Robert and I appeared to be flushed with excitement, overcome by the significance of the event. In fact, our reaction was due at least as much to our relief that the unit hadn’t burst into flames, heightened by the dose of caffeine rushing through our veins. I handed the prototype to Norm Vincent, who inspected it closely. As he passed it around the room, I turned things over to Robert, who gave a tour of the application we had designed to process their automobile claims.
As Robert explained, the application used our notebook metaphor. The claims estimator would turn to a page for the type of car being examined and tap on the quadrant of the car where the damage had occurred. The application would then display an exploded view of the car’s relevant parts. By tapping with the pen on an individual part and checking a box to indicate whether it should be replaced or repaired, an automatic estimate of the job’s cost would appear on another page in the notebook. If there was a cheaper alternative replacement, or if repairing a part would cost more than buying a new one, the computer would flash a warning. The completed report, possibly including a scanned “snapshot” image of the damaged vehicle, could be transmitted to the State Farm mainframes. The audience recognized that our approach was simple, elegant, and practical.
But there was still one important matter to attend to: heading off the competition. I needed some way to extend the emotional impact of our presentation through the following day, when this same group would be wined and dined by IBM, their familiar and trusted partner.
“Tomorrow you are going to visit IBM. They are flying you to New York on a corporate jet; they will feed you a wonderful lunch; they will show you some exciting demonstrations. Their work is very good, and given enough time, I’m sure they can do a creditable job on your application. However, throughout their show, I ask you to bear in mind just one fact: there is a difference between a dog and a cat.
“Now, State Farm has mice in the house, and we have designed a cat.” I pointed to the prototype. “It is tuned precisely to meet your requirements. It’s the right tool for the job. But IBM has several dogs they want to sell you.” The audience started to giggle.
“They will raise the first curtain and show you that they can make a dog that looks like a cat. They shave off the hair, put it in a fur coat, and tack whiskers on it. Then they will close that curtain and open a second one, where they exhibit a dog that is trained to act like a cat. It eats cat food, and meows. Next, behind the third curtain, they will show you that they can fit a dog into a cat carrier. They will cut off the ears and the tail, amputate a leg, and sure enough it fits in the carrier. It can’t run very well, but it fits. Then they will have a noted lecturer explain their plan for breeding a dog that chases mice.” By now my audience was laughing out loud.
“Throughout this process, I ask only that you remember one thing: it is in their interest to sell you a dog, but it is in your interest to buy a cat.” Norm led the group in a round of applause.
On the way to the airport, Robert smiled as he gazed out the taxi’s window. “Dogs and cats.” He chuckled. “Critter would have been proud. I just want you to know that that was great. That was really great.” Coming from him, this was high praise indeed. Our respect for each other usually went unstated.
“Thanks,” I said. “So were you.”
The following Monday, I got a call from Norm Vincent.
“So,” I asked, “how did the IBM meeting go?”
“It was OK, but they couldn’t figure out why our people kept snickering. You really set the tone with that dog thing.”
“I hope you didn’t mind.”
“Not at all,” Norm said. “The presentation was excellent. You really set your cause ahead last week.”
“Thanks, I’m glad you liked it.”
“Jerry, I don’t know how you’re going to take this.” There was an ominous pause. I braced myself to hear that he had nonetheless decided to work with IBM or HR “But how would you feel about working for one of our regular suppliers?”
This was so far from my mind that at first I didn’t understand what he meant. “You mean you want me to quit and go work for them?”
“No, I’d like GO to team up with them on the project—that is, if you’re willing.”
This was the best of all possible worlds. Not only would we get State Farm’s business, but we could neutralize potential competitors by becoming its pen computer supplier. State Farm could open the door to a strategic partnership with a major computer company. With one big customer in the bag, we could likely get the partner’s support on a wide variety of fronts—like approaching other customers, manufacturing our products, promoting applications development, and most important, funding GO. I was thrilled. But there was one missing detail.
“We’d be happy to, Norm. But which supplier do you have in mind?”
“I’ll leave that completely up to you.”
As if things weren’t good enough already, he was indifferent about whom we worked with. I felt like a grateful teenager who had just been promised the use of a Rolls for the prom, only to be told that he could also take his pick of the town’s most beautiful girls to escort.
“You mean I can choose either IBM or HP?”
“That’s right.”
Now that I was firmly in the driver’s seat, there was no reason to rush things. I might as well relax and enjoy the ride. “I’d like some time to discuss it with each of them.”
“Of course. You should expect a call. Two calls, to be precise.”
When I arrived at the office the next day, there were messages waiting from senior salespeople at IBM and HP. I wondered which of them was the better dancer.
