Softwar

by Matthew Symonds

  “In many ways,” says Lane, “the whole CPG thing was wildly successful. Larry would probably say the problem was it was wildly successful—it sold to too many clients, and we didn’t know how to do it. Well the truth is, if Ron Wohl wasn’t in charge of applications we could have done it. It can be done. IBM does it today, others do it today.12 Larry started off very enthusiastic about this stuff, and then halfway along he started turning on it because Ron would go to him and say, ‘IMI needs us to design this thing and that’s not in my budget, I need more money.’ So these partners became irritants. Larry has always hated channel partners, and Ron was the only one who had his ear. Best of breed was very difficult, hard to do, but we were doing it because our own applications sucked. He doesn’t acknowledge that.”

  Even with the benefit of hindsight, it is difficult to know exactly where the truth lies. The only question that really matters is whether, despite all its problems, Oracle was better off doing best-of-breed. Lane believes that Oracle still has customers that it would have lost forever without his strategy, while Ellison believes that the business was so unprofitable that it was a disaster. Ellison says, “We’d sell all this stuff from Manugistics and i2 and the others and package it up with services. I’d say, ‘Ray, this doesn’t make any sense to me, we’re not making any money.’ And he’d say, ‘Well, Larry, it’s what customers want to buy.’ But Ray, the fact that customers want to buy it is not very interesting if every time we sell it we lose money.”

  From a sales perspective, Jay Nussbaum says, “The real danger in doing these smart things was that they were forgetting their mission. Their mission was to sell the damned database and sell applications, sell software, not create solutions that were one of a kind. Compensation became ludicrous because you’d be paying two or three people—salespeople, applications people, solutions people—for doing the same thing. There was too much energy being spent without enough margin.”

  Not surprisingly, Robert Shaw doesn’t think it was all bad. He says, “I see both Ray and Larry’s point of view, and I was in the middle of having to make it work. There was failure at Kellogg, but there was success in other places. I think Larry was being crowded by Jeff Henley to make money from applications. He kept saying this was a black hole; how much was it going to cost us to maintain these applications based on the contracts we were signing? We were making a ton of money on the consulting side, were also dragging database, so overall it was probably a pretty profitable business. But I think what Larry figured out was that when he started handing out the assignments to the development managers to take responsibility for integrating the suites, he kept seeing failures.”

  Jeff Henley had initially accepted that some margin erosion was the price to be paid for “playing the game and keeping some momentum.” However, “What turned out to be horrifying to all of us was that we underestimated what we were getting into. When everybody started wanting slightly different permutations of this or that best-of-breed combination, it became clear that supporting this stuff and migrating it was just horrific. We knew we weren’t going to make a lot of money reselling software—that was okay—but when we discovered the real implications, how much it was really costing us, we were horrified.”

  An example of how it was possible to see the same thing very differently at Oracle was something set up by Lane called the Max Account to deal with the cost overruns of implementing buggy applications or creating interfaces with partner software that hadn’t been delivered in time by development. Shaw says of his consultants, “We wanted relief in our budget because we were spending so much time building around the product to make it work in a customer site.”

  Shaw’s argument was that it was unfair for either the consulting team or the customer to eat the additional cost. Consulting managers were paid on the profit margin of the engagement they were working on. The idea was to steer cost they felt was not their fault into the Max Account. Lane expected development to cooperate, putting its hands up if it had failed to produce something that had been promised. According to Ellison, a consequence of the Max Account was that the consulting group dramatically improved its profitability: “The company’s profitability was deteriorating, but according to our accounting reports each of Ray’s groups was controlling expenses fabulously well. I didn’t understand how this could be possible. But I found out. Ray had a system where his managers could spend money and not be charged for it. For example, when Ray wanted to encourage training he told his managers they didn’t have to pay for the cost of training their people. Suddenly every sales meeting became a ‘training’ meeting. There were lots of ways to get expense relief. It was like the goddamn tax code. It created financial incentives for people to do what you wanted them to do. Unfortunately, like the tax code it was loaded with perverse incentives and people learned how to game the system and increase their profit bonuses.”13

  Another side effect of the Max Account was that it increased the already dangerous amount of finger-pointing. Ellison says, “The idea of people not having to pay for things that aren’t their fault sounds fine. But it meant we had to set an adjudication process to determine fault. ‘Oh yes, I’ve spent the money, but it wasn’t my fault.’ How much was your fault? Are you just pretending it’s not your fault? Was about seventy-three percent not your fault? The whole notion of finding, quantifying, and assigning fault within the company was perverse and poisonous.”

  Jeff Henley says, “Larry always felt like the Max Account was a dark hole. It’s a difficult issue. Ray was right that if it was a product problem, why should the poor consultants have to fix it for free? Ray wanted to expose the product problems to development. But Larry was right that if you give them an account to charge they’ll go crazy charging everything they can get away with.”

