Your letter was a great inspiration to all of us at GO. Be assured we will take your comments to heart.
Sincerely,
S. Jerrold Kaplan
Knowing Holli was concerned about confidentiality, I assured her it would be OK if she wanted to show the exchange of letters to some other people in the company before sending my response. When she gave me the final copy of my letter to sign, it had already been signed by every GO employee.
Making a business deal is like having sex: the more people are involved, the more difficult it is to consummate. Most negotiators prefer to make a single deal with one other party at a time—presumably a reflection of their personal preferences. That’s why Randy was just the guy to pull together AT&T, Active Book, Kleiner Perkins, and GO at the same time, to form a pen computer hardware company based on AT&T’s RISC chip, the Hobbit. Bernie Lacroute had graciously committed himself to serving as CEO for at least the first year, as long as Kleiner Perkins got to invest, and he decided to name the new company EO—Latin for GO. Bob Kavner agreed to represent AT&T on EO’s board of directors.
After weeks of intense negotiations, it looked as though the agreement was ready to be put to bed. Randy approached the deal with the firm but naive conviction that GO needed only an additional $10 million to get to profitability. This was based on an already outdated set of business assumptions and the questionable proposition that significant revenues would begin flowing from IBM and others by the fall. So he devised an arrangement whereby GO would get the needed $10 million as “advances on royalties,” in return for a source-code license to Penpoint and cooperation in porting it to the Hobbit chip. Upon signing, $5 million was to come directly from AT&T, and the rest was due in installments from EO after GO met certain development milestones. When the deal was done, GO would have its money without giving up any equity; a new, committed hardware partner impervious to Microsoft’s influence; and jobs for the people on the hardware engineering team. It all looked like a dream.
Believing that the agreement was complete, Bill left the office one balmy summer evening to go home and baby-sit for his newly adopted baby daughter. Randy was left to review the final terms on a conference call with the other three parties, who had gathered for this purpose at AT&T headquarters in Basking Ridge, New Jersey, the geographical midpoint of each of their offices. On the other end of the phone were Bernie Lacroute of KP, Hermann Hauser of Active Book, and Dave Atkinson, a tall, easygoing staffer of Bob Kavner’s who had been assigned to handle AT&T’s investments in new computer technologies. Normally, Atkinson’s relaxed, affable manner made discussions with him feel like a Saturday afternoon at a Pebble Beach clubhouse. But in combination with Lacroute’s tough negotiating style and Hauser’s concern that he might somehow be cut out of the deal, Randy had his hands full.
An adroit negotiator himself, Randy knew that all good deals come right down to the wire. Otherwise, someone has left something on the table—and he liked a clean table. It was nine P.M. in San Francisco, and as eager as he was to escape and get some dinner, he knew that the others must have felt worse: it was midnight in New Jersey. Hauser planned to return to England the next morning, around the same time as Lacroute’s flight in the other direction. They had resolved all the issues except one, a requirement that GO release future versions of Penpoint no later on Hobbit than on other platforms—that is, Intel’s. Aware of our symmetrical agreement with Intel, Randy realized that this would dramatically complicate GO’s product development, since both versions would have to be perfected and released in lock step. He was understandably reluctant to tie our hands.
“Sorry, but no way I can go with this,” Randy said. “What I can offer, however, is that we will tell you if we can’t meet this requirement and give you the right to finish and release the Hobbit version yourselves.”
Back in New Jersey, they were ready to explode, and Hauser lit the fuse. “What you’re telling us is that you’re not really serious about this, not really committed to Hobbit! You expect me to give up control of my company, throwing our lot in with you, but you can walk away if you feel like it?”
Lacroute echoed Hauser’s point. “No version parity, no deal. This is it.”
“Yeah,” Atkinson agreed, “this is it. This is really it.”
Despite the hardball tone, Randy concluded they were making a mistake. The right way to play this would have been for one of them to leave the room while the others loudly objected, so they’d have a graceful way out if they changed their position later. This was why Randy had let Bill go home.
