If Ellison’s claims for 9i are borne out, it’s not only bad news for Microsoft’s and IBM’s database products; it will also have some pretty serious implications for the hardware makers. It’s a point that hasn’t escaped him. “It will fundamentally change the high end of the computing business. Today, our big customers, like eBay [the Internet auction site], are running big Sun 10000 servers. When one of them goes down, part of their service is down. Now they can run a cluster of four 6500s, or run six 4500s rather than one big 10000. The cluster is cheaper, faster, and a lot more reliable. Why wouldn’t you run a cluster?” Ellison had just met with Scott McNealy, the combative boss of Sun Microsystems and an important Oracle ally. Does he realize that Oracle 9i is a big threat to his business? “I don’t think he does right now. For the first few years people will cluster two big machines together to double their performance and get fault tolerance. But as they gradually see how well clustering works, they’ll start replacing their big machines with clusters of small machines. That’s not good news for anyone who builds big, fast computers. The Sun 10000, the HP Superdome, the IBM S80, all these expensive sixteen- to sixty-four-processor computers, why buy them? You want sixty-four processors of power, buy sixteen four-processor PC servers. You don’t need to buy these really big machines anymore. They’re expensive and they’re unreliable, because a single machine is a single point of failure. If it breaks, you’re down. Why would anybody spend more for something that’s slower and less reliable? Big boxes are going to be replaced by arrays of small boxes. And we’re the ones with the software that makes it all work.”
Ellison needs 9i to be the success he’s so convinced it will be. Although Oracle has maintained its overall share of the relational database business, a report from Gartner that’s due out in a few days suggests that IBM is growing its share of the UNIX market a good deal more quickly than Oracle, thanks to aggressive pricing and the naturally much cozier relationships it enjoys with the other ERP and CRM vendors thanks to the fact it doesn’t compete with them. Meanwhile, through its willingness to bundle SQL Server with its other BackOffice products, Microsoft has overtaken Oracle for the first time in the Windows NT/2000 segment. Clearly, 9i has the potential for Oracle to open up more blue water between itself and its database competitors, while also providing an economic case based on total ownership costs to justify its much fancier price tag. But can the strength of 9i be leveraged to the advantage of the applications business?
“Well, I do think 9i will be a huge success. This is a little bit like when Oracle 7 showed up and it just creamed everybody. I think 9i will sweep away the competition. 9i is also very important for our applications business. Clustering is so fast and so reliable, now even our largest applications customer can move to a single global database. In fact, we’re thinking about repositioning our applications and focusing on the single database message. At the heart of the E-Business Suite is the one database idea: one customer database, one supplier database, all your information is in one place. We have this wonderful unified ‘information architecture.’ With SAP and Siebel, you have dozens of customer databases, dozens of data warehouses; it’s really hard to find the information you need to do your job. Positioning our unified information architecture sitting on top of our clustered database will help us explain the benefits of the E-Business Suite. We’re looking at the same picture, but we’re looking at it through a different lens.
“Focusing on the database as the center of our applications strategy has another advantage: it allows us to tackle some new application areas where we’re not going to write all the applications ourselves. Instead we’re going to provide and publish a database schema with the necessary APIs to allow other apps vendors to write on top of our data model. We’re going to start with health care. I think we need a national health records database where doctors input information, hospitals input information, laboratories input information. All your health care records are in one database. It’s highly reliable, it’s highly secure, it’s auditable, privacy is guaranteed. And rapid access is assured in case of emergencies. We have a variety of partners like GE Medical who we’re trying to get to build applications on top of our national health records database.”
If Oracle were to succeed in persuading lots of independent software vendors to develop their applications using the same data model as Oracle’s own applications, it would be big stuff. It would make the E-Business Suite limitlessly extensible by offering a different approach to integration that still kept all the information in one place. Paradoxically, it would even be a way of making best-of-breed work. Unfortunately for Oracle (and maybe its customers), Ellison will probably have to wait a long time for rival enterprise application vendors to allow Oracle the kind of leverage over them through its database that Microsoft exercises over the battered and submissive PC software industry through Windows. Working with niche players that target particular industries, such as health or education, with highly specialized software is a smart way for Oracle to fill out the functionality of the suite.