7
The Partner
AS WITH MANY of the world’s religions, it takes a bit of extra faith to fully accept that the IBM corporate culture is completely benign. It instills in its employees a sense that they are tough, fair, and among the chosen. Yet individuals are occasionally expected to do questionable things for the greater good, like a CIA operative ordered to defend democracy by assassinating a head of state. IBM has a tendency to create internal myths that protect its sense of order. If it is outwitted by some fast-moving upstart competitor, it is probably because someone didn’t play fair. If it doesn’t have the best products, it may merely be trying to protect its customers from unproven technology.
Like the Marines, IBMers are taught that they must be ready to do whatever it takes. The IBM “do or die” culture has evolved a curious strategy for dealing with its own occasional failures, a technique that might be called “do or lie.” Rather than accept a painful defeat, whole divisions sometimes continue their work as though nothing had happened—as though products had worked, as though customers had bought them, as though employees’ efforts contributed to the bottom line. Company executives simply declare that the battle is still raging and they are still in it, like the last desperate days of the Vietnam War. OS/2, IBM’s answer to Microsoft’s Windows, is a prime example of the “do or lie” technique. Well over a billion dollars had gone into developing this operating system, an investment that continues today with no realistic possibility of a return. The technique works well as long as enough money pours in to buoy up moribund projects. But by mid-1989, IBM was so laden with the walking dead that its highly profitable mainframe monopoly could no longer keep the rest of the company afloat. The problem was that nobody knew it yet.
Within hours of Norm Vincent’s passing the word, both IBM and HP had arranged to visit our modest offices to discuss potential teaming arrangements.
IBM showed up in force: close to a dozen people crowded into our conference room, led by Al Johnson. As the most senior IBM salesperson assigned to working the State Farm account, he knew that his most difficult task would be convincing his own firm that State Farm was serious about working with GO. “Do you realize that this matter got all the way up to the MC?” Johnson said to me before we got started. He was referring to IBM’s management committee, a handful of top executive
s who met regularly to resolve the most important problems facing the company. Given IBM’s mammoth size and far-flung offices, this four-man corporate equivalent of the United Nations mainly mediated territorial conflicts and coordinated initiatives so colossal that they crossed divisional boundaries. Most employees could work at IBM for a lifetime and never so much as shake hands with people that far up in the organization. Johnson was clearly impressed.
It soon became apparent that the participants were from different provinces within the company and had little desire to cooperate. The group included researchers from IBM’s prestigious T. J. Watson Research Labs who had worked on handwriting recognition for years; development engineers from the entry systems division, an enormous product-development organization responsible for IBM’s PC products; and, of course, the sales team assigned to keeping State Farm happy.
The meeting took all day. I later learned that this was standard procedure at IBM: a simple review takes one day, and if there is something substantive to discuss, it requires two. In contrast, GO’s meetings were usually scheduled in fifteen-minute increments.
During a break, Al Johnson took me aside. “Just between you and me, what’s it going to take to get you to bid this thing with IBM?”
I wanted to play it a little coy. “The issue for me is to avoid having to raise our next round of financing from other investors.”
“And how much would you be looking for?” Johnson asked.
“On the order of ten million dollars,” I responded. “We could take five as a loan and five as a straight equity investment.”
He didn’t blink. “And if I get this approved, will you notify Norm Vincent of your decision?”
“I believe the GO board would look favorably on such an offer.”
By the end of the day, Robert and I were exhausted by the endless pointed questions and tired of watching them fight among themselves. Johnson spent much of the time attempting to convince his people that State Farm was, in fact, intent on using GO’s technology and not theirs—a conclusion many found hard to accept.
By contrast, the Hewlett Packard meeting was a breeze. The earnest and plain-spoken head of the HP’s Bloomington sales office assembled a small collection of knowledgeable engineers and account reps. Their questions were relevant; their team obviously understood the customer’s needs.
Robert and I caucused following the meetings. “So, who do you think we should work with?” he asked.
I sat there and stared at him. “Well, let’s see. We have one potential partner who’s skeptical of our technology, divided on whether they should work with us, and resentful of our success with State Farm. The other one is cooperative, has no competitive projects to speak of, is headquartered down the road from us, and is anxious to work with us.” Now we were both staring at each other. We knew what we had to do. In the final analysis, what mattered was not how much we liked the partner, but how much weight they could throw behind our bid to establish our operating system as a standard. In this respect, there was no contest.
Robert broke the silence. “OK, OK, it’s IBM,” he sighed.
A common misconception is that business deals are cold, arm’s-length financial transactions between faceless institutions. In truth, business relations are much like personal relations: good ones are built on trust, respect, and mutual benefit, and bad ones are based on deceit, exploitation, and betrayal. Curiously, those closest to business transactions, the lawyers, are often the most oblivious of the subtleties of the process. Their professional aversion to taking risks is at odds with the essence of business, which is placing shrewd bets that may return more than the wager. They also tend to confuse “the deal,” the working understanding between two parties, with “the contract,” the written words that attempt to capture that understanding at a point in time. Words are good for capturing some things, such as the rules of chess, but not for others, such as how to ride a bicycle. What makes deals work are not the written words but the personal relationships between the individuals charged with making them work. And that’s why it’s very difficult to do business with IBM.