  If the best-of-breed strategy had started coming apart at the seams, Wohl’s other track, to deliver homegrown products with the features needed to catch up with SAP, wasn’t doing much better. Early in 1997, Wohl finally managed to push Oracle’s first full-fledged client/server suite out of the door; it was called Oracle 10.7 Smart Client (it sounded better than Fat Client) because all the program logic was located within the application on the client. It turned out to be something of a misnomer. Smart Client was both buggy and unstable.

  The problem, according to Byron Miller, an analyst at Giga Information Group, was that the software absorbed too many resources of the desktop client, with the result that performance deteriorated and the ability to run the applications in a distributed, multisite environment suffered badly. In other words, 10.7 didn’t scale and was likely to crash your PC. Oracle was not only late to client/server, it didn’t seem to be very good at it.

  Part of the reason for the inadequacies of 10.7 lay with the need to divert developers into the unending catalogue of problems thrown up by the effort to integrate the best-of-breed suites. But the larger reason was Ellison’s growing conviction, following his September 1995 speech in Paris launching the idea of the network computer, that client/server was not the way to go. It was, as he put it, “an evolutionary dead end.” Instead of defending 10.7 Smart Client’s shortcomings to irate customers, Ellison agreed with them, condemning client/server as “an awkward, expensive, and slow technology.” By bringing together the distributed complexity that was inherent to client/server with the attempt to achieve repeatability with a number of industry-specific, pre-engineered best-of-breed suites, Oracle, always with the best of intentions, had created a quagmire for itself. By 1997, the applications business was continuing its strong revenue growth and Oracle was a clear, albeit distant, second in the applications market. But nobody at Oracle thought that it had anything other than a continuing embarrassment on its hands.

  Worse still, the rows over applications strategy were having an increasingly corrosive effect within the company. Even after the release of 10.7, which was by far Oracle’s most complete suite to date, the sales and consulting organizations would frequently recommend that clients buy
third-party software rather than the homegrown stuff. Ellison claims that the sales force was equally compensated whether it sold Oracle applications or a partner’s products. Lane vehemently disputes this, saying that it got paid only on the royalty that the third-party vendor handed on to Oracle for making the sale—typically 30 to 50 percent.

  However, despite the weakness of Oracle’s applications—perhaps partly because of it—the consulting business was continuing to grow at a tremendous clip. By 1997, Oracle Services, which wrapped up consulting with support and education, accounted for half of all Oracle’s revenues. It was a success of a sort, but many within the company were becoming increasingly concerned at its side effects. Jay Nussbaum says, “We were content to grow the revenue number even if the margin was playing between sixteen and eighteen percent. As long as revenue was growing, we fed the monster. But the reality was that nearly all the margin was coming from technology licensing. We may actually have been losing money while consulting was growing at forty percent. We were taking on projects we should never have done because of their sheer size. It was alienating all the other distribution partners, but Robert was a take-no-prisoners type of guy.”

  Lane’s dissatisfaction with Ron Wohl was also becoming an obsession. He could not restrain his criticisms even when speaking to journalists and analysts. What made him even wilder was his frustration with Ellison. Although Ellison had given Lane control over nearly all of the company, development was still a no-go area. When Lane had proposed that Nimish Mehta take over from Wohl, Ellison ignored the idea. But he still seemed to have little interest in the nitty-gritty of applications, and from Lane’s point of view he seemed perfectly willing to let the drift continue. While Lane and Shaw were dealing with constant complaints from the field and from their relationships with major customers—which were far closer than Ellison’s—and learning how serious the shortcomings were, Ellison seemed fired up only when he was talking about Microsoft or how the network computer was going to take over the world.

  By the late summer of 1997, database license growth had slumped to little better than flat while applications had hit a wall. Growth in the previous quarters, artificially pumped up by sales of third-party products, had been respectively 60 and 95 percent. But with the wheels starting to come off the best-of-breed implementations, applications managed only a feeble 7 percent, while SAP had turned in an impressive 62 percent. If Oracle lost its second-place spot in apps to either the Dutch company Baan or PeopleSoft, which were both breathing hard down its neck, the long-term damage would be immense.

  The perception within Oracle, particularly among Lane’s people, was that customers had heard about all the problems with 10.7 and were sitting on their hands. To make matters worse, Ellison, who had bought his first megayacht the previous winter, the 192-foot October Rose, from fellow billionaire Kirk Kerkorian, had decided to spend most of the summer cruising the Mediterranean. The idea of Ellison sunning himself on his $10 million new toy for months on end (he says that he was actually away for no more than five weeks) while the company once more seemed to be imploding did nothing for employees’ morale.

  Gary Bloom, a fast-rising young executive vice president who had been running both the platform group and the partner programs and was about to take on management of the database team, says, “Larry had simply carved out a subset of the CEO job that he wanted to do and really disengaged, so Ray was pretty much running the company. . . . He wasn’t paying any attention to the engineering teams, the applications were getting further and further behind the competition. It was driving Ray wild.”