To try to draw a line under its early instability, Mark Jarvis, the senior VP for marketing, had told me that the E-Business Suite might soon be given a symbolic 2.0 release. Ellison was skeptical: “Yes, I thought about that, but it’s a bad idea. Instead, we’re just going to try to do a better job of explaining how we approach applications from a databasecentric point of view. We don’t need a relaunch. We just need to tell the market about how much specific customers have saved and how fast the software went in because they didn’t modify the code. The other thing we need to do is talk more about the notion of databasecentric applications—all your applications built around a single central database. People already believe Oracle’s good at database stuff, so they might also believe we’re good at databasecentric applications. And you know what, it’s not only believable, it’s true—we are good at that.”
Despite Ellison’s confidence about overall strategy and the quality of Oracle’s products, the impact of the slowdown in technology investments and the deteriorating economy were plenty to worry about. With one of the trickiest final quarters for years ready to close, had the “visibility” that has been so clouded a month ago improved? “It’s very odd, our business in Japan is doing extremely well, and they’ve been in a recession for around ten years. It’s hard to understand exactly what’s going on now that the bubble has burst. Those who enjoyed the ride up are now having an unpleasant ride down. Those who missed some of the ride up, like IBM and Microsoft, don’t have as far to fall. But it is what it is. We can’t control the market. I don’t think it makes much difference whether our stock is flat or down a bit; in the long run what really determines our success and our stock price is how good our products are. And our products are really, really good right now. All we need are more references. References are the key to selling software. That’s what gives people the courage to buy a complex, almost incomprehensible piece of technology. Selling the architecture only works with a few early adopters. In most cases, even when they agree that our applications architecture is much better, they’ll still go with the market leader. People buy Siebel because ‘everyone buys Siebel.’ People have an innate belief that there is safety in numbers. It’s deep inside everyone’s biology. So until we have a significant number of references, including people who threw Siebel out, it will be slow going. So for now we’re just going to take it hill by hill, deal by deal, until we have a strong enough foothold for our sales force do what it does best, reference selling.”
Something I’d been wondering about is the effect on Oracle’s business of Ellison’s bad-mouthing of systems integrators. Ellison had sort of apologized and recognized his error, but was he doing anything positive to rebuild bridges with the consulting and SI partners whom he has managed to insult and annoy? Ellison says, “I’ve had meetings with all the heads of the big five consulting firms, so the answer is yes. More importantly, we’ve made a couple of important changes in our business that they seem
to like. First, we changed our motto from ‘Just say no to systems integration,’ which was stupid, to ‘Just say no to custom code.’ That helped. Second, we’ve been gradually downsizing our consulting business because of our focus on rapid, low-cost implementations, so we don’t compete with the big integrators very much anymore. We specialize on getting in and out very quickly.”
That made sense, but there’s been no letup in deriding Accenture and IBM Global Services. Were they more resistant than the others to what Oracle is trying to do? “Oracle and IBM are point, counterpoint. They say ‘best of breed,’ we say ‘integrated suite,’ they sign ten-year integration and outsourcing contracts, we want to implement in six months or less. It’s unlikely that IBM will ever do much applications integration business with Oracle,3 even though they do a lot of work with our database. The IBM hardware and service guys love our database. Accenture has been even more difficult to deal with than IBM. I really don’t understand why, but they seem to think meeting with us is a waste of their time. Maybe the hard feelings date back to the Robert Shaw days, or maybe it was something I said. I really don’t know. I’ve had a couple of conversations with George [Shaheen] before he went to Webvan [a struggling Internet grocery firm that would soon fold], but they led nowhere. We’d love to work with Accenture, but so far we’ve been unable to get them interested.” Accenture, it’s also worth pointing out, is an investor in Siebel.
