Sybase CEO John Chen, in an interview shortly before SAP announced plans to buy Sybase for about $6 billion, gives clues about why his company might be worth so much. The mobile enterprise has a lot to do with it.
Few companies get a chance at a second life. When John Chen signed on as CEO of Sybase in 1998, the database software vendor was, in Chen's own words, "a very, very dead company." Once a strong competitor to Oracle, Sybase had lost its way, in part because it missed the opportunity to enter the enterprise application market Oracle now leads.
Over the next decade, through the efforts of Chen and his team, Sybase turned around and reinvented itself as an enabler of the "unwired enterprise." Then, in mid-May, enterprise software giant SAP signed a merger agreement with Sybase, citing Sybase's leadership in both mobile and in real-time analytics.
Since the merger announcement, Sybase has not granted interviews to the press. But in March, before rumors about the merger began circulating, John Gallant, chief content officer for IDG, and Eric Knorr, editor in chief of InfoWorld, sat down with Chen for an hour-long chat as part of the IDG Enterprise CEO Interview Series. The interview explored how Chen was able to rescue Sybase from the brink and establish the company as a key mobile enterprise player. This abbreviated version of the interview focuses on the company's mobile business and insights. To hear the complete interview, which goes much more in-depth on Sybase's history and analytics business, listen to the interview in its entirety:
Gallant: You've got a pretty interesting and diverse product set. You have the database, analytics, mobile management tools and mobile tools that are pretty widely deployed among the service providers. Help us understand how all of those different technologies fit into one cohesive strategy that makes Sybase unique.
Chen: Sybase traditionally has been an infrastructure software provider to the enterprise. We started as a client-server database company and then moved into development toolsets when we acquired PowerSoft.
And after I came on board, we continued to develop our database. Gradually, we positioned ourselves in high-growth areas, like analytics, mobile middleware and mobile services. And how we envision that it all comes together is that we believe mobile enterprise computing is going to be the next really big thing. If you think that e-commerce was a big sea change in the early 2000s, m-commerce will make e-commerce a very small thing. M-commerce reaches almost the majority of 6 billion people around the world.
Knorr: What characterizes an enterprise mobile service as opposed to a consumer service?
Chen: Applications must talk to applications; machines must talk to machines. There has to be intelligence gathering, mobile analytics, real-time reporting, all these things. We could be a leader in providing those kinds of infrastructures. It starts with back-end infrastructure in a data center to people sitting here with a device.
When you talk about the database, there are many aspects of it. An OLTP database sitting underneath applications and collecting records. That's a database. Data warehousing, analytics. That's a database. A database could be an embedded persistent store in the field, like if you're driving a truck, with a FedEx machine, or in U.S. Navy or Army deployment units -- they have on-board computers. That's an embedded database. So to us a database doesn't really conjure up the image of a dark room and lights blinking. It's that plus many, many things.
And we really do believe in [a mobile] persistent store. Five years ago, that was a laughable subject, because there wasn't enough capacity, there wasn't enough battery to sustain it. Well, five years have gone by and things have changed quite a bit. Things like the iPad are coming out. Who knows what the next two generations of the iPad are going to be like? I mean this is not only an Apple story. How does an enterprise take advantage of all these technologies? How do we manage these things? What about security concerns, about who's going to use what to see what? And how do we build intelligence in?
I'm holding a device. I'm as good as what you tell me, right? Well, how do I know what you want, at what time? You can go on Amazon and they remember the history of your preferences and tell you what you ought to be buying next. That's just the beginning. You have to be able to deliver that information in a remote way. And all these applications, they have to be able to talk to each other agnostically.
You may have an iPhone and somebody else has a BlackBerry and somebody else has an iPad. Somebody else is staring at a TV; somebody else is looking at the voice-over-IP devices on my desk from Cisco. How does that all come together? Unified communication is the software architecture that's going to bring it through. We'd like to be a participant in that. And that's why we invest in messaging; that's why we invest in how to channel content. That's why we invest in how you manage devices; that's why we invest in all the encrypting of that information that goes over the airwaves. And that's why we invest in analytics, and that's how it ties back to what our core strength has been.
Gallant: That's quite a mouthful. How do you roll all that together?
Chen: To us, this is the unwired enterprise concept: Having a layer of architecture that translates to products and services that people can actually deploy and employ. Not the entire stack, a layer of the stack. You could use my analytics engine over an Oracle database. You could put my services engine over any other company's engine, maintaining an open heterogeneous computing environment, which was the principle of Sybase to start with. So that's how it all ties in.
Knorr: You led the company into the mobile space very early. What did you see that other people weren't seeing?
Chen: I knew we were going to get to that. That has a little bit to do with history. Most other CEOs will claim that they're visionary. The fact of the matter is that we backed into it somewhat, and to be really honest, we didn't have another choice. Because in the e-world, we lost it.
When I took over this company, we were behind in various technologies, not only in the e-world. We made our college effort trying to get there; we had our app server. But by that time, we were the seventh app server in the universe, where WebSphere and WebLogic already had 75% of the market. So you can't rebuild a company based on that.
You have to have something that you are uniquely qualified to do. I always ask the question -- what is our right to be a company?
We didn't have applications, like ERP and MRP. That was before my time. We decided not to go there. It turned out to be a near-fatal mistake. And when I came in, the e-world was big. Everybody's IPO'ing their business plan and everybody's making millions and Sybase had deteriorating revenue and was losing money and all that.
Fortunately, we had a good installed base and a good loyal set of customers. And they were questioning our survivability; so were a lot of people. So I knew [I had to] be making money and stabilizing the troops and stabilizing the customers [who said]: "OK, you guys are going to be in business, I'll take my time moving away from you."