He decided to stand firm. “Gentlemen, I guess there isn’t much point in talking further. I’m going to dinner now. Thanks for all your efforts.” To the amazement of the three men waiting to close the deal at the other end, he hung up the phone.
When I got to the office the next morning, I found Celeste Baranski and her husband, Mike Ouye—our directors of hardware and of software—barricaded in the office of our recently hired VP of operations, Paul Hammel. The three of them were to form the nucleus of the new ten-person EO engineering team, and had already negotiated employment packages with Lacroute. (Kevin Doren had decided to stay with GO, as we still needed some in-house hardware expertise.) We were sorry to lose Ouye on the software side, but he and Celeste were a package deal. I was the only person they would allow in.
“What’s going on, you guys?” I asked, squeezing into the room.
Hammel’s eyes were red for lack of sleep. They stood out against his chalky face, giving him the drawn look of an apparition. “You’ve got to do something. Randy’s being an asshole. He’s blown the whole deal!”
“How’s that?”
“Bernie called last night,” Hammel said. “He told me it was off, because Randy wouldn’t agree to maintain parity on the Hobbit. You’ve got to talk some sense into him.” The three had whipped themselves up into a paranoiac state.
Celeste was on the verge of tears, “Jerry, you of all people know that we’ve been in this together from the beginning. Don’t stand by now while Randy and Bill betray us.”
I headed down the hall to see what I could do. Randy was relaxing in Bill’s office, seemingly oblivious of the trauma he’d caused. “What’s the scoop?” I asked. “Our people are pretty upset.”
“Unfortunately, that’s just Bernie’s way of putting pressure on us,” Randy said.
“Maybe he’s just keeping his people informed about the negotiations.” Since I wasn’t the one on the spot, I could be more objective about the situation.
“Next, I expect he’ll call Bill.” Randy’s face softened a bit. “I hate to see them so worried, but it will all work out fine. Do your best to calm them down. Celeste won’t listen to me.”
Sure enough, within hours Lacroute had called Bill and worked out a compromise along the lines of Randy’s proposal, and the deal was done. But the hardware team never shook off the feeling that they were somehow cast aside, and this lingering suspicion was to taint GO’s relationship with EO.
11
The Switch
“IF I’M GOING TO get shot, I want it between the eyes, not in the fuckin’ back!”
I had never seen Bill Campbell so angry. He had just gotten off the phone with Kathy Vieth of IBM and stormed down to my office. He was so enraged his eyes were bloodshot. Bill could suffer just about any indignity with grace, except when someone had abused his trust. And Vieth had hit his hot button with a sledgehammer.
“Jesus, Bill, calm down.” I couldn’t imagine it was all that bad. After all, Kathy had been doing her best to support our cause.
“I just caught her in her car, leaving Apple.”
“What are you talking about? I spoke to her on Friday—she was back in Minnesota for a week, visiting her brother.” As a cost-cutting measure, everyone at IBM had been given the week of July Fourth off, a sort of forced vacation. Kathy had told me how nice it was to see her brother and how great the weather was back home. She had said that her beeper didn’
t work there, that it would be tough to get in touch with her, so not to try to contact her.
“Bullshit,” said Bill. “I’m telling you I just talked to her. She’s staying in San Jose!” By now Randy was in my office to see what the ruckus was.
Bill unloaded his tale from the beginning. “I get a call from one of my buddies at Apple, see, tipping me off that Cannavino and Vieth are in with Sculley. He’s pitching them on dumping Penpoint and joining forces with Apple on Newton, along with a bunch of other joint deals they’re making.”
“Oh, Christ!” Randy’s eyes narrowed. “Nah, I don’t believe it. Not after all the promises they’ve made to us. IBM’s difficult, but not devious.”
“So I figure what the hell,” Bill said. “I’ll call Kathy’s car. Sure enough, she was dumb enough to pick up the phone. You should have heard her—scared the shit out of her. I actually caught IBM red-handed, looking to sell us out.”
