Softwar

by Matthew Symonds

* * *

  1. LE writes: They waited, and they liked what they saw. Today, most of our fifteen thousand applications customers have very happily upgraded to the E-Business Suite. It wasn’t easy, and it wasn’t on time, but for our customers 11i was worth the wait.

  2. LE writes: POSCO is the best-run—most profitable—steel company in the world. When they told us they were getting $600 million a year in savings because of our E-Business Suite software, we got very excited. Many companies spend a fortune on enterprise software and get no measurable return on their investment. Now we had a large and respected company saying that huge savings were possible if you used an integrated suite. We saved a billion dollars a year at Oracle using our applications, and now we had a key customer confirming that the savings were real.

  3. LE writes: Now Siebel’s “hit the dust” too. Best-of-breed is dead—except for dog shows.

  4. LE writes: Most people don’t realize just how hard it is to build software as complex as the E-Business Suite. But I know. I’d been through it before with Version 6 of our database. Software megaprojects are very late—or they never finish. Ron and his team can be proud of finishing what they started. It was a major feat of engineering.

  5. LE writes: Maybe I’m mistaken, but I think Mark was frustrated because the ERP products were outselling the CRM products.

  6. LE writes: I agree. We have in fact moved to a more conventional consolidated management structure in development. We now have one person responsible for all of applications development and another person responsible for all of technology development.

  7. LE writes: After we hired Jim Finn, our media relations improved markedly.

  8. LE writes: This was a huge philosophical victory for us. When I first started telling customers and consultants not to change our code, it was considered a radical idea. Now everyone was beginning to understand the risks and costs associated with too much customization. Plain vanilla was becoming the favorite flavor for applications software. Suites were literally killing best-of-breed. We had gotten a few things right after all.

  9. LE writes: Of course Daily Business Intelligence cannot take the place of good management judgment; it simply provides better data so managers can base their decisions on hard facts rather than gut feelings. Those “hard facts” are very difficult to come by in most companies, because current computer systems are a bunch of loosely connected process automation systems that are terrible at providing useful information about the business to management. A single, unified process automation database with integrated Daily Business Intelligence moves a company out of the dark and into the information age.

  10. LE writes: Your best people want management to know how well they’re doing their job. Keeping relative performance a secret is not fair to them.

  11. LE writes: Not quite. Customers still have the option of buying the applications à la carte, which they do about half the time.

  12. LE writes: As we started to put in Daily Business Intelligence, I realized that Oracle had the wrong product-quality metrics; we had been measuring the wrong things for years. Oracle had held engineering accountable for product defects—bugs—rather than service requests. A product can be bug-free and still generate too many service requests because it’s difficult to install or difficult to use. We decided that our primary measure of product quality would be the number of service requests, not the number of defects or bugs. Once we got the metrics right, the squabbling between support and development ended, the number of service requests started trending downward, and customer satisfaction went up dramatically. Daily Business Intelligence is a powerful tool for changing behavior, but be careful of what you measure. If you measure the wrong things, behavior will change the wrong way.

  13. LE writes: Not all the analysts. A few good ones actually survey customers.

  25

  A PERFECT STORM

  For Larry Ellison, Thursday, May 2, 2002, was meant to be a routine working day. Through the morning he had worked on the deck of his home in Atherton, taking the usual mix of calls from his senior executives, business partners, and customers. Much of the rest of the day, as had been the pattern for several months, was scheduled to be spent in the eleventh-floor boardroom at 500 Oracle Parkway hammering out problems with the newly unified applications development team. But by late afternoon, Ellison found himself engulfed in what soon became known within Oracle as the “perfect storm”—a vicious combination of political feuding driven by large egos and the impending gubernatorial election, a fashionable post-Enron obsession with “corporate wrongdoing,” and a hyperaggressive media that had worked itself into a feeding frenzy. It was the worst crisis the company had faced in more than a decade.

  The telephone call announcing the storm’s arrival was from Ken Glueck, Oracle’s Washington-based senior political operative. Glueck is a big, unflappable man whose calm and geniality were never threatened even at the height of the “Dumpster diving” brouhaha (when Glueck had helped to expose Microsoft’s secret bankrolling of purportedly independent industry lobbying groups that just happened to take the software giant’s side during the epic antitrust case). But now Glueck was agitated.

  Alerted by an Oracle lobbyist in Sacramento, Glueck had been watching live TV coverage of the California Highway Patrol raiding the state’s own Department of Information Technology (DOIT) “to secure all shredders and trash.” The news bulletins were also carrying reports that DOIT’s director, Elias Cortez, had been suspended by Governor Gray Davis pending a criminal investigation and that the governor’s e-government adviser, a longtime political aide named Arun Baheti, had been fired. Baheti had accepted a $25,000 campaign check from an Oracle lobbyist just as the state was completing a big software contract (worth between $95 million and $122 million depending on whether it ran for five or ten years) with Oracle. Although there was nothing wrong with Oracle’s making a political contribution, Davis had issued orders to his administration staffers not to accept money personally. It was clear to Glueck, Ellison, and Safra Catz that Oracle was now embroiled in a major political scandal. Ellison told me later, “I could see a headline with words like ‘police raid’ and ‘shredders and trash’ with ‘Oracle’ right in between. I didn’t look forward to being served up as the tasty part of that media sandwich.”

