It took a while, but we tallied them all up, and here is what we got:
1. At what point would you say goodbye to socket A for your main system and say hello to a different platform (Clawhammer or PIV)?
10% Under 4GHz:
26% Over 4GHz
8% Would have to be cheaper than the prices listed
38% I only replace when my system sucks.
6% I only replace when I have extra money for it.
11% I only replace when I get the hots for a new machine.
You Are Quite Forgiving and Hardly Slutty
Stripped down to bare numbers, one might think these would give AMD great cheer, since half the population (the suckers and the hotties) would seem susceptible to marketing magic.
That’s why we don’t use a Java vote counter. It’s the comments you make along with your votes that are really what makes this surveys informative; they say not only what you voted, but why.
Read the comments, intentional and otherwise, behind the votes, and you get quite a different picture.
In general, for those who said their system had to suck before they replaced it, that system has to suck hard and sucking your face off before you’ll replace. While you’ll go with a cheap CPU replacement, you want to see at least double the nominal performance or more before making a major move.
Even those who say they only buy when they get the hots did not show much sign of nymphomania. Many if not most of those in the “suck” and “hots” categories who indicated their current rig were still using TBird systems, and not showing much sign of breaking down big-time.
If you forget the numbers and read the comments, this audience can be described as “fat, happy and cheap.” Cheap as in inexpensive, not cheap as in shoddy. You want to see a major bang for your buck, and it still better not cost a lot.
So you’ll buy a TBredB or Barton when you feel like it/when the price is low enough for you, then await future events. At some point, you’ll want a cheap Clawhammer.
Loyalty To A Low Price
2. What platform do you think you would buy?
32% Whoever gets to my buy point first.
Here’s another case where numbers can be deceiving.
On the one hand, 59% is nowhere as good as you might think. Remember, we’re asking socket A owners this question. This is like asking registered Republicans if they’ll vote for George W. Bush in 2004. This number should be skyhigh.
On the other hand, very many of those in the AMD and the “undecided” categories took pains to point out that it was price/performance (weighted heavily towards price) that would govern their selection.
In short, the only real difference between many AMD and “undecided” voters was that the AMD voters felt that Intel would never beat AMD on price/performance, and the undecideds thought it was conceivable. Certainly most of the undecideds considered bang for the buck the decisive factor.
This raises the most crucial point of the survey: just why are people loyal to AMD?
From the comments received, the overwhelming reason for AMD loyalty is simply and only that they sell processors for less than Intel.
There really were few “AMDroids” out there. Many AMD adherents were frankly critical of AMD’s recent performance. The logic was more, “AMD is on “my side” because AMD sells processors for less than Intel.” Beginning, middle, and end of story.
No doubt some of you will say “Well, DOH,” but look into this a bit deeper.
At least three-quarters of those who spoke about this listed price as the main or only reason for AMD loyalty. Some mentioned siding with the underdog, but practically no one (outside of a few dual-processor users) cited performance or any other non-financial reason alone as a reason for their loyalty.
Again, you may say, “So what?” but look at what AMD has been doing with pricing lately.
If AMD isn’t cheaper than Intel, most of the “brand loyalty” just melts away.
Of course, “cheaper” can be measured a number of ways. The average person responding couldn’t care less what an XP2700+ costs so long as he can buy a $50 or $90 processor that will do about as well or better.
But get rid of that $50 or $90 option, and the buyers get scarce. A few will nibble at the cheapest Barton, but just a few.
Athlon 64? They aren’t against it, but they’re certainly in no hurry. A .13 micron Clawhammer will have relatively few buyers.
Get it above a 4000+ rating (which looks like .09 micron territory), and people will start to show interest, provided the platform change is cheap.
Going To Intel
They usually cited non-performance reasons like stability or noise for their “defection,” and even those who cited performance cited the fear of AMD “not being competitive” at the time of their upgrade (i.e. no desktop Hammer).
What was interesting about those who thought their next system would be an Intel system was their relative insensitivity to minor differences in performance. So long as AMD stays somewhere in the ballpark with Intel and keeps the CPUs cheap, it’s going to take major differences in performance and minor differences in pricing for Intel to make big inroads here.
If Intel truly wanted to destroy the AMD fan base, all they would have to do would be to come up with a sub-$100 Prescott overclockable to 4.5GHz+ with no AMD equivalent.
AMD needs to make more money, but they’re not going to get more out of you. You’re in no rush, and can wait them out.
You don’t want the best CPU in the world, you want the best cheap CPU in the world. So long as it’s fairly close to Intel, and costs a lot less, you’re happy.
That makes you a nut too expensive for Intel to crack. Not so much you per se, but if Intel started offering $50 processors to you, it would start giving a whole lot of other people very, very bad (to Intel) ideas.
So Intel’s problem is they can’t afford to get you. AMD’s problem is they can’t afford to run the company off you.
That’s why AMD seems to be trying to do the exact opposite of what you want by offering new products at sky-high prices. They need somebody to pay a lot for their muffler, but who and how many?
No, it won’t be the Opterons, either. Opterons will be more like Hamburger Helper than a meal unto itself. They’re a nice profitable sideline, and certainly worth doing, but they can’t support the company all by themselves. The number of potential sales just aren’t there.
Unfortunately for AMD, they’ve never been able to attract customers with anything but price, even during the Thunderbird age when they had unquestioned superiority over the PIII.
Is increasing the price going to help that? Or will the average desktop and server buyer continue to hear just the Intel ding-dong, leaving AMD will alienated regular customers?
Not that Intel’s in any great shape in the long-term, either. People are getting resistant to the ding-dong, too. Celeron sales continue to increase their proportion of Intel sales.
Yet Intel continues its plans of spending more and more money to build faster and faster processors people see less and less need to buy.
In contrast, what people seem to want is smaller, quieter, and cheaper, and if that happens, then the CPU revenues and profits really go to hell.
Yes, you’re cheap, but what you have is growingly contagious, and in their own ways, both AMD and Intel are bucking a trend.