For my first few visits, I whistle-stopped my way from town to town, seeking out executives, making my way up the myriad corporate ladders that converged on Bob Kavner’s office in corporate headquarters. At every opportunity, I wove a tale about the personal communicator as the successor to the cellular phone, and how the software that drove its user interface was as strategic an asset as the systems that connected the millions of long-distance calls made every day. “Today, dialing a call is like programming in assembly language,” I would say. “You punch in a bunch of digits, and it rings a bell at a specific remote location. With the personal communicators of the future, we will address people, not machines, wherever they happen to be. And if they aren’t available, you can simply jot them a note.”
After working through an endless series of gatekeepers, I finally secured an appointment to meet with a key strategic advisor reporting to Kavner himself. A trim and stylish man with the aspect of an Ivy League professor, he had as his assignment the development of a master plan for the future of what AT&T executives fondly called “the cloud”—the amorphous, ubiquitous web of wires, satellites, relay stations, and cells identified by a familiar audible bong followed by the musical rendering of “AT&T,” spoken by a sparkling female voice.
On the morning of my appointment, I was awakened at six A.M. by the distorted voice of an evangelist booming from the hotel’s clock radio, pitching his own personal formula for getting to heaven. I arrived half an hour early at AT&T corporate headquarters, a magnificent array of stacked, truncated pyramids tucked in the hills of affluent Basking Ridge.
A uniformed guard questioned me at the entrance to an underground garage, constructed as a series of connected caverns, with spiral ramps and gates requiring coded cards for access. “You’re going to the executive offices,” he exclaimed after finding my name on his list, as though this were an invitation to the Vatican. From the visitors’ rotunda on the ground level, I was ushered onto an escalator that rose past a waterfall to an atrium at the top. Here a polite young woman filled out a pass for me to carry.
Her perch looked out at the rear end of an enormous golden sculpture of Winged Mercury standing atop a globe. I noticed that he was naked. “How do you like staring at the butt of a Greek god all day?” I asked.
“That’s why I took this job,” she said, pointing me toward a bank of elevators.
Ascending to the top floor, I was greeted by an attendant at another way station, who directed me to a hostess who managed the fully equipped offices made available to visiting executives. She, in turn, buzzed me through a glass security gate and pointed me toward a magnificently appointed office, paneled in walnut and decorated with abstract art. “I was asked to tell you your meeting may start a little late. Please let me know if you need anything.” Here I waited.
The advisor showed up fifteen minutes late, apologized for being tardy, and announced that he had an important conference call in fifteen minutes that should take only a short time. He looked distracted as I launched into my pitch. He soon withdrew to take his call, only to reappear forty-five minutes later. Unfortunately, he was already late for another meeting, so he asked if I wouldn’t mind waiting for an hour and a half until it ended. He returned two hours later, and I continued my pitch. Within minutes, he got another conference call. This one lasted over an hour. When he came back, he told me that his wife had been waiting for him downstairs and he had to leave immediately.
I begged him to take a moment to compare our calendars for another open slot. As it happened, he wasn’t available again for more than a month. But I found a loophole: he would be in Tokyo ten days hence. Coincidentally, I was scheduled to talk at a convention of Toshiba’s marketing partners in Japan that week, and arranged to meet him at his hotel upon my arrival. When he left, I calculated that I had been waiting for him a total of five hours.
Ten days later, after the eleven-hour flight to Tokyo, the two-hour bus trip into the city, and a half-hour cab ride from my hotel to his, I called his room. It was already seven P.M. in Tokyo—two A.M. in San Francisco—and I was starting to hallucinate from sleeplessness. He said he would be down in fifteen minutes. An hour later, he showed up with his wife. We withdrew to a sushi bar, where I launched once more into my pitch while his wife looked on impatiently. By now I had become obsessed with delivering it. I waited for his reaction.
“I hear that IBM has the source code to your software,” he said.
