it’s really analytics. That it isn’t just the data anymore, it’s bringing the value from the data and being able to use that analysis to make better decisions to do smarter things.

Knorr: And value in some cases you didn’t know was there, right?

Clark: Exactly. Right.

Knorr: Because you’re talking about all of these vertical areas where it takes domain expertise to be able to make sense of the data, the data visualization component becomes very important.

Clark: Right, exactly. Visualization is a very powerful way to help make sense of data and we consider that an important part of our analytics strategy. In our Smarter Cities area, we’ve got something we call the IOC (Integrated Operations Center). It’s a dashboard for a city.

Imagine you have a city … where you have all of these stove-piped operations. You have police, fire, water, transportation, and they’re all kind of verticals. Of course, the tragedy of most cities is that there is no real connection. They kind of exist by themselves, and we as citizens can’t understand this, but that’s just the way it’s kind of grown up.

What we do with Smarter Cities is — I like to call it City as a Platform. If you think about it, it makes sense. This is an app platform model. So the police can go and build some systems here and fire can build some systems. And guess what? Entrepreneurs are coming in and literally building out this kind of infrastructure. It’s essentially leveraging and channeling the enthusiasm of a lot of entrepreneurs. I mean, Angry Birds is a fun game, but what if you could take some of those people, those talented programmers, and get them working on helping us do a better job of managing transportation or responding to emergencies or helping us sort out the terrible parking situation?

So when we talk about Smarter Cities, we’re really thinking about — can we come into a city and help build a platform? And everybody would be on the same page and understand what’s going on and there would be a lot more collaboration and flow among the agencies.

Knorr: That was the promise all the way back to the early Web services days.

Clark: Well, I think we’re finally getting there. I don’t want to oversell this, but we are starting to deliver lots of these IOCs to cities, and they love the idea it’s not just a dashboard that’s kind of hardwired, but it really opens up this kind of platform thing. Of course, we host a version of this on our cloud now — I call it “city as service.”

The idea that we can host this whole thing and provide connectivity across agencies and then it becomes a much more agile system as it wants to grow and expand. The other beauty of this is that there’s another set of APIs, and that’s for people. That’s for people who live in the city because they should be able to get access to their government.

In so doing, we take another burden off the city, which is the citizens constantly demanding access to more services the city doesn’t know how to provide, given the way they’ve traditionally built their city infrastructure.

We’re trying to do this one city at a time. We think the city is the perfect unit of work for trying to do this because it’s the nexus of a lot of these problems that you see. If you can make the city work, then you can start to say, OK, I have city A, and what about city B, and maybe city C, and now maybe I can interconnect those.

Knorr: And when you get down to the pothole level and people care about it. That helps sustain it.

Clark: Who doesn’t want to be the mayor of a Smart City? The way this works is that we present this to the mayor. They typically love it. And their first reaction is — well, I’ve got to go find a way to raise the money to go do this. So the mayor becomes the champion and the mayor wants to be the hero who built this thing. It’s a very compelling proposition because he or she can go to the people who might fund this cause and say: “I want to build a Smart City, are you with me?” And nobody wants to say: “No, we want to keep it dumb. Everybody wants to have a Smart City.”

Knorr: Do you think the sum of what you learn here will be able to be cycled back into enterprise?

Clark: I think it’s an extension of the enterprise. I think it’s the enterprise moving out to the network. This is all very much an enterprise play for IBM. It’s enterprise moving to the edge.

Knorr: How does all this relate to mainstream trends?

Clark: You have mobile, social, localization, and cloud in this amazing confluence. And it’s happened so quickly. A lot of observers have said that they’ve seen individual trends happen this fast, but they’ve never seen this kind of confluence of trends happening all at once. We have four things all converging at the same time.

Knorr: On the other hand, there’s also an enormous pressure on it to create a private cloud, to become as efficient as Amazon or whomever.

Clark: It might even be unrealistic.

Knorr: It is unrealistic.

Clark: I mean, how much can you really scale some of that stuff before you lose the whole advantage?

Knorr: Since you’re a longtime observer of the scene, I’m interested in your qualitative take on what’s happening with mainstream IT. Traditionally, the center of the world has been the data center. What’s your sense of the mood there? Or do you deal with that culture that much anymore?

Clark: We do. IBM holds CIO conferences every year, and we bring in the CIOs from our top customers. They were way ahead of us. They were already deploying clouds. They were already figuring out the balance between public and private clouds. Let’s just say it’s been recommended to them that they really start to adopt the cloud and deploy and mobile devices and all that. I know in some cases it’s felt a bit like a loss of control. But in other cases, I think they’ve gleefully done it because it offers a chance to really align with the business. That’s really what IT craves. And I think that the adopters of cloud, the advocates, are the CEO and the business units, not the CIOs.

Knorr: That’s exactly right.

Clark: Sometimes the CIO comes naturally along with them, sometimes they drag their heels. But in a lot of cases I think the CIO has turned a challenge into an opportunity. And I think that all of our major customers have a program somewhere and they’re somewhere along the path.

