Tesler began his tenure with a thorough housecleaning. He asked AT&T to make changes in the Hobbit’s design and to produce a delivery schedule and price list. Finding AT&T’s response unsatisfactory, he canceled the Hobbit project and switched over to the ARM, to the horror of Bell Labs engineers and even some of his own staff.
Then Tesler did the unthinkable. At Sculley’s request he considered canceling Apple’s own software project and working with GO instead. The idea of buying technology rather than developing it internally was entirely contrary to Apple’s culture. This technical xenophobia was known in Silicon Valley by the initials NIH—not invented here. NIH must have originated at Apple, because by nature the company was incapable of adopting the term from outside.
To my delight, Tesler had hired a consultant to advise him on the future of the Newton project. It was Peter Miller, the person whom Mitchell Kapor had originally asked to join his team, and who by that time was a personal investor in GO. I helped things along with an unbeatable financial proposal for the licensing and control of our software, similar to the IBM arrangement.
But by the time Peter delivered his report, Tesler was facing a growing revolt on the Newton team. He sympathized with their desire to maintain internal control over the technology and identify this project solely with Apple. The prospect of buying a key component from the outside—particularly one that IBM had beaten them to—wasn’t very palatable, and the team also had concerns about the suitability of the GO technology. So Tesler recommended to Sculley that they pass on working with GO. Sculley reluctantly concurred, and the Apple team found themselves starting anew with a fresh mandate from management.
Although it seemed impossible, the pace of activity at GO continued to increase. There were so many urgent matters demanding my attention that I took to staying at a nearby hotel, to save the few critical minutes spent commuting from the city. There was no such thing as a weekend off.
Engineering schedules were slipping again, so we had a round of meetings to decide how we could remove certain features from the final product. Sue King continued to schedule review meetings. The last stock-purchase agreements permitted us to raise an additional $4 million within ninety days of the most recent investment, so I hosted an unending stream of potential investors. Following the success of the IBM announcement, many developers wanted to begin writing applications for our operating system, but most of them insisted on meeting with me first. I was needed for press briefings, recruiting key employees, performance reviews, reports to investors, and, most of all, for visits to other personal computer manufacturers that might follow IBM’s lead in licensing our software, many of whom were in Japan. Despite my constant efforts, I was falling behind. I called John Doerr for advice.
“John, I’m burning myself up. I’m working the equivalent of two full-time jobs. What should I do?”
“That’s easy, let’s split your job in half.” The logic of this simple observation was compelling. “Engage a headhunter and we’ll look for a chief operating officer. And get out and do something this weekend.” I took both suggestions.
Coincidentally, I had been invited to a dinner party at the apartment of a woman recently hired in our sales department who lived a few blocks from my condo. I accepted. After the meal, the guests decided to take a walk. As it happened, the woman had a friend who was giving a party nearby, on the way back to my house. The two of us trekked up Russian Hill—no small feat for someone with little sleep who hadn’t worked out for months. By the time we arrived at the top of the hill, I was out of breath. She pointed to a five-story walkup.
“What floor?” I asked.
“Five,” she said.
I shook my head. “No way. No way. I’m going home.”
“She’s got lots of female friends . . .”
“If I’m going to climb those stairs, I damn well better meet my wife at the top.”
We went on up. After catching my breath and checking out the desserts, I scanned the room with the practiced eye of a thirty-eight-year-old bachelor. My rusty radar settled on the host, a bubbly, slender woman named Layne. She had sparkling blue eyes and long blond hair pulled back in a ponytail reminiscent of a cheerleader, which in fact she once was. She worked in marketing at a PC company.
I introduced myself. Layne, in turn, introduced me to her boyfriend, an athletic-looking lawyer who worked at Next. As I left, I slipped her my card. “I know I look like a nerd,” I said, “but I’m only a partial nerd. If you ever dump that hunk, give me a call.”
Three months later, she did. The next August, Layne and I would be married.
“Seven thousand!”
The crowd was going wild. I had donated a one-hour piano recital—to be performed by me, in my condo on Lombard Street—to a charity auction for the Foundation for Educational Software. It was one of many crazy items, including a software code review with Bill Gates and an appearance as a character in a video game. The event was staged one balmy evening around the pool at a resort in Laguna Niguel during Agenda, the exclusive invitation-only conference held each September by the computer industry analyst Stewart Alsop. The top four hundred movers and shakers in the business were there.
Steve Ballmer of Microsoft was the auctioneer, and he had whipped up the audience after a fabulous dinner of Chateaubriand and cabernet sauvignon, hosted by an investment-banking firm. The items were going for staggering sums, a testament to the fortunes made around personal computers. I had boldly estimated the value of my contribution at $250.
“Eight thousand!” someone else yelled out.
Ballmer pounded his gavel. “Sold.”
“I better start taking piano lessons!” I yelled across the pool to the lucky winner, who slapped his forehead in mock disgust. The crowd loved it.
