Time to market can mean the difference between making your numbers and losing your shirt in our business, especially for PC hardware OEMs.
In the case of Microsofts Windows Vista, the delays stretched from weeks to months and then to years, and now it looks as if the company will bring out its flagship upgrade right in the middle of a product-cycle trough, potentially putting the companys partners through a period of deep pain.
I pored through some history to see when Microsoft said Longhorn would launch originally.
As some of you may have discovered, finding history in the Internet Age is fairly difficult, since the Web has no memory, and all those “inconvenient” facts can disappear from Web pages unless theyre archived somewhere.
Nonetheless, I found evidence that the company first said that Longhorn would ship in the third quarter of 2004.
Some quick arithmetic shows that Microsoft is likely to be just about two years late to market; much of the delay, I suspect, is based on how long it took the company to come to Jesus about its programming methodology. And the situation wasnt helped by its dysfunctional culture.
Before learning about the wholesale changes the company has finally undertaken to address the out-of-control Vista programming effort, I had a suspicion that a culture had been developing in Redmond in which workers could bring only good news to management or risk being shot as the messenger.
Weve all heard rumors of famous temper tantrums on the part of its top managers. Steve Ballmer himself once barked in my direction. So, Ive had a direct taste.
But a corporate culture can stand only so many dead messengers before live change grinds to a halt.
How far Microsoft has come since its early entrepreneurial days was brought home to me in spades the other day when I was scheduling some meetings in Redmond.
I offered slots at 8, 9 and 10 a.m. since I had a flight out of SeaTac at 1 p.m. But my contact warned me, “Frankly, youll have a hard time getting anyone here at 8 oclock in the morning.”
Remember, this is from the company where the halls used to smell funny because everyone was sleeping under their desks so as to put in 23-and-a-half-hour days.
And I know that its hard to stay lean and mean when youre a monopoly. There are only so many pull-ups you can do before flopping down in the chaise lounge for a massage.
After AT&T was broken up, one of its top managers whined to me one day about his changed circumstances.
“Dont let them fool you,” he said. “Monopoly was great. Wednesday afternoon golf. The whole bit.” Less golf; more mattresses under desks.
OK, I can accept the changes: the need for Microsoft to grow and mature, to settle down. And Im sure that the stock price growth hasnt really justified employees putting in those long hours recently.
So, whats a couple of years among friends, right?
The PC Life Cycle
Well, heres the rub: PCs have a life cycle of three to five years, averaging pretty close to four.
Most of the time, individuals and companies are busy minding their own business and the upgrade cycles are all random and overlapping, giving demand a pretty stable and steady tone.
But late in the last century, the industry fooled everybody into thinking that their PCs would blow up at 12:01 a.m. on Jan. 1, 2000.
Businesses responded, just the way the industry hoped they would, by buying, from about the fourth quarter of 1998 up through New Years Eve 1999, spanking new equipment that would be “compliant.”
This wave of buying —made coherent by the specific date of the millennium—had the effect of synchronizing everyones purchasing, if only temporarily.
The Millennium Upgrade led to the 2000 to 2001 trough. Everyone had new PCs and didnt feel it necessary to buy again for a few years. The recession didnt help the situation, stretching the cycle out perhaps longer than usual.
However, enterprises and consumers alike were back again buying in 2003 to 2005 in what could be named the Post-Millennium Refresh.
It was a purchasing wave somewhat less coherent than the preceding one, but it still had the characteristics of a wave.
The next wave will be still more dispersed, as individuals and companies respond to their own imperatives, and time purchases slightly differently. Lets call this next swell the Longhorn wave for sentimental reasons, and lets say that the newly established periodicity is four years.
Thus, we can expect the Longhorn wave to arrive in 2008. The history of Windows XP adoption shows that the consumer segment will cut over fairly quickly, since consumers buy what they find on the shelf, and OEMs will turn their consumer lines rapidly.
However, corporate customers hesitate, waiting for the OS to stabilize and taking time to qualify their new images.
If you estimate that Vistas Service Pack 2 wont arrive for 18 or so months following the 2006 release, its a fair bet that 2008 is when well see widespread corporate adoption.
The company is trying to make the Windows Vista launch a big event by tying as much to it as possible in the form of new hardware partners, applications and services.
But the release would have played much better for everyone if it had dropped at the peak of the wave in 2004, rather than floating in the middle of the trough in 2006.
Roger L. Kay is founder and president of Endpoint Technologies Associates.
Check out eWEEK.coms for Microsoft and Windows news, views and analysis.