THE LAST FOUR years of NeXT, following the demise of the IBM deal in 1992, were something of a tragicomedy. Steve tried so many different strategies that the company meandered without direction. He created a cheaper, pizza-box-shaped computer called the NeXTstation, but it never gained any traction. He and his team decided to design another model, based on a new microprocessor called the PowerPC chip—the same one that Apple’s newer Macs would be using. But eventually they decided there was no real market for the machine—not a single one was ever manufactured.
Mike Slade, who was head of marketing for a time, would occasionally think fondly of his old employer, Microsoft, “which was like the Yankees,” he remembers. “Going to work for NeXT was like being a starting pitcher on the 1998 Florida Marlins, a team that won, like, what—fifty games? In those days, Steve was kind of a forgotten guy. He was like Brian Wilson [after he had walked out on the Beach Boys], someone who had faded, a has-been. He became a pretty irrelevant guy in the high-tech world. Steve was in the wrong business now. He was put on this earth to sell to consumers, not corporate IT managers.”
Steve had great marketing instincts, but they were wasted at a company that had not created a competitive product. One day he told Slade that he wanted to “pick a fight” with Sun. So he had Slade ask two programmers to create a fairly basic database application, one using a NeXT computer armed with the company’s software, the other using a Sun workstation and Solaris, Sun’s implementation of Unix. Slade had their work videotaped. The NeXT programmer completed his task so much earlier than his counterpart on the Sun workstation that he had time to play a bunch of computer games. The video that they eventually released showed the Sun programmer muttering, as time ran out, “Um, I’ve just got a couple more things to work on.” NeXT followed up with eight spread advertisements in the Wall Street Journal, spending the company’s entire marketing budget for that year in one swoop. The result? “A shit storm of publicity, just as Steve had predicted,” Slade recalls, with a suppressed guffaw. Sun’s Scott McNealy went public, whining about NeXT’s “immature” marketing. “What people didn’t understand,” says Slade, “is that Steve could be just as brilliant when he had to think small. I came up with this elaborate marketing strategy, and he said, ‘Nope. The only thing that counts is picking a fight.’ And he was right.”
If he was brilliant at moments, he was still confounded by the ins and outs of running the company. His series of managerial miscues reached its climax when he hired a garrulous Brit named Peter van Cuylenburg to run the company’s day-to-day operations. The tale of PVC, as he was called, speaks to how unfocused Steve could be. Having impulsively decided that he really had to hire a president, Steve ran through a series of barely vetted candidates before turning to van Cuylenburg, a veteran of Xerox and Texas Instruments, who had rejected by fax a previous job offer. Steve professed adoration. “If I was about to get run over at a crosswalk,” he told the New York Times, “I would feel good about leaving [Peter] in charge of NeXT.”
Eventually it was van Cuylenburg who got run over, figuratively speaking. He had promised to give NeXT a clear strategy, but that’s not what happened. Zeroing in on details, van Cuylenburg found resistance from some employees who felt he was more interested in process than products. Worse yet, he and Steve seemed to often be at odds. Investors like Canon (that had invested $100 million in 1989) would complain that they didn’t know who was running the company—Steve or van Cuylenburg. The staff, too, got mixed signals—at least a couple of top executives believed van Cuylenburg tried to sell the company to Sun Microsystems without telling Steve. Van Cuylenburg denies this, and Sun’s CEO at the time, Scott McNealy, denies that the two companies ever came close to a deal. But there’s no doubt that the two did not make a successful management team. PVC was not at NeXT for long.
Shortly after he left, Steve, the ultimate “hardware guy,” made the painful decision to end production of the NeXT computer. The physical design of computers engaged him more than anything else, and he took great pride in the beauty and functionality of the machines he oversaw. But the sleek NeXT computers weren’t selling. Steve reluctantly shut down the hardware division, fired half the staff, and shunted the remaining hardware and factory assets off to Canon in a deal overseen by Jon Rubinstein. The Fremont factory building itself was put on the market to be leased out or sold as what it had been before—basic warehouse space. The original dream—that NeXT would create the world’s next great computer—was over. “We got lost in the technology,” Steve would later tell me.
There was no hiding NeXT’s failure, and there was no hiding the fact that NeXT’s failure was primarily Steve’s doing. This was the low point of Steve’s career. He was distraught over having failed, and, uncharacteristically, he let his disappointment show. One day, Ed Catmull read a NeXT press release about, he says, “how NeXT is really happy to be selling software to control government information servers, or data centers, or something mundane like that. I read this and thought, Oh, shoot, this has got to be killing Steve. So I called him up. We met at a Japanese restaurant in Palo Alto, and I said, ‘This isn’t you, Steve.’ And he went, ‘Ohhhhh, I know! I hate this so much. I mean, CIOs are nice guys, but God is this awful!’ ”
In public, Steve tried to portray this shift as a bold bet on the company’s software, especially its NeXTSTEP operating system, which had, he said, “no competitors.” But this time his sophistry was recognized as such by the media—and by those competitors, like Microsoft, who supposedly did not exist.
Steve did not shut down the entire company. Just as he had never given up on Pixar, he never quite gave up on NeXT. And just as he had at Pixar, he decided to play out two separate end strategies. He halfheartedly pitched the company to Sun (again), Hewlett-Packard, and even Larry Ellison’s Oracle, but nothing ever came through. At the same time, he kept pushing Avie Tevanian and his software team hard. Steve genuinely believed he had the sharpest team of operating system software engineers in the business, and he still hoped that the workstation world might embrace the NeXTSTEP operating system. So the software engineers kept beating the bugs out of it, and porting it to other microprocessor architectures, such as Intel’s Pentium family, or the PowerPC chip from IBM and Motorola. Steve worried deeply about finding a way to repay his investors, who had provided nearly $350 million in working capital. Not making them whole would have mortally wounded his credibility as an entrepreneur if he ever tried to start another computer company. So Steve waited to see where NeXTSTEP—and Avie’s crack team of engineers—would lead him.
By 1996, it began to seem as if their efforts might pay off in at least a modest way. Avie’s team had developed another software product that was drawing accolades. WebObjects was a tool for building commercial websites and other online applications out of modules of prebuilt code, called “objects,” that sped the development process and allowed the reuse of standardized components. This capability was especially helpful in building online stores, and the World Wide Web was now teeming with independent software developers and corporate coders building interactive websites with a commercial component. Business had grown so quickly that sales of WebObjects licenses now generated more revenue than NeXTSTEP. Finally, NeXT could truly say it was generating a small operating profit. Steve even lined up Merrill Lynch to back a potential IPO. Once again, a company of Steve’s had found its footing by transforming into something other than what he had intended.