By this point, the guests were on their feet. John Doerr was on one side of me, Vinod Khosla was on the other. Vinod leaned over and spoke into my ear. “I hear that Bill Campbell might be loose in the socket at Apple.” Campbell, who had hugged Steve Sakoman in an Apple parking lot when I first met John Sculley, was the CEO of Claris, the applications software division of Apple Computer that Sculley had decided to spin off as a separate company three years earlier. But when Claris was ready for an IPO, Sculley decided that Apple should be in the applications business after all, so he killed the plan and folded Claris back into Apple. Campbell hadn’t made a secret of his displeasure.
“Do you think we have a shot at him?” I asked. Campbell had a reputation as a charismatic leader and an outstanding manager, but seemed more like a big-company guy.
“We’d have to make him CEO, if that’s OK with you,” Vinod yelled over the noise of the loudspeakers.
I thought for a minute. I wasn’t in this for the job title. In fact, I rather disliked the responsibility of managing a growing staff, which now numbered more than a hundred. Besides, we hadn’t turned up any good candidates for chief operating officer, despite many months of interviewing. “No problem—if that’s what it takes to attract someone of his stature.”
“You can have the position of chairman,” John said.
“Sounds good to me.”
“Sixteen thousand!” Ballmer called out. “Come on, everyone, this is the last item!”
“What’s the item?” I asked John. Instead of answering, he grabbed my hand and stuck it into the air, holding my bid card aloft.
“Seventeen thousand from Jerry Kaplan!” Ballmer was going hoarse.
“Oh, hell. What’s the item?” I asked again.
“Sold! A two-hour business plan consultation with John Sculley. Good luck, Jerry!”
I was stuck. Vinod and John were laughing so hard they could barely breathe.
As the evening ended, someone asked me what I planned to do with the time. “Break it up into five-minute segments,” I said, “and sell them to everyone who wants to tell Sculley off.” I was telling this to the wrong person, a widely read newspaper columnist. To my embarrassment, he published my remark the n
ext day.
By late October, a rumor was circulating in the company kitchen that our engineering schedules were going to be delayed again, possibly into 1991. This had become the midlevel technical managers’ standard technique of gauging the reaction to a schedule slip before actually proposing it. GO was rapidly gaining a reputation as a purveyor of “vaporware,” the industry’s term for products that are promised but never delivered. I had to take some action or we were never going to get a product out. So I called a meeting of the relevant managers: product marketing, quality assurance, documentation, operating systems, applications, technical support, hardware engineering, manufacturing, and public relations.
“It’s been three months since the IBM announcement, and all I hear on the phone is ‘When are we going to see a product?’ We can’t hold out on nothing but promises much longer. It’s time to set a product launch date.” I could see everyone freeze. “Carol, what do you think?” I already knew what she thought. She spent much of her time fielding calls from journalists who teased her with information about our potential competitors.
Her voice was strained and tense. “We just got word that Microsoft is going to announce in February that they’re planning some pen extensions to Windows. We can’t let them steal our thunder. We have to announce in January.”
I was hoping to build a consensus, but instead a revolt broke out. The operating system wouldn’t have enough test time. The application developers’ documentation, which had grown like a cancer to several thousand pages, wouldn’t be finished. There was no guarantee of how many demonstration machines we could manufacture by then. Only some of the promised applications would be complete. User testing indicated that the pen computer wasn’t yet ready for sale to end-user customers, only for applications development.
“I’m glad we’re all in agreement,” I said. A nervous laugh rolled around the room. “Look, gang, it’s time to bet the company. IBM is getting antsy again; Microsoft is breathing down our neck; the investors are anxious; and our reputation is fading. We can’t afford to delay any further, period. It’s time to stand and deliver.”
Carol interrupted. “I can get the Meridien Hotel in San Francisco for January 22, but only if we make a fifty-thousand-dollar deposit now. Then we can take a redeye to Boston, where the Boston Computer Society will let us present at their January 23 meeting.”
I looked around the room. Everyone’s face was white with fear, wondering if I had the nerve to make this substantial deposit to reserve the hotel. “OK, Carol, book it. Listen people, figure out what you can get done by that date. This is it—we’re either going to make fools of ourselves or take the high ground in the coming debate.” The room was silent.
“I’ll do my part.” It was Phil Ydens, the stoic operating systems engineer. “Whatever it takes, my group will be ready.”
“Same here,” said Celeste Baranski, the director of hardware. “Tell me how many units you want, and we’ll deliver.”
“Maybe we can quick-print copies of the key manuals for the launch. That’ll save us several weeks,” said John Zussman, the director of documentation.
It was working. With a fixed endpoint in sight, the team was pulling together. After some discussion, we agreed to pitch the product as a “developer’s release”—an early version aimed at applications developers—and to call the event a “developer’s summit.” There was one last thing to do.
“It’s time to name the product. Carol, show everyone our proposal.” She pulled out a large piece of foam-core board with our new product name and logo on it.
“Penpoint,” it said in large black letters. A line ran across from the e to the t, and an inverted red triangle hovered above the i, dotting it like the point of a pen. Underneath was a tag line: “The Pen Is the Point.” Everyone loved it. We hung it up in the kitchen as a battle standard.
“Just one thing to watch out for when you demo,” Carol said. “As you know, the handwriting recognition tends to run words together. Be sure and space out your words when you write the tag line, or you’ll wind up describing male anatomy.”
