Softwar

by Matthew Symonds

  Ellison goes on to predict that previously high-flying Internet software firms, such as Ariba and Commerce One, which are already suffering from the sudden downturn in demand for e-business applications, will not survive. “These aren’t one-product companies,” he asserts. “They’re one-feature companies. Ariba has a little piece of the procure-to-pay flow—Internet purchasing requests. It’s just a tiny component part of a procure-to-pay system.” Oracle has already drawn attention to Ariba’s difficulties with a nasty piece of copy on its Web site entitled “Ariba Derci.” Siebel, by comparison, has a full CRM suite, but Ellison claims that it’s the product of acquiring other software companies and putting a Siebel label on their products. “It’s complete, but it’s not integrated. It’s a lot easier writing checks than writing software.”

  The reporters are politely interested in the “CRM in ninety days” message; the next day there is widespread, albeit fairly muted, coverage. But it’s clear that they see it more as smart marketing than as anything fundamental—although Carleen Hawn of Forbes asks an interesting question about the impact on Oracle’s business model of switching to a high-volume, low-price approach. (Ellison’s answer is that the cost has never been in the software but in the labor needed to put it in—Microsoft got it right, the enterprise guys got it wrong.) They’re also pretty used to hearing Ellison banging on about Oracle’s “war on complexity” and the dreadful consequences of fragmenting data, even though he’s always funny and they can’t help laughing at his jokes.

  What they most want is to get Ellison talking about the economy and its effects on the competitive landscape. Marc Boslet of The Industry Standard (a tech magazine in Silicon Valley that at the height of the dot-com boom was running to more than two hundred glossy, ad-packed pages but is now in what turn out to be terminal financial difficulties) asks in what way the B2B market has changed. It’s time for Ellison to start enjoying himself. “A lot of wild things were going on,” he says. “People were watching the stock market more closely than they were watching their businesses. Commerce One was giving big customers stock in their company if they bought its B2B exchange software. Companies were buying Commerce One products because they expected to make a lot of money on Commerce One stock. It seemed a brilliant plan because every time one of these big deals was announced the Commerce One stock would go up. On top of that, some of these companies would then spin off their purchasing departments and call them B2B exchanges. They thought they’d get rich taking their purchasing department public. It was crazy.”4

  Another question, from the San Francisco Chronicle’s Kelly Zito, is about whether Oracle is finding itself squeezed in the database market by IBM (the following day, IBM announces that it’s buying an old but much-weakened survivor from the brutal database wars of a decade ago, Informix). No problem, says Ellison. Oracle 9i with real application clusters runs SAP or Siebel five times faster than IBM’s DB2 or Microsoft’s SQL Server, with ten times the scalability and one hundred times the reliability. “RAC [Real Application Clusters] is the most important new technology from Oracle in the last ten years. It’s game, set, and match in the database business,” he crows. The briefing is already running half an hour overtime, and the Oracle PR people are trying to pull Ellison out. Appropriately, one of the last questions is about the battle between Siebel and Oracle. The previous week, Tom Siebel claimed that he didn’t often see Oracle win big new accounts when the fight was on. Ellison is both outraged and amused. He says, “Very interesting. Possibly Tom needs an optometrist.”

  The meeting that Ellison is late for is the regular Monday afternoon PDMC (product development management committee) meeting attended by the heads of the key engineering teams. It starts off with fairly routine pricing and licensing issues. Among them is a proposal that Oracle’s e-mail server (which, Ellison says, is faster and more reliable than Microsoft Exchange because it runs on top of a database) should cease to be sold as several separate products. Oddly, there’s a feeling that telcos that license Oracle Internet Directory “product” in the millions at $5 a seat might not like the change. “It makes no sense to price every feature as a separate product,” says Ellison. “But that’s the way the enterprise software business has been for a very, very long time.” Another mystery to Ellison is the poor reception of a proposal from Ron Wohl to stop talking about exchange software and to refer instead to “hub-and-spoke” software. Wohl convinced Ellison to make the change, and Ellison has been using the new terminology for several weeks. Unfortunately, nobody else agrees. Ellison says jokingly: “Okay, okay, I give up. I promise never to say ‘hub’ or ‘spoke’ again. I just want to get out of here before midnight. What’s next?”

  The next item on the agenda is rather more serious. It turns out that one of the “fast-forward flows” that Oracle is formally to launch the next day and that Ellison has just previewed with the press doesn’t work as it should. It’s a package called “incentives to performance” that is designed to allow members of a sales team to calculate their compensation precisely. The software has to deal with a fiendishly complicated set of variables, determining rules for who should get what as a lead is turned into an order and eventually becomes a sale. In the approved manner, Oracle has been using its own business and processes as a laboratory for developing its applications.