“So what’d she say?” I asked.
“First she denied the whole thing. Then she fessed up about the meeting. I asked how it went. And she just said, ‘I can’t comment on that,’ like I was some goddamn reporter. So I told her that if they were going to take a different path from us, that was fine with me, I couldn’t dictate their business direction. But I want her to tell me to my face. So I told her she had till the end of July to make up her mind. Then I told her I wanted to talk to Cannavino about this, and she just about jumped down my throat. She said, ‘Don’t do that, I’ll handle it.’”
Bill’s conversation with Kathy Vieth must have shaken things up a bit. No doubt she contacted Jim Cannavino immediately. Within a few days, Cannavino had called Mike Quinlan out of retirement to deal with us, presumably thinking that we were more likely to trust him than Vieth, now that her credibility was damaged.
Quinlan came in to smooth things over and, ostensibly, to brief us on the upcoming joint IBM-Apple announcement. He was dressed for golf, and said he was on his way to a family reunion in Hawaii. He took the credit for patching things up. “I want you to know that I spent many hours with the IBM team evaluating Newton, and I convinced Cannavino to stick with Penpoint. I told him that by the time Newton could ship, Microsoft would most likely have taken over the pen market. Also, we’ve already promised Penpoint to a number of key customers. Besides, with all the other joint agreements—a new chip, a new object-oriented cross-platform operating system, a new standard for multimedia—we shouldn’t try to boil the ocean in one shot.” He kept using this unusual phrase over and over again. These other projects ultimately became known as the Power PC, Taligent’s Pink, and Kaleida’s Script-X. “So I want to reassure you that we’re going to stick with GO.”
To retain our attention, Quinlan revived a discussion that had more or less died after he left to retire: arranging a sizable loan to GO as additional financing. “I can’t make any promises, but confidentially, things are a lot looser than they were the last time around. It looks like we can earmark perhaps ten million.” Randy looked as if he had just been offered money in return for a night with his wife. Nevertheless, he was destined to spend much of the next year in negotiations over the loan while IBM added ever more onerous terms.
After Quinlan left, Randy, Bill, and I regrouped in the conference room to compare notes. “It’s so hard to believe they would consider dumping us—behind our backs too,” Randy said.
For once Bill was the calmer one. “Actually, we’re better off now that we know what kind of partner IBM really is. If this happened once, it will happen again. We’ve got to close the AT&T deal right away.”
“You bet. I can’t wait to tell IBM to get lost,” Randy said.
Bill walked over and stood directly in front of Randy, leaned in dramatically, and put one arm on his shoulder. “It’s not time yet to burn any bridges. We need to keep IBM on the hook for the time being.” Randy sighed, nodding in agreement as Bill continued. “We’ve just got to suck their ass through a Flavor Straw until we don’t need them anymore.”
By October of 1991, the hoopla about pen computing had continued to escalate, despite the fact that there were no real sales in sight for anyone. After repeated delays, IBM and NCR had yet to release their pen computers for sale. The few prototypes they had in circulation were in constant use, on short-term loan to one company or another, creating an impression of overwhelming demand and shortage of product.
Once again, GO was running out of money, and the only real prospects were to agree to the IBM loan or to attempt another round of financing. Despite Mike Quinlan’s optimism, IBM was as difficult to deal with as ever. Randy spent endless days in discussion with a newly assigned negotiator. Every time Randy thought they were close to a deal, the negotiator would change the ground rules. In its latest demand, IBM wanted the rights to use our notebook’s “look and feel” on top of OS/2, DOS, Taligent, and Unix—for free. The negotiator argued that IBM was unlikely ever to use these rights, but the various development groups insisted that they be allowed to maintain a level of consistency across their product lines. IBM’s strategy was simple: wait until we were out of money, and then we’d have to accede to its demands.