When word spread inside IBM that we had committed ourselves to an announcement date, it was open season on renegotiating our agreements. Out of the blue, I got a call from John Kalb, the IBM negotiator who had disappeared after my last meeting with Jim Cannavino. “We never completed the agreement to share in your revenues,” Kalb said.
“What agreement?” I protested. “Quinlan and I worked out a completely different arrangement!”
“That’s not Cannavino’s view. We simply haven’t finished the negotiation. He wants twenty-five percent of your revenue from others as an offset against our royalty payments.”
I was shocked. When I met with Cannavino in Armonk, I had proposed giving IBM a credit against royalties for a portion of our revenue from others—but only in conjunction with IBM’s buying 25 percent of GO at a high price.
“No way, John,” I said.
“Not only that,” Kalb said. “If the credit exceeds what we owe you, we get the difference in cash.”
“I want to talk to Cannavino.”
“That isn’t possible. Until you agree, there will be no more public or private support for GO.”
Later, I asked David Liddle to check this out with Cannavino. He did, and was amazed to confirm that Cannavino was quite serious.
We were screwed. And now Kathy Vieth, who was supposed to be our emissary on such matters, was unavailable, despite my understanding that she would call a number of potential Penpoint licensees and applications developers who were eagerly waiting to hear from IBM. That night I discussed the situation with Layne, who had an uncanny knack for guessing people’s intentions. “Don’t kid yourself, this is no coincidence,” she said. “Kathy’s been told to disappear, to put pressure on you. Remember, when push comes to shove, her allegiance is to IBM, not GO.”
I called a board meeting to deal with this emergency.
“You’re going to look pretty silly standing up there in January without IBM,” said John Doerr.
“In other words, we have a choice,” I said. “Go out of business now, or go out of business later.”
“Sometimes a rat has to gnaw off its leg in order to live,” David Liddle added.
“Sure,” I replied, “but rats don’t gnaw off their legs when they’re dangling by them!”
Vinod Khosla summed it up. “You have to make a deal one way or the other.”
I had only one piece of leverage, my rights under the “joint walk-away” provision in the license agreement. I knew IBM was committed to State Farm, and by now several other large customers as well. I called Kalb.
“John, as it stands right now, we have the right to cancel our contract with you and not deliver any further versions.” He checked the contract, and was surprised to discover that I was right. “So here’s what I propose. We’ll give you twenty percent of our revenue on a sliding scale, but only after we have income of fifty million dollars on Penpoint. And it’s only a credit against your royalties—no cash. In return, I’ll cancel GO’s right to walk away from the relationship.” After several weeks of further discussion, he consented.
These negotiations went right down to the wire: IBM was still threatening to back out a week before the January announcement. We had managed to avert the crisis, but I secretly knew that there was only one way for us to survive, and that was if IBM failed in the market, leaving its royalty credits unused. IBM had inadvertently made it in our interest to decrease its market share at the expense of other licensees, who would pay us real money in royalties.
After completing final press briefings for the dailies—the local newspapers plus the New York Times, the Wall Street Journal, and USA Today—I rushed over to the Meridien Hotel to check on the preparations. We were allowed access to the facilities only one day in advance of the product announcement, so Carol Broadbent had to lay her plans with the care of a police chief about to storm an orphanage held hostage by terrorists.
>
The cavernous main ballroom was filled with teams of specialists, each moving feverishly to complete their assigned tasks. Construction crews worked everywhere, rigging giant projection screens, hanging Penpoint pennants, planting row after row of chairs, and building a custom stage set. Following close behind were audio-visual technicians, securing wires with yellow-striped safety tape and adjusting the sound levels of numerous wireless lapel microphones, each labeled with the name of a speaker. The cramped area behind the stage looked like the control room of a nuclear power plant: pillars of equipment were set to project videos, live images and slides, and provide synchronized music. Stagehands reviewed minute-by-minute scripts in small groups, practicing their cues and sequences. Overseeing the fray was a director, in constant communication with each crew via a wireless headset, like an electronic coxswain.
But this was only one of a dozen rooms being prepped for the event. The two-hour announcement was to take place in the morning, then, after a buffet lunch, more detailed technical presentations were scheduled throughout the afternoon in the hotel’s smaller conference rooms. Maps and signs hung everywhere, directing foot traffic to these “breakout sessions.”
One room was marked “No Admittance: Charge and Load Station.” Inside, the GO buildmasters—who had taken the program modules submitted by the programmers and assembled them into a software “build”—were cooking up their final system and loading it into sixty-five spanking-new, sleek, black pen computers, wrapped in leather cases similar to my original maroon portfolio. The product-marketing crew had created an electronic notebook full of wonderful goodies, from traffic-accident forms to architectural drawings to memos recommending new recipes.
The rehearsals began at three, and I was up first. The director showed me my mark and video cue, then motioned me to the podium. I started practicing my speech. “Look directly ahead more, so our camera unit at the back of the room can pick up your eyes,” he said. When I finished, Robert went through the paces for his demo. Next up were presentations by the ISVs, the independent software vendors.
Startup Page 19