“That’s right, but they have to keep it compatible with our APIs.” I could see that he wasn’t clear on what this meant. “I don’t think it would be an impediment to anything AT&T might want to do with us.”
“Thanks for the input. I’ll keep it in mind in our next cycle of planning.” He collected his wife’s jacket and placed it around her shoulders, then they left.
By mid-April, it was obvious to everyone that something was terribly wrong with the way things were organized. GO, EO, and AT&T Microelectronics had their own separate sales forces calling on the same customers, who were growing increasingly confused about who was responsible for what. GO and EO were each doing software development and courting ISVs to build applications for them. EO could not agree with Microelectronics on a common direction for future Hobbit development. The hidden cost of decentralized management and interorganizational communication was bringing the joint effort to its knees.
To address these issues, Dave Atkinson and Bill Warwick began proposing a concept they called the virtual corporation. Mainly by gaining a larger degree of control over EO and GO, and gathering key managers from all three companies, they hoped to streamline operations. Believing that Kavner supported their approach, they called for meetings to plan the details.
Behind the scenes, Atkinson knew that EO’s corporate culture would have to be redirected, so his first priority was to gain a greater degree of control over EO. He began discussions with Bernie Lacroute, EO’s chairman, to increase AT&T’s equity stake from 24 to 52 percent, on the assumption that this would enable AT&T to direct EO’s activities. But owning a majority interest in a startup like EO doesn’t necessarily result in control: the company still has to be managed as an independent entity in order to respect the rights of the remaining minority shareholders—something that AT&T was unaccustomed to doing.
The EO board accommodated Atkinson’s wishes and worked out a deal in which AT&T put up nearly an additional $40 million to increase its percentage ownership, giving EO a generous valuation of $150 million after the transaction. But instead of investing the full amount directly in the company so the funds could be put to good use pursuing the project, about half of it went right into the pockets of EO’s common and preferred shareholders. The reason for this unusual arrangement was that it cost AT&T less to reach its goal by purchasing existing investors’ shares than if it had purchased enough newly issued shares to dilute all other stockholders down to a 48 percent stake.
Kleiner Perkins agreed to sell half of the EO stock it had purchased barely two years earlier, for two and a half times as much as it originally paid for the whole lot. KP netted about $7 million—as much cash as it had invested in GO and EO combined.
The deal proved to be a bonanza for many of the individual stockholders, despite the fact that the company was far from turning a profit. Hermann Hauser of Active Book sold one third of his stock to AT&T, netting him several million dollars. The earliest GO employees, who had stoically suffered through five years of perpetual development, were aghast to learn that some of their colleagues who went over to EO had pocketed small fortunes while they were still toiling away. As one GO software engineer observed, “AT&T is the Beverly Hillbillies of venture capital.”
What AT&T didn’t realize was that it had just wasted its money, since EO maintained business as usual. But to those responsible for budgetary planning, it looked as though AT&T had just invested an additional $40 million in its personal-communicator initiative. So when Kavner’s staff began thinking about taking a similar position i
n GO, he wouldn’t hear of it.
Under growing pressure to focus GO solely on AT&T’s goals, Bill Campbell was increasingly concerned that Kavner might not be sold on the virtual-corporation concept. To allay Bill’s fears, Atkinson and Warwick arranged to have him meet with Kavner while he was on the West Coast for an EO board meeting. What Bill didn’t know was that the EO board had just rejected the contractual revisions and $3 million payment that GO and EO had recently agreed to—something he assumed Rossmann had the power to decide on his own.
Warwick ushered Bill into the main conference room at EO, where he took a seat at one end of the large conference table, opposite the end where Kavner was holding court. Warwick began the meeting with a detailed explanation of the virtual-corporation plan. Bill focused his attention on Kavner, who was squirming in his seat, plainly skeptical about the idea. Unaware of Kavner’s discomfort, Warwick turned the meeting over to Bill.