Ellison and I then turned to discussing the future of the spin-off from Oracle that makes network computer devices—now known as New Internet Computers—the same ones that were meant to replace the PC. He’s thinking of merging the company with a little firm called Be that’s focused on developing software, particularly operating systems, for Internet appliances. Ellison relaxes, and his tone changes to that of someone talking more about a hobby than a business. “Be is a public company, and they’re having business problems, not good. But they’ve got great user interface technology and an important relationship with Sony. NIC is doing OK, $2 million last quarter, $4 million this quarter; they’re tiny but growing. All NIC really is is a $200 desktop computer running Linux and a Netscape browser. If we put the Be user interface on top of Linux, the NIC would be easier to use. Maybe we should glue these two companies together.”4
A potentially much more important initiative is Ellison’s plan for Oracle to buy NetLedger, a start-up run by a former Oracle programmer named Evan Goldberg that has produced an award-winning suite of Web-based applications aimed at small businesses. It’s a little awkward for Ellison because he has backed NetLedger with nearly $40 million of his own money and consequently owns most of the company. But now that the bursting of the Internet bubble has killed any hope of an IPO, Ellison wants to bring NetLedger and its five thousand subscribers into the Oracle fold and rebrand the software as Oracle Small Business Suite. He hopes that Intuit, the firm that makes the PC-based QuickBooks accounting application for professionals and small businesses, will make an offer for NetLedger and thus establish fair market value for an Oracle purchase. He says, “My conflict of interest prevents me from getting involved with an Oracle purchase of NetLedger. That said, if NetLedger proves to be successful, they could play an important role in Oracle’s applications strategy. We’re talking to a large auto company that’s interested in providing an online service to all their dealers. The NetLedger system is designed as an online service for small businesses, like car dealerships. It’s an ideal fit. It’s also an interesting new business model. By connecting the NetLedger service to our E-Suite software we can link a bunch of small companies that are doing business with a large company. In other words, we can automate networks of companies, rather than one company at a time.”
NetLedger/Oracle Small Business Suite could turn out to be important for other reasons. Microsoft has been dipping a toe in the small-business applications market with its bCentral Web site. In December 2000 it also spent more than $1 billion to buy Great Plains, an applications firm with a decent franchise in the small and medium-sized business sector. It’s clear that Microsoft wants .NET to offer a range of services and applications to what it sees as a largely untapped market. Ellison agrees. Market research predicts that the small-business software market will be worth $47 billion by 2006, and Ellison thinks that he can use NetLedger both to get a big slice of the action and to head off Microsoft’s efforts to extend its applications franchise beyond the desktop. He says, “Intuit is one of the few surviving PC software companies. They’ve done a terrific job. But NetLedger is an online service company, not a software company. It’s a totally new business model. An online service has a much lower cost of ownership and offers a lot more features than Intuit’s QuickBooks or any other small-business software, including Microsoft/Great Plains. Nobody is going to beat Intuit or Microsoft in the small-business software market. But if it’s a war between service and software, then I think the service model will win.” I wondered what would happen if Microsoft decided to offer $200 million for NetLedger. “I’d tell them to get fucked. I suppose Evan might take a swing at me, but I own 55 per cent of the company, and there’s no way in hell Microsoft’s going to get it.”