But I think the change we’re starting to see is that IT departments have now got a vision for how they’re going to make it work. And I think the BYOD thing is playing into that. Some of the more interesting things I heard from these CIOs were where the BYOD and the cloud come together — because they see this as a unique opportunity to kill a lot of birds with one stone. They can see that if they go and they build a private cloud, let’s say, a platform, and they industrialize some of their systems and put APIs in place, that they can accommodate a lot of the BYOD stuff a lot more easily.

Knorr: That’s an interesting point. How does that work?

Clark: Instead of having them come to me one by one with the latest app of the day that they downloaded over the weekend, that they want me to support, what if I just install an app store on the cloud and kind of let them bring this stuff in and let them self-provision it?

A lot of customers are starting to think of building that nexus where those things come together and seeing that that is really a very efficient use of their resources. And it’s delighting the end users because they feel like, OK, now this guy is on my team and we’re doing it right. The CEO is happy because they’re getting the cloud systems built, and they’re getting what they want.

Knorr: Yet on the ground, there’s the data center itself, which is still a mess in a lot of large organizations. And there are various private cloud efforts there that are trying to rationalize that. OpenStack, for example — that’s just an explosive ecosystem.

Clark: It’s just the hottest thing in cloud right now. It’s good-quality stuff, too. I think it’s really shaken up conventional wisdom about the cloud.

Knorr: IBM has been kind of tentative as far as OpenStack is concerned.

Clark: I think … well, we have a lot of … what should I say? We have a lot of work to do in delivering the solutions that our clients are asking for now, just within the framework of what we offer. But I think looking longer term, OpenStack has to be factored into what we’re doing because our clients are starting to ask about it also. I think you’ll see a day when our systems will work with an OpenStack configuration. I can’t be more specific than that.

Knorr: What about the public cloud? IBM’s already into it, although not in a very public way. IBM has not talked much about public cloud services.

Clark: I think that’s going to change, and I’ve always been open with you about what’s changing. Just last fall we announced a SmartCloud enterprise. And they had a lot of elements of public cloud in them that we maybe didn’t articulate loudly enough, but they’re there. But I think you’re going to see more of that in one particular area at least, and that’s the PaaS area, which is a very dear spot of mine.

The value of the cloud is unlocked at the PaaS level, so we’re very eager to start putting more of our industry frameworks and our middleware up at a level that can be more easily leveraged by entrepreneurs.

Knorr: I think the IBM brand can infuse a lot of life into PaaS because right now it’s really small, independent ISVs, game developers. There’s not a whole lot of enterprise activity on PaaS.

Clark: For apps that make sense to be in the public cloud, why not attract enterprises too? We also think it would be interesting to provide a platform for IBM to expose some of its own things that we haven’t before — things like our middleware, like Tivoli, that kind of stuff.

Imagine what you could do if you had analytics as a service running. We’ve been talking about all of these [SmartCities] apps being analytical apps. How are they going to take advantage of the heavy lifting we do with SPSS and Cognos and other things, if we don’t make [analytics] available as a service to them, as an API?

Knorr: That would make a heck of a cloud service.

Clark: You might find this hard to believe, but the app development part of it is probably the least important piece of it. I mean, we have partners. There are good companies out there that can do the app [dev platform] — we can get Engine Yard and Zend and those guys. They can do a lot of the DevOps and app development piece here.

What’s interesting, though, is the value that we think IBM could add is all the other pieces. Analytics, storage — who knows? I mean think of all the as-a-services that we could stub into here and make available to all three of these guys.

That’s what we think will be most interesting, when we start thinking about PaaS — and maybe this will be our differentiator — as an enabler for all kinds of services, not just a place to develop and deploy apps in a more leverageable way. I mean that’s really what PaaS is today, right? It’s development in the cloud. It basically seems like kind of a strange value prop. What we’re trying to do is say: We want to push more services, call them cloud services, up into a PaaS so that you can get more value.

Knorr: One last question: With startups in general, are we in a bubble? Particularly with social media?

Clark: Just looking at social media … I think any objective observer would say there’s certainly a bubble there. I’m not picking on any particular company, but the whole space is crazy.

I think on the enterprise side, I don’t see [a bubble] at all. I think that what’s interesting is … a lot of venture guys are coming back and embracing a lot of the opportunities in some of these Smarter Planet spaces. We’re opening up and enabling new markets.

That rubs off on the enterprise also. Maybe that says that more attention will be focused further down where these things have to run. I think it’s always healthy to grow the opportunity high up on the value chain because it trickles down to making sure the underpinnings are working. It’s going to keep investment in cloud strong. I think it’s going to keep pushing the envelope on performance of hardware and underlying systems. I think it’ll keep pressure on security to keep up with all of this, if it can. I think that a lot of the areas in the food chain will continue to grow.

Knorr: I’ve always thought you had one of the best jobs in the Valley. It seems like you’re enjoying it more than ever these days.

Clark: I am, because I get to meet all these great new entrepreneurs who have all of these great new ideas. The good news is that we have, we think, a good value prop for them. We actually have something that we think has a nice impedance match between their new thing and the enterprise, something that could help pull them through and get them to market.

Knorr: And you get to be right there.

Clark: Absolutely. It’s fun to do.

This article, “IBM’s startup guy: We’re cracking new markets,” originally appeared at InfoWorld.com. Read more of Eric Knorr’s Modernizing IT blog, and for the latest business technology news, follow InfoWorld on Twitter.