Microsoft made one big, wrong decision that led to Vista's delays, Microsoft CEO Steve Ballmer told financial analysts during his meeting with them last week. The company took a Big Bang approach and tried to overhaul all of its operating system's core components simultaneously, an approach that eventually led to a fiery development crash.
"We made an upfront decision that was, I'll say, incredibly strategic and brilliant and wise -- and was not implementable," Ballmer said. "We tried to incubate too many new innovations and integrate them simultaneously, as opposed to letting them bake and then integrating them, which is essentially where we wound up."
Ballmer's blunt assessment of Vista's development problems came in response to analysts' skepticism about how well Microsoft can execute on its ambitious visions. Cranky about Vista's protracted incubation and the challenges a company Microsoft's size faces in trying to fuel continued growth, Wall Street has kept Microsoft's share price stagnant through most of Ballmer's six-year reign as the company's CEO.
Microsoft executives used their meeting to outline a blizzard of major initiatives, all aimed at furthering Microsoft's goal of being a "multicore" company with stakes in a number of diverse markets, including desktop software, online services and entertainment. Despite stumbles like its Vista execution problems, Microsoft has never failed with any of its major development investments, Ballmer claimed.
"We've been fortunate. There is nothing that we have undertaken -- with a couple of exceptions like Microsoft Bob that I'll cop to in advance -- where we have decided that we have not succeeded and let's stop," Ballmer said. "We've either succeeded, or we're still telling you we're going to succeed."
Trying to persuade analysts to trust Microsoft as it sinks billions into new "high growth" projects like Zune and its expanding unified communications and business intelligence software lines, Ballmer argued that the company won't repeat its Vista mistakes.
"I'm very confident that we have learned the one key lesson we needed to learn from that time frame," Ballmer said. He pledged that future development will happen more modularly and incrementally.
One Microsoft project manager said that's the message Microsoft is also spreading internally. At the same time as Microsoft met with analysts, it held an internal meeting with several thousand of its worldwide sales and field employees.
"We'll never again do a Windows update this big," said the Microsoft staffer. He expects future releases to follow the path mapped out by on-demand hosted applications, with small, frequent updates rather soup-to-nuts overhauls.
Analysts had mixed responses to Microsoft messaging. The company's stock dropped 2 percent on Thursday, the day of the meeting, to close at $23.87. It rebounded Friday to close up 1.6 percent.
One money manager at the event said he had hoped Microsoft would do more to demystify its Live strategy, which he views as little understood on Wall Street.
Industry analyst Mike Gilpin, of Forrester Research, echoed that view, saying he thinks Microsoft hasn't yet explained how its Live services will better those of its rivals.
"I would like them to explain the differentiation they'll bring," he said. "How are they going to nail Google?"
Building out Live and Microsoft's array of other new projects will be a long, pricey commitment. That has some financial analysts nervous. "Microsoft's business mix shift is changing as its underperforming businesses are growing faster than its profitable segments," UBS analyst Heather Bellini noted in a report to her clients on the conference.
But Microsoft's message to the world is that it's up for the challenges it faces, with its robust financial resources and in-house expertise.
Craig Mundie, Microsoft's head of research and strategy, argued that the industry at the edge of a sea change in software, one that will help Microsoft manage the issue of maintaining its ever-advancing software products. New research on formal composition methods for building software systems will better equip companies to deal with escalating complexity, as physical engineers and architects have learned to do with their creations, he said.
"Many of these questions of thinking that complexity grows either linearly or exponentially with the code base size will essentially be dealt with," Mundie said. "If that was true in other disciplines, you'd have a tough time building skyscrapers and bridges and other things, because they get bigger. We still manage to get them to be built and stand up."