Although I understand your line of reason, I think you are overlooking the fact that Apple's slice of the $1000+ market is more like 90%.
Go back and take a closer look.... those figures are
workstation only, which is what the MP is. So those figures exclude the consumer systems, premium or otherwise, which is what the linked articles are referring to (laptops, iMac and so on /= workstation). Even server sales are excluded from those figures, which are higher than workstation sales.
I could not find more recent reports, but if the general trend of the last two years is anything to go by, I'd expect it to be fairly similar or better now. As such, your assumption (Apple's 10% over-all marketshare = similar workstation marketshare for the MacPro) is flawed.
Since you cannot give Apple a larger share than they actually have (greater than the overall marketshare value I listed), it's the best-case scenario. In reality, it's not even that high for the MP sales (no way Apple sells an equal number of MPs as they do iMacs or laptops (where the 3% value came from; actually figured 25% of 10%, which is 2.5% and rounded up).
Again, this basically comes down to you not providing proof that Apple is either:
a) Losing money on the Mac Pro
b) Not selling enough Mac Pros.
That's not what I said, and I've covered this before.
It seems to me there's an issue with
Reading Comprehension, which I suspect is the result of skimming the content, rather than taking enough time to digest the
actual meaning.
From the 1st part of the previous response:
Now for the moment, these numbers should be sustainable (means the 76K figure as well as the 272K figure).
Now if you weren't sure, sustained = makes a large enough profit that the MP is still viable ATM (if this caused an issue, I apologize, but it seemed reasonable to me you'd have understood the meaning correctly).
But to reiterate one more time...
I have not said that the MP is currently loosing money.
What I have said,
is that in the near future, this scenario is extremely likely, due to economic reasons that have been thoroughly explained.
As mention above, Apple owns the high end market, while Wintel dominates the low end market. Your math is bad, and honestly, Apple only needs to make a profit.
You're making the same assumption as
zephonic; that it's about consumer systems.
The sales figures linked are WORKSTATION ONLY. Nothing else is accounted for in those numbers.
As per the marketshare %, all that is available publicly, is for Apple's entire line of computers. Realistically, the MP is not on equal footing with the rest of Apple's computers in terms of % either, and is where the lower estimated figure comes in (see the response to
zephonic for further details).
The actual math applied to sales volume was only to put a realistic number to it (proves it's not as high as millions of units, or as low as 1 for the entire year). It neither shows profit or loss. ONLY THE UNIT VOLUME SOLD for 2009. That's it.
Nothing else can be inferred from that information alone. Other figures are required to put this to use to demonstrate a profit or a loss, such as total cost for the entire order (what they're contractually obligated to pay to the ODM, and any internal costs actually paid, such as salaries for the software development).
Where things can get a bit more confusing, is when a very low profit exists, say 3%, for a low volume selling product (Quarterly profits are low, say in on the order of a few thousand, yet they've millions tied up in it). In such cases, such a low figure could easily cause a vendor to EOL it anyway for other reasons.
The most likely of such reasons, is they'll determine that they could get a higher ROI on the principle spent (millions spent on the ODM costs, ...) by investing it elsewhere.
I do not think this is the case right now, but could be part of/will be the actual basis for any future decision to EOL the MP.
It's that Apple was shipping a crappy product, and they either had to fix it or let it go, they let it go.
You're seriously thinking that crappy product issues = out-weighed any financial implications? Are you serious?
If a product is developed that's crappy (bug laden, whatever), but has a high profit margin, then that
high profit = sufficient motive to fix the problems. Fixing products in this situation can actually increase sales, and therefore increase profits!
But when the financial aspect /= incentive to improve the product by solving the various issues it has, they milk it until it reaches the "magic" volume figure that = EOL.
It really is this simple, and how business actually works.
Corporations are purely about money, first, foremost, and finally. Any other concepts were buried in a landfill somewhere the moment that company was formed.
This is how businesses justify what they do on a personal level, and why we get statements like "it's just business" or "it's business, nothing personal",... Greed trumps ethics in less than a heartbeat. Why do you think they break the law, get busted, pay a fine, and do it again?
They look at it like tipping the waiter at your local sit down restaurant; only on a much grander scale. They still walk away with more money after the fine was paid vs. had they not broken the law in the first place.