• • •
When we land at Chicago’s Midway Airport, there’s a limo waiting to take us to an Italian steak house downtown where we’re going to meet Jimmy Linn, or “the Judge,” as Ellison usually refers to his nephew. Linn is the son of Doris, the much older sister he grew up with and the daughter of his adoptive father, Lou Ellison. Although there’s no blood relationship, the Linns are Ellison’s family and they’re a very important part of his life. By the time we get to the Erie Café, a former meatpacking plant near River North, the dinner’s already in full swing. There are about twelve people at the Judge’s table, nearly all of them are lawyers, and most of them seem to have met Ellison before. It’s a very traditional place with dark-paneled walls and bare brick, every plate groaning with bleeding slabs of meat (the twenty-four-ounce T-bone is a specialty) or mountains of steaming pasta. And the lawyers don’t really look or sound like lawyers. From their appearance (loud sports jackets, open-neck shirts, and plenty of gold hanging from wrists and necks) and conversation (mostly about criminal mayhem of one kind or another), you’d think it was a scene from The Sopranos. The Judge, who’s only a few years younger than Ellison, has a dark tan, long black hair, and a Zapata moustache. It’s the first time I’ve met him, and although he’s outgoing and friendly, he’s also guarded. He doesn’t like the treatment Ellison has received from journalists, and he’s protective of his famous uncle.
Although Ellison is, inevitably, the center of attention, he’s not accorded any special treatment. There’s a lot of joking and telling of mostly horrifying anecdotes about Chicago lowlifes. There’s also quite a bit of grousing about the fall in Oracle’s stock price. Ellison has made Jimmy wealthy by giving him shares in the company (it’s the reason he could afford to follow his father onto the bench), and several of the others at the table have personally invested in Oracle. Ellison assures them that the business is in pretty good shape, but he’s not prepared to say when he thinks the stock will start to recover. The only thing that fazes him is when a massive plate of lamb chops arrives for him. Ellison loves his food, but the sheer size of the portions seems to rob him of his appetite. He’s been in California too long. When the dinner breaks up, Jimmy, who’s picked up the tab, puts a braceleted arm around me and says that unless I’m nice about his uncle, he’ll personally kill me. I think he’s joking.
• • •
The next day’s roundtable event turns out to be for chief information officers rather than CEOs. There’s some embarrassment about this on the part of the Oracle people who’ve made the arrangements. According to protocol, Ellison really shouldn’t be wheeled out in front of this lower-level audience. Maybe Sandy Sanderson, who heads Oracle’s consulting organization, should take the meeting instead. Ellison decides that he will show up, but not until after lunch is out of the way. He’s d
ecided that from now on, even with a room full of CEOs, they should eat before he makes his grand entrance. It’s difficult to make the pitch, he says, if you have already used up all your best lines on the people sitting next to you. In fact, perhaps because the audience is made of CIOs, who are more technically savvy than their bosses, it turns out to be a rather good occasion.
Most of those present are in service industries such as health, education, and the Chicago Police Department. When Ellison finishes his standard confessional riff on why information fragmentation is the reason that expensive computer systems promised much but delivered little, he’s drawn into an interesting dialogue with the CIOs about what they need to do if they’re to get their own organizations to follow Oracle down the path to e-business. There’s near unanimity that, in the words of one, “From an IT perspective we have islands of automation—little fiefdoms that like spending money on their own bespoke systems.” The CIO of Western Digital, a maker of hard disks and storage for computers, says that he figures his firm is running on a total of thirty-eight different applications when it could be using just six—five of which are within Oracle 11i. But when he tried to change things, eight functional directors “slapped me down—the message was, you’re in my penalty box.”5
Without having the CEO “in their camp,” Ellison admits, they’ll find it tough. “New software is a great catalyst for change, but unless people are willing to change the way they work, the benefits will be elusive. New business processes are at least as important as new software. At Oracle we began with a project to put in new software. The software made us change the way we worked. Those new business processes led to a more efficient global organizational structure. The sum of all these changes was a new Oracle culture in constant pursuit of higher quality and lower costs. It was a surprise ending to what began as a software project.” However, the message that Ellison most wants to leave this technical audience with is the idea that the database comes first. “Count your customer databases. Every year you should have fewer. You’re not done until you get to one. Once you have a single customer database, your company has entered the information age. Technology is important. Applications are important. But chief information officers should focus on delivering better information to the business. That’s what Oracle applications and Oracle technology are good at, managing and delivering information.”
Softwar Page 33