  Increasingly ominous looking clouds had been gathering for some time, but Ellison hadn’t noticed them—it had never occurred to him that Oracle had done anything wrong.1 The origins of the crisis had been more than a year earlier. California’s DOIT had hired Logicon, a consultancy and systems integrator that was a subsidiary of Northrop Grumman, to look into the state’s future IT requirements and propose a cost-effective solution. Logicon had recommended a six-year so-called enterprise license agreement (ELA: a kind of heavily discounted “all-you-can-eat” licensing arrangement sometimes favored by big corporate and public sector customers, in which they pay a lump sum for an unlimited number of database seats combined with long-term maintenance and support) with Oracle, with a further four years’ option. The state and its 104 agencies were already big Oracle users, and Logicon reckoned that it could leverage their combined buying power to get a very good deal out of Oracle for both California and itself. Which is exactly what Logicon and the DOIT believed they had gotten. If the deal was allowed to run its full ten-year term, they calculated that the savings compared with buying Oracle database licenses on a piecemeal basis were likely to be around $110 million. It also meant that agencies could bring as many users online as they wanted, when they wanted—a big incentive to developing innovative and cost-saving e-government initiatives.

  Why Oracle was willing to do the deal is another matter. It was getting paid only $30 million to allow up to 270,000 people to use its software for ten years, while support and maintenance, which included free upgrades to the latest versions of Oracle’s technology, were pegged to a flat rate that worked out at $5.2 million a year for the life of the contract. Unknown to Oracle, there was an escalator cla
use in the maintenance part of the contract that would give Logicon $12 million for doing little more than man a help desk. Logicon was also doing well out of the transaction: from the $52 million the State of California was paying for licenses, Logicon was getting a $10.7 million sales commission, while the remainder was accounted for by sales tax and the charges of the leasing firm that had been brought in by Logicon to finance the deal. But from Oracle’s point of view it made little financial sense. Ellison later complained to me, “It was a spectacular deal for Logicon and a good deal for the state, but it was not a good deal for Oracle.”

  Even at the time, plenty of people at Oracle, including Ellison, thought the contract was far too generous and an example of precisely the kind of “scorched earth” deal in which future earnings were sacrificed for the sake of making a quick buck in the present, which Ellison and Catz had actively campaigned against. If Oracle itself had negotiated the contract, the terms would have been very different. However, Jay Nussbaum, still the head of Oracle Service Industries, reasoned that although the State of California had been give very substantial discounts on Oracle products, there was something to be said for an arrangement that would help lock in a big customer for the foreseeable future.

  But if the deal was controversial at Redwood Shores, it was rapidly becoming political dynamite a hundred miles away in Sacramento. The DOIT had a good many enemies among Sacramento legislators and, facing a reauthorization that many thought it would fail to get, Cortez was desperate for a big IT win to prove its worth; Barry Keene, the director of General Services, whose department was responsible for negotiating the contract, was eager to get the publicity that a deal of this size would bring. Sure enough, Cortez, Keene, and Arun Baheti had all appeared, grinning from ear to ear, on the pages of Info Week. Predictably, the pictures, and others like it in a host of publications that exist to massage the egos of senior government employees, were, to the DOIT’s many critics, like a red rag to a bull. At the same time, issues began surfacing that were troubling to even neutral observers.

  The Department of Finance suggested that neither it nor the other offices involved had exercised proper diligence in determining whether the cost savings that the ELA was meant to bring were probable or were based on spurious estimates provided by Logicon after it had been hired as a consultant. The accusation was supported by the haste with which the deal had been put to bed and the ambiguity of Logicon’s status, which stemmed from changes in its commercial relationship with the state. Logicon was now receiving all its remuneration from commission both as a distributor for Oracle and for arranging financing with a leasing firm. Consequently, there were legitimate doubts about how disinterested its advice had been while in its original consulting role.2

  And that wasn’t all. Suspicions were further raised because the contract had not been put out to tender. At the last moment, Cortez, determined to wrap things up as quickly as possible, had insisted that Oracle, rather than Logicon, be the prime contractor. Oracle was one of a select group of companies on a General Services Administration list that qualified it as a sole-source vendor. Ellison says, “The best place to buy Oracle software is from Oracle Corporation.” Up to a point. The problem was the suddenness of the switch from Logicon to Oracle and, once again, uncertainty about what Logicon was bringing to the table and how much it was being paid. Finally, there was the nasty little kicker of the $25,000 check that Ravi Mehta, a political consultant on Oracle’s books, had handed over to Arun Baheti while the two men were having a drink in a Sacramento restaurant. At the very least, the timing of the payment, the details of which had only just emerged, was unfortunate.

