Rude, your analysis is pretty much right on. The 5 to 1 ratio is
darn close as well. I'm not a financial analyst or IT manager but do work with those guys and have a strong sense of the complexities involved with asset management. Needs are diverse. You don't necessarily gain a competitive advantage waiting 6 mos for prices to drop by 30% or by buying something cheaper/slower with reduced functionality for your company-wide intranet server or mechanical engineering department workstations.
My argument is/was with those who don't understand why anybody would purchase leading edge technology. Their argument being it's "stupid" to pay those ridiculous prices.. people are wising up and they are not gonna buy it anymore cause all they need is a 200Mhz machine to do WP/Email... therefore Dell is doomed. All predicated on
their own usage model. Talk about narrow mindedness and ignorance about how these machines in the hands of others, are re-shaping their lives...
Those of use who work in the industry know the bear argument is flawed. There are plenty of acceptance drivers for power. Price of the box is not as huge a deal to a corporation as it is to price sensitive individuals when all other costs are factored in . Heck, less than two years ago we bought Sun Ultra2 workstations for 25k a piece. Now, we can get more powerful Wintel boxes for about 5K. At the time however, we had no choice but to upgrade and stay with Unix as none of our CAE vendors had successfully ported to NT. Now they have.
I contend that Unix growth is dead, will shrink and Wintel workstations will experience healthy growth and tremendous profitability.
I personally like Unix however.
Another statement you made, "The demand curve is not flat and peaks at technology that is 12 to 18 months old.." is right
on the money.
It's interesting to note that at the leading edge of the demand curve where volumes are lowest, profits are the highest. Dell focus is marginally skewed from their competitors in that they focus more intensely on the left side of that curve. Where technologies are completely mature, they are not as aggressive. This strategy is the only logical choice as long as Moore's law remains intact.
When innovation and advancement slows waaay down or stops then this
game is over. PC's become commodities and the pricing structure will
collapse. Unfortunately for the bears, the technology roadmaps I'm working with paint a different picture. If anything, the rate of
hardware advancement is accelerating... driven by none other than our friends at INTEL.
New bus structures, signaling levels, clocking schemes etc. are providing the means to design beasts that will blow away peoples current concept of PC power. Moores law states that speed/power doubles every 18 months. 24 months from now, leading edge PC's will be roughly 3-4 times more powerful than todays hottest offerings. That's acceleration. It's not just speed mind you, which in and of itself is only a moderate enticement, but capabilities and standards that are non-existent or exclusive today.
Are we going to need/want machines that much more powerful that soon?
Some of us will, we always will... there are'nt enough MIPS in the
universe to satisfy us for the work we have to do. There are many more like us who design everything consumers use that is electronic
or has a moving part.
Simulating thousands of permutations of design possibilities before
comitting dollars to physical prototype... that's where you'll find
most early adopters. How much power and what kind of software does it take to design one of those new, high-tech ergonomically friendly fancy toothbrushes on the market today?? I think the bears would be
quite surprised to how important simulation and 3D modeling is to
the development of just about everything we come into contact with
daily that we don't give a second thought.
The rate of innovative change in everything from diapers to satellites is accelerating by virtue of the power of our beloved PC.
One final point however. Some are under the assumption that the IT department or finance guys make these hardware purchasing decisions
at all the big corporations for all departments. This is not the always case. In newer/leaner organizations, the department manager owns the budgetary responsibilities and makes the call as to what gets purchased and when. If the ECAD manager decides the department will need 30 new workstations in 98 then it's budgeted for, the budget is approved at the beginning of the year and the money gets spent at some agreed upon window of time during the year. Generally, there can be budgetary concessions with finance but rarely are purchases of less capable hardware an option because no intelligent individual would compromise the efficiency and productivity of their department. After all, their review is tied to departmental performance and when you spend that kind of dough, you want every last ounce of power to justify the purchase for as long as possible.
You also want happy CAD engineers right? Just one example of autonomy and how some of these bad boy machines sneak
their way into a company.
IT has only a minor influence on a decision like this and it's generally concerned with networkability and managability. Speed, power and price are not their concern. Servers however, that's another story.
So to summarize, the bears are WRONG. There IS a market for leading
edge, it's healthy, it's the most profitable and it's not going
to shrink or go away.
MEATHEAD |