Rather than empowering the responsible party to make a deal, IBM assigns a professional negotiator, who usually knows or cares little for the substance of the agreement but has absolute authority over all of its details. The negotiator begins by assembling a list of interested internal constituents, all of whom are free to add new requirements at any time or block some minor concession if it doesn’t suit their needs. The result is like watching a photographer attempt to corral a group of hyperactive children into sitting still for a picture: if there is any delay in snapping the shot, some new disturbance inevitably erupts.
IBM’s deal-making culture developed out of one of the greatest legal battles in history, the thirteen-year attempt by the U.S. Department of Justice to break IBM’s monopoly on the market for mainframe computers. In the end, IBM settled the suit by consenting to some minor restrictions—such as not announcing in advance products that might chill competitors’ sales—but at a cost perhaps as high as $100 million. A legal staff of hundreds was employed over the duration of the dispute. This ordeal left deep scars on the IBM psyche, resulting in a preoccupation with secrecy, an aversion to putting commitments in writing, and an elaborate system of legal reviews for even the most routine business transactions. At IBM, every press release, speech, and product disclosure is subject to legal approval, leading to interminable delays. This culture is unsuited to the fast-moving environment of the personal computer business.
After agreeing to join forces with IBM, I expected to develop a working relationship with Al Johnson, sort out the details, then draft and sign an agreement. But now the only person who would return my phone calls was a man named John Kalb. And to my surprise, Sue King, one of the more critical members of the original visiting team, was put in charge of the project.
Kalb, thin and slightly stooped, a heavy smoker, with a wave of brown hair and a soft voice, set about to create a Byzantine agreement between our two companies. He approached his work with the meticulous care and uncompromising zeal of an artist, seemingly with little concern for deadlines, cost, or consequences. Under his dedicated stewardship, an army of lawyers, accountants, security officers, escrow agents, and file clerks labored long and hard to create a masterpiece of legalese. Kalb engaged the top New York law firm of Cravath, Swain & Moore to augment his own in-house staff. He resisted my pleas to get the deal done and let me return to managing the company.
Kalb began by reinterpreting the basic structure of my proposal: the $5 million loan was to be secured by GO’s intellectual property, which meant that if we failed to pay the loan back, IBM could take possession of our software, product designs, and patents. We were to receive the other $5 million as an investment upon the signing of a “comprehensive agreement” in which IBM might relabel and sell GO’s products, provide access to proprietary advanced technology, and contract to do our manufacturing. But, of course, there had to be a thorough review of our technology and an approved internaldevelopment plan before IBM would commit itself to the final agreement. “We can make you the Microsoft of pen computing, with IBM as your banker,” Kalb would say when I voiced an objection.
The next step was a series of review meetings with Sue King and her hastily assembled team. In contrast to Kalb’s grand visions, King saw us primarily as another supplier—one unaccustomed to meeting IBM’s demanding requirements. She protested that our schedules were too aggressive, our project milestones were insufficiently detailed, our test plans were inadequate, our product designs weren’t rugged enough. As hard as she was on us, though, her willingness to do battle with her superiors to promote the project helped get the attention and resources we jointly needed to succeed.
Morale at GO began to flag. A seemingly endless stream of calculator-toting IBM engineers would arrive for leisurely tutorials on object-oriented programming and pen interfaces. The visitors would take our engineers to lunch, where they would dar
kly confide that there was no way IBM would ever decide to back an operating system other than OS/2. As one of their minions warned me, “We can blacken the skies with planeloads of curious tire kickers.”
The most comical of the visitors, perhaps because of their gumshoe demeanor, were from the IBM security forces. In anticipation of providing us with confidential documents that never actually arrived, they came to review our security procedures—which were nonexistent. They designated a special locked room for IBM confidential material, containing an even more special locked file cabinet. Only I was to have the single key, and anyone allowed access to its contents had to sign a register attesting to what they had read. They even required us to place a chicken-wire barrier above the tiles of the drop ceiling, in a preposterous attempt to impede any intruder trying to crawl through.
Despite her lack of faith in our ability to deliver, Sue King forwarded our tentative delivery schedule for preliminary hardware prototypes to John Kalb. At our next meeting, he changed the rules again. As soon as I saw the new language, I exploded. “John, what the hell is this new appendix about breaking the loan into four installments, to be paid as we deliver working units to Sue?”
In response to my exasperation, he acted as if this modification were as innocuous as changing a comma. “She is concerned that you may not be able to deliver on schedule. This is just a way to focus your attention on her needs.”
Startup Page 14