  Although Lane was quite happy to allow respectful journalists to take away the impression that he was now the guy in charge at Oracle, few within the company believed that he could solve its problems. Jeff Henley says, “I’ll always remember calling Larry and telling him that the company was blowing up and he had to get off his frigging boat.” When Ellison returned at the end of September, Bloom confronted Larry. “I said, you’ve got to decide whether you’re going to re-engage and run the company as CEO or not. If not, I said, he’d have to go and hire one. Oh, and by the way, Ray’s not it. Ray couldn’t manage the technology groups by any means, and he was frankly creating a scenario that was not effective for the company. A bit earlier I had also talked to Ray about it and said, ‘Your job is not to make Ron Wohl fail. You’re the president, you’re the guy everybody listens to right now because Larry’s not here.’ The real kind of negative turning point for me was when he went public, saying in front of huge audiences and Oracle staff that our applications are a disaster. That’s when it became crystal clear that Ray couldn’t be CEO because you can’t have a leader that’s out there trying to make other leaders fail. It just doesn’t work.”

  Bloom had an offer to become CEO of another software company. If Ellison wasn’t prepared to take the reins, he had decided to leave. “The job offer meant that I had a catalyst to get him interested in the topic. I told him that I wasn’t looking to leave but that I couldn’t work at a place where we were going to fail. I told him that Ray couldn’t run the company and that in fact many of the areas that he was running were out of control. He’s busy drawing attention to other people’s problems when he’s got a million of his own. I think he took it pretty seriously. He asked me to come back the next day to his house. When I arrived, Don Lucas was there as well, and Larry said, ‘I want you to have a private discussion with Don and to tell him your views without me there.’ ”

  At about this same time, Oracle was forced to issue its first earnings warning since the dark days of 1990. The moment markets opened on December 9, there was a rush to unload Oracle stock. Within hours, 30 percent of the company’s market value had been wiped out.

  Ellison soon faced what amounted to another ultimatum, this time from Lane himself. It was about the same issue as usual. Lane says, “The subject was about how Ron had to come out of his job. The thing that provoked us into having this specific meeting was that Barry Ariko, who was in charge of all American sales, and Robert Shaw had both threatened to resign. That’s actually not how they put it to me. They came to me and said, ‘Ray, sign up with us. The three of us will go to Larry and tell him that if he won’t replace Ron we’re all leaving.’ I said that I understood why they were asking me to do that but that as president of the company I couldn’t do that, I couldn’t give an ultimatum to the CEO like that. But I would carry their message to Larry. . . .

  “They were really suffering. They were saying that they were having to sell stuff that was subpar and that if we did sell it, it was the worst of all worlds because then we had to fix it and we had to try and deal with Ron, who would just ignore us. Randy Baker, who was running support, was living in sheer hell. . . . The bug counts on 10.7 were so high that it was an environment that was impossible to manage. This is the worst incrimination of Ron I can think of. At one time we had to stop development completely and put all of Ron’s staff on crunching bugs.”

  The meeting between Lane and Ellison took place in January 1998. Lane said that the situation with Wohl had to come to an end or Shaw and Ariko would quit. “I said, ‘There needs to be a Mr. Applications, somebody who can really go head to head with SAP.’ I wanted Larry to throw Ron out and hire a professional. I said, ‘You brought in me, you brought in Jeff Henley, why in hell can’t you bring in somebody to run applications development?’ ” When, as usual, Ellison raised the objection that actually finding such a person might not be very easy, Lane said he even had a suggestion: Robert Shaw should become Mr. Applications. It wasn’t clear who Shaw would report to. Lane said he could report to Ellison. Ellison had no doubt that Shaw would have continued to have Lane as his boss. Shaw says, “I knew about the proposal, and I was willing to do the job. It made a tremendous amount of sense. I also knew that Larry would never agree to it . . . never, never.”

  Lane insists that he was interested only in finding a way out of the impasse over applications. This was not, however, the way Ell
ison saw it. He believed that Lane was using the threatened departure of Shaw and Ariko to grab still more power. What’s more, in doing so, he had stepped over a critical line. Ellison had been willing to let Lane run just about everything else at Oracle, but development would always be his domain. Ellison says, “Ray asked me whether I thought he should run development. I said no, not really. I think he genuinely believed that he was better equipped to run development than I was. He felt the most serious problems in the company were in development. He had decided he couldn’t solve them without me out of the way and without him running development.” Jay Nussbaum puts it succinctly: “Ray had stepped into the sacred burial ground.”

  Ellison knew he could no longer placate or ignore Lane as he had in the past on the subject of Ron Wohl and applications. Jeff Henley says, “Things had changed with that last quarter. We’d set a record on Wall Street for the most amount of value lost in a single day, and the stock had gone on being pummeled into January. The board was pissed and Larry was embarrassed as hell by the way we’d fallen on our rear end. There were lots of things going wrong, like Ray’s disastrous sales reorganization. But the real recurring issue was the product.” However, Ellison was damned if he was either going to hand apps over to Shaw, whom he didn’t consider to be technically qualified, or get rid of the loyal Wohl.