GO’s alternative was to raise yet another round of financing from investors. We had just about tapped out the venture capital groups. After all, they had put up a total of $23 million, in addition to the $7 million from IBM and the $10 million from AT&T. To attract a new class of big investors, we had to spiff up our image and increase our public visibility. Months of going nose-to-nose with Microsoft and the announcements of several short-lived but highly touted competitors had reduced people’s perception of us: we were seen as just one more independent contender in a crowded field. Fortunately, fall and COMDEX were just around the corner.
In the hopes of finally crushing us while they had the home-court advantage, Microsoft allocated a large COMDEX booth to Pen Windows: perhaps fifty by seventy-five feet of real estate in a prime location. As probably the single largest exhibitor, Microsoft had plenty of weight to throw around. Despite its strong recruiting efforts, I was pleased to see that it didn’t have any more ISV applications to show than we did. And Microsoft’s were plainly of lower quality, mostly minor modifications to existing Windows applications.
The heavy behind-the-scenes pressure continued. I asked representatives of one hardware manufacturer why they ran only Pen Windows on their machine. They told me privately that Microsoft had offered them special concessions and was willing to provide a senior executive to speak at their product launch, if they worked exclusively with Microsoft.
Unable to afford a real booth, GO took a small private room off the main hall. The rent was $53,000 for the week. To have the room vacuumed once a day cost an additional $1,000. To turn on the electrical plugs—which were already installed, but disabled—the fee was $150 per plug. The electricity itself was extra. It felt like the last days of the Weimar Republic.
What we lacked in funds our crew made up for with enthusiasm and imagination. To save money, they carpooled for the twelve-hour ride from San Francisco to Las Vegas, then worked tirelessly from dawn to dusk. They canvassed the aisles and halls wearing sport shirts with “Ask me about Penpoint!” printed on them, handing out passes to hourly shows in our modest room. The presentations were consistently packed.
It was against COMDEX rules to give demos outside officially rented spaces, so our guerrilla demo squad, armed with Penpoint computers, would lure prospects to the nearest hallway to avoid detection. To attract interested parties, the guerrillas would ostentatiously whip out their computers and pretend to take notes on some exhibitor’s presentation. When they discovered a stubborn hardware manufacturer with a pen computer running only Pen Windows, they would borrow one of the manufacturer’s units and run it back to our room, where a bearded, irascible engineer sat all day behind a curtain loading Penpoint, like a short-order cook in an off-brand burger palace.
My mission was to get our product mentioned in the press. It wasn’t hard.
/> The first night, PC/Computing magazine held its Most Valuable Product awards. This lavish one-night Las Vegas show, which featured the magic of Penn and Teller, was broadcast on closed-circuit TV to the hotel rooms of the 130,000 COMDEX attendees. After being escorted to the show in a complimentary limo, I accepted one of the product awards on behalf of Robert Carr and the entire development team. Penpoint was praised at length in the magazine’s next issue, which reached nearly one million readers.
The next morning, I did an interview with Gary Kaye of ABC News. Then, after spending several hours chasing down various reporters, I developed a new strategy. Borrowing a press tag, I set myself up in the press room, where everyone eventually came to me instead.
The following evening, I was asked to go to the PC Magazine awards—not to be confused with the PC/Computing awards, although both monthlies were owned by the same publishing company. This time I had to wait in line for a cab, a forty-five-minute proposition at best. Luckily, I ran into Steve Ballmer and Bill Gates, who invited me to hitch a ride in their limousine.
Ballmer looked a little embarrassed at this apparent extravagance. “The hotel provides it for free,” he explained, “because we rent so many rooms from them.” On the way to the show, they were all business. To my surprise, Ballmer was not reticent about briefing Gates in my presence about what he had learned from hanging around IBM’s OS/2 booth. He listed, feature by feature and application by application, what worked and what didn’t. “They’ve done a better job than I would have guessed at getting Windows apps running,” he said. “But I still contend that full Windows compatibility is virtually hopeless in the long run.” This had become IBM’s approach: OS/2 would be a better Windows than Windows itself, since IBM had rights to the source code—for the time being.
Startup Page 25