CA turns products to services, says Kumar

Computer Associates CEO Sanjay Kumar talks about partnering, privacy, portals --- and his company's new licensing model.

Computer Associates (CA) is on the spot. Its strategy of purchasing other companies, its business model, its financial reporting, and scrutiny from federal authorities was given a going-over at GartnerGroup's Symposium/ITxpo in San Diego, where CA's CEO, Sanjay Kumar was interviewed by Gartner research directors Betsy Burton and Ray Paquet. Immediately after that, ZDNet editor-in-chief Dan Farber and editorial director David Berlind got an exclusive interview during which he discussed what current and future customers can expect from the software giant. ZDNet: You spent a lot of time in your keynote answering questions about your new licensing model. How are customers responding? SK: The new model is built around the premise that software will be sold as a service, and that gives customers two things. First, it gives them a lot of flexibility. They do a short-term license or a long-term license. They can roll it out to a hundred or a thousand workstations. There's a lot of choice. Second, they'll get better products because improvements are instantly available. We have a massive head start towards transitioning the company and the organisation to get behind this model. ZDNet: But there's a higher cost with the shorter-term licenses? SK: Customers will buy on a short-term basis, at the beginning of the life cycle. Eventually, customers that want to run software for a long time will want to focus on cost certainty. They don't want to be too variable in terms of their costing. They clearly want to get to a cost-certainty license. So, I think the model provides you the opportunity to win more business faster. And yes, it is riskier because you don't have the customer committed for some number of years in the beginning. But I believe we have good products and when you have good value, you can win the business faster and that's the tradeoff. ZDNet: And then, through improved customer service and products, you'll get the loyalty of the people coming back? Like after the three-and-half-year average you mentioned? SK: Absolutely, They're going to come back. The renewal rate with enterprise customers is historically very high. So people continue using your products and the focus of this model is to get products in the door and convince customers to run while mitigating a certain amount of risk (Editor's note: Kumar is referring to the risk associated with committing to a product for a long term, only to find out it doesn't satisfy your needs). It's proven to be a good thing for us so far. I'm not sure every company can do it. I am not sure it's a good thing for every company. But for CA, it's a very good thing. ZDNet: You said you had a head start, but your answer was mostly in terms of retooling the organisation to sell software as a service. What about retooling the products? For example, will you be taking the functionality of something like Unicenter and packaging that as a service? SK: We said at CA-World last week that we are taking many of our products that can be services and creating services out of them effectively. ZDNet: Is there a roadmap? SK: There is a roadmap to do that. Especially, in technology that does management of infrastructure, especially in security--an ideal product family to do it in. But it is also a complicated area to have some components of security as a service. Then, you have embedded technology in the desktop that says, "call this Web service to do antivirus scanning or intrusion detection or something like that." But it's incredibly powerful. It's also easier for the customer to administer because they only have one copy of the product and it's a published service. For example, you don't have many different copies. There's a lot of technology behind it and you really have to rethink how technology is built. You have to rethink this idea that it's one copy and many instances of that one copy that will be utilised. Not individual copies. So, take a simple issue about rolling out a new version of product to a thousand desktops. Today, you can stagger the rollout a hundred at a time. But when you publish something as a service, boom, you're there. So, there are many other things you have to keep in mind as to how you go about doing this. ZDNet: Realistically, how many more years until that boom happens? SK: I think you're three years away if it keeps up at the current pace. There's a lot of activity today. And still, Web services does not mean .Net. There are people that are very focused on a J2EE [Java 2 Enterprise Edition] environment. Some folks on .Net and many folks on both. ZDNet: As it relates to .NET or Java Web services models you are neutral like Switzerland? SK: Yes. I think it makes sense. If you go to large customers today and you ask them, "Are you interested in Web services?" the answer is "Yes." If you ask them "Are you doing something about it?" the answer is "Yes." If you ask them "What technology will you be using to build them?" the answer is "Both [.Net and J2EE]." ZDNet: You talked about partnering with consultancies. That seems to be the new world order now when you look at the success that IBM has had in the consulting area over the last couple of years, Novell's merger with a