  Ellison starts off by asking detailed questions about how the software maps and automates the set of interactions that begins with sales forecasts and ends with accounts receivable. He doesn’t like what he hears. Always suspicious about the ability of sales organizations to extract more than they should by counting the same sale more than once or by spreading the commission around to too many people, he concludes that the detail isn’t there. Suddenly he asks crossly, “Just what is it that we’re announcing tomorrow?” The tension in the room is mounting. Is Ellison going to pull tomorrow’s announcement? One of the team involved in the “CRM in ninety days” initiative is licking his lips nervously and starting to sweat despite the air-conditioning. He agrees that the product isn’t ready. Ellison says with rising incredulity, “You want to announce this, but not deliver until May . . . can anyone tell me why we’re doing this? Why should we announce it now? Why shouldn’t we wait until it works?”5 It’s a nasty moment, but Ellison is assured that the other flows are ready. The decision is made to pull the sales force compensation package until the problems are resolved and go ahead with launching the other flows on Tuesday.

  • • •

  When the day is over, Ellison and I drive the half mile to a canalside house he owns in Redwood Shores, where we can sit on the deck and watch the pelicans. Ellison’s earlier belief (sincerely held) that Oracle alone might miraculously escape the recession sweeping through the rest of the technology industry now seems ludicrously optimistic. What really happened during those last few days in February at the end of the quarter? Was it a question of two or three big deals falling through at the last moment, as Ellison had earlier appeared to suggest, or was something more fundamental going on?

  “No. It was a lot of deals—dozens, I’d say. Sometimes a deal was delayed. Sometimes the size of the deal dropped from $6 million to $2 million. All of a sudden people stopped buying in anticipation of their needs for the next couple of years. Instead they bought exactly what they needed for the next three months. There were even cases where companies that were already using our software were trying to find ways not to pay for it. I’m not going to mention names, but a very large wireless company that was supposed to give us $15 million or $20 million this quarter, based on their own published subscriber numbers, decided to dispute the amount. It was just a negotiating tactic. If we lowered their bill, they’d pay us this quarter. We didn’t. Lots of tech companies and telcos are under financial stress, so they’re being much more careful with their cash these days. The dot-com die-out is getting worse. Suddenly, moving all your business processes to the Internet doesn’t seem so important anymore. It seems like e-business and the Internet a
re going out of fashion.”

  I suggested that perhaps companies were no longer worried about Web-based “stealth competitors” coming from nowhere to eat their lunch. “That’s exactly right,” said Ellison. “Wal-Mart is no longer worried about Amazon.com—if Wal-Mart ever was worried about Amazon. A lot of companies have realized they don’t have to move quite so quickly. They can be more deliberate and automate at a more rational rate. Flying the Internet flag isn’t so cool anymore.” Perhaps another factor was that simply making an announcement about what you were doing on the Internet was no longer likely to have a major impact on their stock price? “Absolutely,” replied Ellison. “All these B2B exchanges . . . there were lots of companies that issued a press release about a B2B exchange and saw their stock price shoot up. Commerce One dished out about three of these B2B press releases for every one that we sent out. They killed us in the press release war. But that exchange business is gone now. And it’s not just the exchange business that’s dried up. Companies are much more cautious about buying any kind of new technology.”

  It is almost as if Ellison were talking about some semi-mythical period remembered only through the mists of time. I can’t help reminding him that in mid-February, only ten weeks ago, he still thought Oracle would make its number. “Absolutely,” he says. “We went into that quarter extremely confident. We got off to a great start. If you look at our quarter month by month, we had great growth in December, we were still way ahead in January, and then we hit the wall in February.” When did he first know that something big had changed, and why didn’t Oracle’s own much-vaunted e-business systems provide any warning? “Our sales-forecasting system tells us how many deals are in the pipeline and multiplies their value by a historic close rate. If there’s a change in close rate we detect that, but in our case we don’t detect it until the last day or two of the quarter, because that’s when we close so much of our business. I hate that. Consulting firms actually go out and train software buyers to negotiate right up until the last hour of the last day of our quarter before signing a multimillion-dollar deal. Most software companies will increase the discount on the last day just to get the deal into the quarter. It’s a monster the industry has created by mapping its sales compensation plans to its fiscal reporting. I think that’s a bad practice, and we’ve tried to put a stop to it at Oracle. But that hasn’t stopped our big customers from asking—and taking us right up to the end of quarter before they sign.”

  I asked Ellison what he had expected the reaction to be when he issued the earnings warning. He said, “Well, the reaction had already occurred. The stock had already gone down dramatically. It was one very loud ‘I told you so.’ It turns out that the analysts were looking at the macroeconomic forecast and that proved to be more accurate than our internal systems sales forecast.” So how bad did he think things were going to get, and how did he plan to adjust to the new circumstances? “Well, first off, I thank God that we spent the last two years reengineering our company into an efficient global e-business. If we had just started the process of moving to the Internet now, we would really be in trouble. Yeah, we were disappointed in the quarter, but it wasn’t a disaster: our sales went up, our margins went up, and our profits went up. I’m not happy, but I’m not going to go out and kill myself. We’re really in pretty good shape. We’re making a lot of money, and our products have never been stronger. We have release 9i of the database coming out soon. We have a cool new version of the application server that’s going to drive BEA nuts. Our products are competitive, and our organization is efficient. Those two things allow me to sleep at night. The last time the economy slowed down [in 1991], Oracle lost money.”