Bill aimed his remarks directly at Kavner, as though no one else were present. “Let me tell you what we’re trying to accomplish. We have three organizations represented here. In many cases we’re working together, but in many we’re working at cross purposes. We really believe that the AT&T umbrella ought to be over all three of these. They don’t have to be one organization, but to the outside world it has to look like we’re thinking and acting in unison, and that AT&T is fully behind it.”
Kavner looked suspicious, and quickly served up his concern. “I don’t think AT&T should own an operating system company. Why should we? What we want is strong, independent partners who can stand on their own and make money.”
Bill was getting annoyed. He had been led to believe that Kavner was already on board, but instead Kavner seemed to be accusing him of a hidden agenda. “Whether you want to own an operating system or not is up to you,” Bill shot back. “I’d like to make money on my own, take the company and go public. But I’m being urged by your people to devote all my effort to the Hobbit chip and to AT&T, flying to Japan eight million times and New Jersey eight million and one. But I’ll tell you why you should want to own an operating system. There are two key parts to the intellectual property, the processor and the OS, and right now you only own one.”
As the exchange grew more heated, heads silently followed the action from one end of the table to the other, as though watching a tennis match. Kavner got to the point. “It looks to me as though GO is running out of money. You’re trying to get several million dollars out of EO for some minor contract violation!”
Now Bill believed that somehow, somewhere, he had been set up. Unaware that Rossmann had brought the contract issue to the board, he thought it had been completely settled before he agreed to meet with Kavner, despite the fact that he hadn’t seen any money yet. Nevertheless, he didn’t think it was appropriate for Kavner to get involved in a matter of this sort. “That’s a business issue between GO and EO, and you should stay out of it,” Bill said. “I can see this meeting is a waste of time.” He abruptly ended the discussion and walked out of the room.
Atkinson was close behind, and caught up with Bill just outside the building. “I didn’t know anything about this!” he protested.
Bill laid into him anyway. “I don’t know what cave you’ve been in, Dave, but I could smell this from three thousand miles away. Kavner’s nowhere on this. Nowhere.” Bill got into his car and drove away.
For days after the meeting with Kavner, Bill refused to talk to Lacroute or Rossmann, despite several phone messages from both of them. Things seemed to be at a standstill until Lacroute convinced Kavner to call Bill personally. They had a tense but cordial chat, and Bill permitted discussions between GO, EO, and AT&T to resume. Since it was clear that Kavner wasn’t going to be a sugar daddy for GO, and we still hadn’t got the $3 million due from EO, we continued looking for alternative financing, such as the proposed investment from Taiwan. But on Tuesday, June 22, a bombshell hit.
Bernie Lacroute called to deliver the bad news. “Bill, Kavner’s gotten serious about ditching Penpoint and going with Newton.”
The mere suggestion seemed wacky, given Kavner’s public statements of long-term support. “What the hell’s going on?” Bill said. “Just last week he advised Matsushita to get behind Penpoint!”
“I don’t know,” Lacroute said. “But he’s asked his team—and us—to go down to Cupertino and check it out.”
Bill considered the consequences. “If that happens, we might as well close the doors. And it’s no picnic for you guys either. What on earth could make him even think of such a thing?”
“I’m not sure. But we have to get a proposal in front of him right away to sort out the GO-EO situation or I’m afraid we’ll lose everything. I’ve got a dinner with him next Monday. He’s willing to consider a plan to merge the two companies . . .”
Bill didn’t even bother to put down the receiver before making the next call. He had moles all over Apple, and by the end of the day he had assembled the whole story. Behind the scenes, things were looking pretty sour in Cupertino. Sculley was coming under fire for hobnobbing with politicians when he should have been paying more attention to business. Sculley reportedly met with Bob Allen, the CEO of AT&T and Bob Kavner’s boss, to discuss the possibility that the phone giant acquire Apple outright. But Allen’s staff had decided that Apple’s prospects weren’t rosy enough, so they began considering a less ambitious goal: a joint venture between Apple and AT&T in the area of consumer products. Sculley flew cross-country with Kavner on one of AT&T’s corporate jets to explore the idea of combining their efforts in this area, along the lines of the deals he had cut with IBM. Disappointed with EO’s and GO’s product delays, Kavner apparently wondered whether he should throw EO into the pot and switch AT&T from Penpoint to Newton.