I don't like it, but it's the way the world we live in works for big business.
So can you really sit here and say things other than money are the fundamental reasons for their actions?
Meantime, their sales were falling because:
a) It was a crappy product (software).
b) The Mac Pro was becoming more popular because it cooled better and could fit a full height graphics card if you were doing clustered computing (CUDA, OpenCL).
If the XServe was sufficiently profitable, they would have solved it's issues because it was in their financial best interest to do so. They didn't, so that tells the real story - it wasn't as profitable as they wanted, and decided not to spend another penny than they absolutely had to until it was EOL'ed.
The MP was the opposite; it had the financial incentive for Apple to continue to invest in it's software development.
__________________________________________
BTW, from a hardware POV, there's little differences between the XServe and MP as they're both Enterprise systems, and based off of the same Intel parts and base reference designs.
That is, a Nehalem based DP MP and XServe used the same LGA1366 socket (same Xeons can run in either), chipset (5520), and ICH from Intel. The boards are separate P/N's due to the different format specifications needed to fit the enclosures.
If they screwed up the boards, it was the result of the ODM, such as poor voltage regs, .... Software issues fall on Apple.
Apple is willing to dig products that don't make money out of their holes if they are good products (see: Macbook Air, Mac Mini, AppleTV, etc.) The XServe just wasn't one of those products.
Care to actually prove this?
I ask, as if you look deeper, they weren't really losing money, and Apple fixed/updated them due to a sufficient profit motive.
Lets look at the AppleTV for example (only product you've listed that actually turned up a result during searches). IF you read the
article carefully, you'll see it states there is
~21% Gross Margin. That's a profit, not a loss.
Now they also mention that it
may be losing money due to marketing expenses, which is just speculation on the author's part.
>>>But let's assume
for the moment, the author is correct. Apple's looking further down the road, which indicates that it's still a viable product with their idea of margins, and is directed at increasing the sales figures significantly, = increase their profits significantly for a "loss", they'll actually get back.
This sort of move is a short term loss-leader to increase their total sales volume in the near future for that particular product, with the expectation of positive results in the next quarter or so.
The motives may look like something else on the surface, but they really do have a profit motive behind them.
You only do a re-write if you plan on keeping the product around. A rewrite is much more expensive than an upgrade. For an upgrade, you have to write x number of lines of code, whereas for a re-write you might have to write 30x lines of code to 100x lines of code.
Yes, you're right, a re-write makes your code cleaner for future upgrades. Which is exactly my point. A re-write only makes sense if Apple is planning future upgrades.
Seems you missed the point. I'm not arguing that FCPX will only be around for a year or so. It could go on for another 10+ years if it's profitable enough to keep it around.
I've the impression you've tied FCPX to the MP in such a way in your head, that FCPX cannot exist without the MP, which is
false.
The reality is,
FCPX can already run on other systems, such as the iMac (it even runs on the iPad!; watch the
youtube video in the link).
Which means, that
the MP can go EOL, and
= FCPX has a future without the MP (where the pro-sumer shift argument originated from various graphics pros in MR).
A re-write can also be faster to write, but it doesn't make up for the sheer number of lines of code, and that doesn't factor in:
a) An greatly increased amount of planning
b) Having to do new UI design and artwork from scratch
c) Requiring a far deeper and more thorough Q/A cycle
These factors are accounted for in man-hours.
So less man-hours to develop the re-write vs. those needed to go the spaghetti from hell route = cheaper to do the re-write.
Which I think is exactly what they're going for. It's the same thing they did when they dropped FCP from 50k to 1k and everyone was like "ZOMG how will they afford that!"
This isn't the issue for the pro's, and I agree with their frustration/disappointment over the initial release. It's beta-ware.
As I've said before, this is fine for consumer-ware, but not professional software.
You may not see the need for pros to have the features from day one, but it's absolutely valid from their POV.
If anything, I think they plan to make it up on Mac Pro sales, further putting a stake in the "Mac Pro is dying" coffin.
It's MSRP already includes a profit margin. They just need to make enough sales to make it actually happen (ATM, they're still regaining their development costs).
I even suspect Apple was aware of the issues that would come up as a result of the missing features, and have estimated the development time needed to complete those sections into the current MSRP = no loss at all. And the sales volume will go up, and make a profit. When is the only real question IMO.