  Even if the product cycle was in good shape, it was clear from the “CRM in ninety days” initiative that the marketing message was changing with the times. “That’s true. Without the same sense of urgency about moving to the Internet, we have to sell differently. What we’re selling now are very rapid, low-cost, low-risk implementations of automated business flows. The E-Business Suite contains several separate business flows. We can automate one flow at a time—like “procure-to-pay” or “opportunity-to-global-sales-forecast” and so on. We provide the software and the labor to get it up and running for a guaranteed fixed price, so customers get a guaranteed return on their investment with a minimum of risk. That’s what’s selling these days—return on investment. People have lost their appetite for risk.

  “Selling separate flows solved another key problem for us. By talking about the E-Business Suite as complete, people thought that you have to buy and install the entire suite all at once. It’s an all-or-nothing kind of deal. No, no, no. Not true. But our competitors used it against us quite effectively. So we thought that we had better start talking about the separate E-Business Suite flows inside the suite. That also lets us sell better against the best-of-breed specialists. We’ll sign a contract guaranteeing to do a CRM flow in ninety days for X number of dollars. We tell our prospects to go ask IBM and Siebel how long it will take and how much it will cost for them to do the same thing. It was the best way we could think of to drive home the point that Oracle applications install faster and are much less expensive to run than the competition. Even today many companies make the incredibly expensive mistake of failing to ask for a guaranteed implementation cost and guaranteed annual operating cost of the applications systems they are about to buy. We’ll guarantee everything. The competition won’t.”

  Would Oracle have gone down this route if the economy hadn’t gone bad? “Well, marketing the flows separately is a good idea with or without the downturn. We marketed the complete integrated E-Business Suite as a cost-cutting, productivity-enhancing engine. Now we emphasize the fact that you can install the suite a flow at a time. So we have two applications marketing messages. The first message is, we have a complete integrated E-Business Suite—we saved $1 billion at Oracle using it. The second message is that you can install one business flow at a time, quickly and at a guaranteed price.”

  A big part of Ellison’s 11i campaign has been centered on the idea that systems integration was to be avoided at all cost. Yet the assumption behind being able to install the software to manage a particular flow was that it was being integrated at some point with legacy systems. Didn’t that mean that he was backtracking on something pretty fundamental? “Yeah, the ‘Just say no to systems integration’ slogan was poorly thought out on my part. It wasn’t systems integration that had to be avoided; it was code modifying that had to be avoided. Of course, some systems integration may be necessary. So wake up! Try a little harder!”

  Perhaps another reason for the flip-flop on systems integration could be the damage that was being done to Oracle’s relations with the big consulting firms, which frequently recommend what software their clients should buy. Ellison said, “I don’t think so. We still think it’s important to minimize systems integration. It’s just not always possible to eliminate it completely. ‘Say no to systems integration’ was right some of the time, but it wasn’t right often enough for it to be our primary message. ‘No code modification’ is the correct message. If you’ve integrated our accounting application to your custom telephone-billing systems, you can easily move to the next version of our accounting system so long as you haven’t modified our code. If you’ve modified our code, you’re stuck.” And the integrators, the consultants? “Now that we’ve cleaned up our message, the integrators that compete on price love us because we install faster and cost less to operate than the competition. The integrators that do the giant time-and-materials projects still don’t like our message.”

  How was Oracle doing in efforts to win a massive deal with AT&T that Ellison had recently been pitching in New York? “They’ll make a choice in very, very short order. I think we’re ahead. We’re the only one that can do the business flows. By the way, we did a very similar business flow for their ADSL division. The business we’re trying to get now is fixed-line. They’re trying to decide whether they’
re going to do the flow with us or take a more traditional, functional piece of sales automation from Siebel. We’ll know in a week or so.” As is often the case with Ellison, he was being overoptimistic about the time frame. The negotiations with AT&T dragged on interminably as key executives on both sides left and the decision-making process at the troubled telecom giant stalled. Apart from the sheer size of the deal, winning it would create a vital reference point in Oracle’s struggle to catch Siebel.

  How was the fight against Siebel, the market leader in sales automation and CRM software, going on other fronts? What about Tom Siebel’s remark, quoted earlier in the day, that he hardly ever saw Oracle competing with CRM on major accounts? It’s the kind of dismissive claim that incumbents always like to make about the pretensions of challengers, but it had seemed to needle Ellison. Obviously Mark Barrenechea lived and breathed the idea of one day destroying Siebel as a serious rival, but what about Ellison? Did he get up in the morning thinking about how to take the fight to Siebel? Ellison replied, “At one point I ranked Siebel as our most important applications competitor, because they’re the largest best-of-breed software company and they have a close partnership with the largest best-of-breed integrator, IBM. I thought we had to beat Siebel for the suite approach to triumph over best of breed. But even now the outcome of that war is now fairly certain. We’ll end up with a couple of suite companies, and SAP will be our only major competitor.”