So we decided to look at what we can do and what the world looks like. And fortunately, at that time, we partnered with a couple of companies, at least the people who were willing to chat with us. The big one was really Intel. They were pushing Centrino chips in the early 2000s. So they came up with this "unwired enterprise"; they spent $300 million to do this campaign. I had no money. So I immediately called myself "unwired enterprise." I see somebody's branding this and I don't have money to brand it. By the way, we didn't have credibility either. If I came up with "unwired enterprise," nobody would listen to me at that time, right?
But the hardest thing is to keep the sales force focused. The sales force still needs to put food on the table while I go develop this grand scheme. My sales force calls and says: "Hey, I like this stuff you're calling 'unwired enterprise,' it's sexy, it's whatever. But what the hell is it? Because my customer doesn't know what the hell you're talking about." Those are the kinds of issues we had to deal with.
So, in some ways we backed into it, but we didn't really have a choice. We needed to leapfrog it, and we knew if we leapfrogged it correctly, it would be big time. If we leapfrogged it wrong, well, it's no worse than when we were given no respect anyway at the time.
Gallant: So how were you able to make the transition?
Chen: The key was that we had a team of people who were willing to do this, who would hang in there. Because there were going to be a lot of ups and downs. And there were, but we have a great team. If you look at our executive team, we've been together for a very long time. I've been here 12 years; some of them have been here longer than I have.
If you remember, we acquired a company called PowerSoft. PowerSoft acquired a company in Waterloo, Ontario, that does compilers. In the year 2000, nobody was buying compilers.
Now, at the same time, coincidentally, it was the early days of Palm [which had been bought by 3Com]. 3Com wanted to have a lot more intelligence on a device. Apple had failed at that with the Newton, if you remember. I said, well, these kinds of devices don't have a lot of memory or processing, but it's got to be able to do more than just keeping contacts, address books. I mean, it's got to be able to do more than that, right? So we started taking an interest in that.
I came from a chip design background. I used to design VLSI chips for Motorola. So I said -- well, we could apply the same concept, take the compilers and compile application-specific code. So take the hardware concept of ASICs, application-specific ICs, into application-specific code.
So I challenged the team in Waterloo. I said: What is the smallest database you could build? And this is very interesting historically. We went to 3Com and 3Com came up with some ideas about the smallest usable applets. Apple obviously is a much better marketing company, but we had that concept way back when.
So we were generating applets with our compilers. We had the first ultra-light Java compilers that could compile things down to 60,000 bytes. [Those apps] actually did something. They didn't do a lot. But 60 kilobytes is all they gave us in terms of real estate. Today that's nothing. But in those days, 10 years ago, it was a different world.
So we got really interested in that, and then we won a lot of embedded business because of that. So at the time when IDC and Dataquest were tracking that market, called the embedded database market, the whole market was only $300 million. We owned almost 70% of that market. We were in all kinds of software, even Siebel -- you know, all the CRMs -- and Remedy. We were in handheld devices like FedEx and in all kinds of hardware, software. In fact, the government used it in the system for international travelers, with the palm reader. That palm reader is on our mobile database. High-end chain stores were using it to scan prices, like Andronico's. So we ended up doing very well.
[Editor: That's the history, what about now?]
Right now I go to my existing customer base, the database customers, and say: "You need to use devices, right? Yeah, OK. So do you use one type of device, like a BlackBerry, or do you use more than one type of device?" Well, you know, three years ago, they said a BlackBerry is good enough. Today, they say no, no, no, we want the iPhone.
So we said, "OK, now how do you manage that?" And they said, "Well, that's just the problem." So I said, "OK, let me tell you what I'll do. I will manage it for you because I'm going to provide you a server that will allow you not only to manage these devices from a security point of view, but also develop apps."
You develop the app one time and deploy, because we will write the device-specific code. You could deploy it on a BlackBerry as easy as you could on an iPhone.
That's where our technology called the Sybase Unwired Platform comes in, which is both development and deployment. We take the traditional database customer and go in and say: "I'll sell you analytics, because you've got all this data and you need to analyze it." OK, good, that's obvious. "By the way, if you ever use devices, I will sell you the server to manage all these devices."
Gallant: John, since you see the direction for this mobile market, talk about what you think enterprises aren't thinking about today that they ought to be thinking about.
Chen: Today their [concept of] mobility is still by and large an information delivery mechanism. E-mail is a perfect thing. But you know, it's not really an action-based platform.
So what does that mean? Well, taking a cell phone and paying for something is action-based. I don't think there's a serious push yet into payment, remittance -- that whole area of the mobile commerce world.
We will have to clear a couple of hurdles to get there. Government regulations are a huge hurdle, but it will be fixed, because otherwise Western Union couldn't have been in business for 150 years. It will be fixed. And there's a big fight between the telephone company and the banks -- who owns the wallet. Now that fight is somewhat political; different countries have different regulatory environments. It will be interesting.
We provide mobile unwired technology to Celcom in Malaysia, for example. In Malaysia you can trade minutes with 13 different telecoms in Southeast Asian countries using our technology. You can buy things. You can give people money or something of monetary worth, like minutes.
So that whole area -- the machine-to-machine, point-to point-of-action space -- is still in a very, very early stage of life. Partly because the high-end PDAs -- about 13% of all the devices out there -- are still relatively expensive. I could use my Web-based phone (which you could probably buy for 15 bucks) for SMSs and all that. I could use it to tell a bank to deposit something. I could do that, but I can't really do a hell of a lot more than that. And so these high-end devices need to become cheaper. It's a matter of time. Very shortly in the near future you'll be able to buy PDAs very cheap.