consultancy, and at the last Symposium, Carly Fiorina talked about taking HP down that path. Is this the new channel? SK: I'm not sure it's the new channel. For us, I think it's important because we're going to be limited in growing our services capability. Our services capability is capped. We have architects for our products. We have very good implementation people for our products. But we're not in the consulting business. IBM is in a different kind of consulting business. They're into massive outsourcing and big deals and those kinds of things. We think partnering with people who are focused on best-of-breed solutions in their consulting like the Ernst & Young partnership in security. We will have a couple more in the next few months. That's a very good direction for us. ZDNet: But a lot of people are moving to this model and talking about it. Fiorina talks about it. SK: I think she wants to do it internally. She's saying, "I'm going to build my own services consulting business internally," right? ZDNet: They are number three now. But at the last Symposium, they talked about how they'd be partnering with large consultancies. SK: Yes, but at the same time, the whole premise for the HP-Compaq merger is that they want to be the next IBM. IBM doesn't partner much. They do a whole bunch of things internally on their own. My strategy is I believe consulting is important. I'm not going to go build it. I'm going to go find the best people with the best practices around the world in different geographies to help us do that. It's a different strategy but it does support the premise that consulting is important. ZDNet: Your security product, eTrust 20/20, is an interesting product in that it provides capabilities like intrusion detection. What kind of feedback are you getting from customers on the monitoring capability, the ability to monitor everything an employee does? SK: I think the whole monitoring thing is highly overrated. ZDNet: Do you mean that no one has time for it? SK: I think it is way overrated. Our premise is that customers are already collecting incredible amounts of data today. All customers collect this stuff. It's there, already built into the products they are using every day. For example, log files are created every time you and I log into or out of an e-mail system. Do I, as an individual, want someone sitting there and piling through my stuff because I looked at the guy the wrong way today? I'd rather have some technology that says, "Here's something unusual, now go look at it." We talked to a bunch of customers and they said that they collect gobs and gobs of data and don't know what to do with it. They have no desire to go traipsing through what an employee does all the time. They want to put in some rules to pick up the anomalies. That's what we want to do. I think the whole monitoring side is overrated because our technology actual monitors zilch. That's the part that people are missing. All we do is we go collect data from the logs that other systems are already creating. We aggregate those logs, apply some intelligence to it, and then we decide based on a set of rules whether there's something that somebody should look at. That's it. That turned into "oh my goodness, there's all this monitoring, monitoring, monitoring." 20/20 has zero capability to monitor anything. It collects whatever you tell it you already have. If you have Web logs, fine. If you have access cards to get into your office, fine. These logs are all over the place. ZDNet: It's like what Unicenter does for systems and networks? SK: Right, which is assimilate all this stuff, pull it together in one place, and figure out whether there's a problem. In the case of Unicenter, we built the rules as to what's a problem and what's not. In the case of 20/20, we're telling the customer that we'll give you the tools to build the rules. The beauty of the product is the visualisation combined with the intelligence. It has zero monitoring capabilities. ZDNet: Let's talk about portals. There's some discussion about how portals are core to applications and no longer isolated interfaces. How do you look at the portal business going forward about two to three years? SK: Portal is a market in transition. The portal business as we knew it last year won't be the portal business of the future. Customers are starting to take portals very seriously as a large-scale enterprise application. One of the reasons that the little portal vendors are having a tough time is because all the big names are in the business now. I think portals will come three ways. They will come bundled with applications from people like SAP and they've already announced that. They will come as stand-alone generic frameworks in which you can plug in a lot of content. But the definition for content will change. It won't be the content that we think of publishing today. There's a clear sea change in terms of the definition of content. It's applications, data, mail, messaging, and what you do. Third, portals have the opportunity to become the next super-user interface. How do you navigate between a wireless device, a PC, and a network appliance? Portals will more and more become the common user interface. It won't replace Windows. It won't replace many other things.


Have your say instantly, in the Tech Update forum. Tell us what you think in the Mailroom.