  For Dean Florez, the Democratic chairman of the state Joint Legislative Audit Committee (JLAC), the target was too big and too tempting to ignore. Florez, a political maverick who saw himself as California’s answer to John McCain, had a burning desire to make life as unpleasant as possible for Governor Davis and his administration. What Florez wanted and what Florez got was a full review “of the State’s contracting practices in entering into an enterprise licensing agreement with the Oracle Corporation.” The surprising thing is that nobody at Oracle seemed to realize that this was a portent of serious trouble.3

  It was months later, toward the end of 2001, before even muffled alarm bells at Redwood Shores started ringing. State Auditor Elaine Howle was a long way from completing her report, but Oracle executives were beginning to worry about the kind of questions they were being asked. Belatedly, it occurred to them that Howle might be set on reaching a hostile conclusion. In January, fearing the worst but still convinced that the state had no grounds for complaint, Oracle took the precaution of asking Howle’s vastly experienced predecessor, Kurt Sjoberg, who was now in private practice, to carry out his own “independent” report on the deal to see if, in his view, California would, as Logicon had claimed, save a lot of money. On April 10, Sjoberg Evashenk Consulting reported that over the full ten-year life of the contract the state stood to save between $110 and $163 million. In other words, Logicon’s estimate of the contract’s value had been respectably at the low end of the range. Surely even a skeptical Elaine Howle would have to admit that California was getting a pretty good deal after all.

  Just six days later, Howle reached a very different conclusion. In a stinging report, the bureau argued that over six years the state would lose $40.6 million (although it admitted that this would be reduced to $5.6 million if the contract ran the full ten years). The state auditor made no mention of the campaign check from Oracle, but in just about every other respect, it was as bad as it could be. She accused DOIT of ignoring the results of its own survey of 127 state departments, undertaken prior to signing the contract, which had provided evidence that “few state workers might need or want any new Oracle Corporation products.” All three departments responsible for overseeing large IT projects were lambasted for failing to assess the state’s actual need for the contract. She further inveighed against “General Services’ unprepared and inexperienced negotiating team” for agreeing to a contract “that left the state unprotected against numerous risks.” “Lacking an in-depth understanding of whether the ELA might fill a legitimate need for state departments, and without knowing the true costs and benefits of the contract, the state committed millions of taxpayer dollars to a questionable technology purchase.”

  It got worse. Howle argued that because of Logicon’s “undisclosed role, actions and compensation,” questions had been raised about the validity of the contract. Howle calculated that Logicon stood to make no less than $28 million out of the $95 million the state would be paying for the first six years of the contract. Legal advice to the audit bureau suggested that despite Oracle’s status as a vendor that could be exempted from competitive bidding requirements, the same didn’t apply to Logicon. Given that it was taking 30 percent of the contract’s value, a court might judge the contract to be void. But for Oracle, it was the second half of Howle’s report that was most damaging; it read like a sustained attack on the software firm’s business practices. The report claimed that Oracle was well known for:

  • Aggressively selling to the highest levels of an organization by basing its arguments on the “positive impact on the customer’s business” rather than the technical details of the database and the competition;

  • Using high-pressure sales tactics to close long-term, high-value deals quickly—saying, for example, “If you buy databases for ten computers this year and promise to buy databases for fifty computers over the next five years, we’ll give them to you at the special rate we have now. Prices are going up; it’s the fourth quarter, and we’re ready to deal.”

  • Practicing the Oracle maxim, “Lock customers in, lock competitors out,” by getting from customers a long-term commitment that encourages migration to Oracle products and helps establish Oracle as an organizational standard. The high cost of future transition away from Oracle products to those of competitors discourages future competition.

&nb
sp; • Changing the way it licenses its products, making long-term commitments problematic. For example, the most recent version of the Oracle enterprise database software, Version 9i—released two weeks after the state executed the ELA—has certain features that are separately licensed. If the state wanted to upgrade from 8i to the 9i version and also wanted these special features, it would have to pay a separate license and maintenance fee for them.

  The analysis continued, “According to our technical consultant [who had also provided the insights above], other Oracle business developments that were occurring shortly before the contract was finalized included pricing pressure from its customers. In 2000 and the first half of 2001, Oracle’s users and prospective clients were putting tremendous pressure on the company to lower prices. Oracle’s database market share was being threatened at the high-end (large, complex systems) by IBM and at the low-end (small, departmental systems) by Microsoft.” The findings of Howle’s “technical consultant,” it emerged, were being assisted by none other than Ellison’s old enemy—the Gartner Group, probably the technology consulting firm most consistently hostile to Oracle. Not surprisingly, Gartner “questioned the wisdom of establishing Oracle as a standard for the entire state.”

  The auditor’s report was a nasty shock to Oracle. Ken Glueck said to me later, “The section of the report about our business practices simply couldn’t be reconciled with what had happened. Our salespeople were not engaged at all in the negotiation of the deal, and the four guys from our side who were involved in it, you would instantly realize if you saw them, are just not capable of aggressive tactics. And besides, Kevin Fitzgerald, who had succeeded Nussbaum, kept telling us that the terms of the contract were the most favorable he’d seen during the twenty years he’d worked with the public sector. But the way Howle’s report was written provided a lens for looking at the contract that was intended to make us look bad.”