Bill called an emergency board meeting for six o’clock Tuesday evening to consider the options. John Doerr was ill with the flu, and so attended by phone. Without being able to read Bill’s body language, John was less restrained than usual in proposing new partnerships, but the converse was also true: Bill was freer to express his reactions with animated gestures. After a while, the meeting looked like a game of charades.
Although no one liked the prospect of merging with EO, the board worked out a message for Lacroute to take back to Kavner: Yes, we were willing to talk about a merger, but even better would be a joint investment on the part of AT&T along with another partner, such as Novell or even IBM, or at least AT&T’s support in finding another partner.
Given Lacroute’s personal stake in EO, Bill wondered whether he could fairly represent GO’s interests to Kavner. Bill put the question to the outside members of the board. John and Vinod Khosla stood behind Lacroute without reservation. David Liddle concurred. “If Bernie agrees to put on his KP hat, he’ll wear it with Kavner.”
John, over the phone, attempted a bit of levity. “If we do wind up merging the two companies, let’s call it GEO and not EGO.” At that moment, David opened his Apple Powerbook, which was sitting next to the speaker phone, and it said “David’s got a headache” in a Monty Python-esque voice. It seemed he had replaced the usual startup beep with this customized announcement.
“Who the hell was that?” Doerr asked.
“David’s Powerbook doesn’t like your joke,” Bill said.
After the meeting, Bill called Lacroute at home, despite the late hour, to deliver the message. Lacroute sounded exasperated. “You don’t get it, you just don’t get it. Kavner doesn’t want options, he wants us to merge the two companies. It’s that or he’s going with Apple!”
“OK, OK, I hear you,” said Bill. “But Bernie, we both know there’s no such thing as a merger in Silicon Valley. One company eats the other, and only one survives. And under these circumstances, it’s important that we don’t look like we’re fighting among ourselves.”
Lacroute mistakenly took Bill’s comments as self-serving—that he was looking out for his own job. He decided to sound Bill out on running the merged companies.
“Bill, would you consider being the CEO?”
Bill took a deep breath. “No way. Rossmann can handle the job.”
Bernie was astonished. “Why not?”
“AT&T has already committed to making EO their flagship company in Silicon Valley. And besides”—Bill let the receiver drop around his chin—“life’s too short to work for AT&T.”
The following Monday, Bill and I flew to New Jersey with several other GO people to make a presentation the next day to AT&T’s global business systems division. The flight arrived late at Newark Airport, delivering us into the steamy summer heat. A heavy shower had just ended, and the sticky pavement seemed to glow black under the harsh bluish light of the streetlamps. Bill cursed loudly when our rental car wasn’t ready, forcing him to wait in a long line of sweating salesmen and charter-flight tourists.
Bill drove west like a maniac to our hotel near Basking Ridge. For some unfathomable reason, the New Jersey highway authorities were in the process of relabeling the exits on Route 237, with signs like “Exit 35, formerly Exit 33.” Unclear which turnoff we wanted, Bill took a chance, so we wound up backtracking through what appeared in the dark to be endless pastures. After dropping us at our hotel, Bill went careering off to the Short Hills Hilton, where he was to meet Lacroute after his dinner with Kavner.
The meeting was considerably more formal than Bill had imagined. Not only were Lacroute and Kavner there, but also Rossmann, Atkinson, and representatives of a number of other AT&T divisions. Bill cooled his heels in the lobby of the hotel, pretending to read while the fate of our company was being decided in a rented conference room above. Each time the elevator bell rang, he would quickly lower his magazine, raising it back up slowly when it turned out to be a false alarm. Finally, after midnight, Lacroute, Atkinson, and Rossmann emerged wearing broad smiles. “Let’s go sit in the bar,” Bill suggested.
Startup Page 31