But the issue of
FCPX = tied to the MP to exist = FALSE. FCPX can actually run on an iPad for crying out loud... The "must be run on a MP" argument is invalid logic.
Apple's profit margins are high enough on the Mac Pro that they'd seriously have to be doing something wrong to not be making a profit.
You haven't at all proven your economics. Until you do, I'm not accepting this point.
You're making an assumption here.
We know their current avg. Gross Margin (41%), which held true on the 2009 and 10 models in a reverse analysis (it's in here somewhere if you search). What we don't know, is the actual MP sales volume.
So I used what I could find that was relevant and publicly available (2009 Workstation sales), and do some math to figure out how much of that would belong to the MP. That's as close as we're going to get, unless Apple suddenly changes their minds, and splits out the MP data from the rest of their computer systems. Unfortunately, I suspect pigs will grow wings and fly before that ever happens.
But for the moment, I do suspect the MP is fine. I've said this multiple times.
The issue has to do with what's coming up (systems not yet available), as there is plenty of information that the LGA2011 socket parts are going to be expensive. Then when you add in Apple's profit margin, it's not going to be pretty.
_________________________________
LGA2011 vs. LGA1366 = 47% increased pin count, so let's look at this from a simple linear calculation in order to illustrate things ...
Now if you take the X5670 (fastest DP Hex used by Apple), which sells for $1440, and simply multiply by 1.47, you'd get
$2117 per for the equivalent SB-E5 replacement (DP @2.93GHz) in terms of market point. As you need a pair, it
ends up as $4234 in CPU's alone.
This is the first reason the MP's MSRP will have to increase (let's call it the first punch). Pay attention, as it's going to get ugly... fast.
Simply looking at the current 2.93GHz DP @ $6199 and the quantity pricing on the X5670, you can get the following:
- $6199/1.41 = $4396.45 cost to Apple (takes out the 41% margin).
- 2x 5670's = $2880
- System cost - CPU cost = $1516.45 (covers case, PSU, board, drives, ...).
Now let's just use the current system manufacturing cost, sans CPU's for the moment.
Bare system cost = $1516.45
CPU cost = $4234
Total system cost = 5750.45
Now for the second punch...
Multiply by 1.41 to add in a profit per unit, and it comes to
$8108.18.
Now the board will be more complex for LGA2011, and there'll be an additional pair of memory sticks, so that should be ~15% increase in costs on those ends. Which generates a no CPU per unit cost of $1743.92.
So adjusting for this:
Bare system cost = $1743.92
CPU cost = $4234
Total system cost = $5977.92
Now figure in the margin @ 41% = 1.41 * $5977.92 =
$8428.87.
Not that much of a change in terms of % (4%), but it's still an increase that potential buyers would notice ($320.69).
Intel's pricing combined with Apple's margins aren't pretty at all. It's a one-two sucker punch, and it's going to hurt (at this product point, it's really a sucker punch + kicked in the proverbial family jewels).
Obviously this is the high end, and these figure's aren't entirely accurate. But it's in the ballpark. To get closer, we'll have to wait for the Quantity Pricing List from Intel on SB-E5.
Now it won't be this bad on the lower end, and Apple will have to select lower clocked parts to keep the pricing from getting this extreme. For example, the base models could be 2.66GHz for the SP, and 2.26GHz for the DP. Lower clocks is the only way Apple could get machines out at a price people would buy (don't expect a reduction in margin). Users will complain, but its the only way to get the math to work out (MSRP's users will buy at, even if grudgingly).
I see this as price creep, and it's real (price increases in excess of inflation; causality is the complexity, but the costs are increasing, but incomes aren't for most, even professionals). Consider how this is going, and think another generation ahead (Haswell), it will reach the point I've been discussing (price creep pushes the MSRP too high to sustain the MP sales at a sufficient level to be profitable enough for Apple to stick with it).
Now I'm not saying they're going to skip SB-E5, or Ivy Bridge versions. But I suspect these will be the last models, when you consider both the price creep issue and that Haswell is expected to produce a single die 8 core part suitable for an iMac (+ TB for peripheral bandwidth). Not an ideal solution, but it is practical for Apple (fastest platform available for their professional software, and can also handle internal software development).
Makes the most sense due to economic reasons.
