February 15th, 2006

What Mactel means for Apple

Posted by Paul Murphy @ 3:03 am

Categories: Apple, Enterprise Policy, General, Government, Hardware

Tags:

When Apple first announced its switch to "Intel Inside" most people seemed to rationalise the change on the grounds that users fundamentally don’t care what’s inside as long as the product meets their needs, that IBM wasn’t meeting Apple’s supply requirements on either volume or performance, and that Intel would.

In addition, most analysts glossed over the reversion to 32bit CPUs brought on by the switch to make one or more of the following supportive arguments:

1. that the switch would lead to faster, cheaper, systems;
2. that the switch would make Apple more "mainstream" and thus lead to wider adoption;
3. that the switch would lower costs for software developers and thus increase the range of software for the Mac; and/or,
4. that the loss of brand differentiation consequent to making the Mac a more mainstream product on the hardware side could be offset by re-focusing marketing on software and "user experience."

That was then, today Apple is shipping an Intel based iMac and is taking orders on an Intel based laptop. Unfortunately what is most obvious about these two new Macs is that they contradict expectations built up around the operation of Mooore’s law by more or less matching, instead of significantly exceeding, The right to watch Microsoft Word run more than twice as fast on the other guy’s Windows/XP machine isn’t going to sell a lot of MacBooks or iMacs. the products they replace and costing exactly the same as, instead of being cheaper than, their predecessors.

Look at Apple’s bill of materials on these products and the reason is obvious: the low end Core Duo CPU from Intel costs Apple more than twice what it paid IBM for a high end G5: about $265 for the 1.83Ghz Core Duo versus an estimated $104 and $78 respectively for the 2Ghz G5 and the 1.67Ghz MPC7447A (G4).

That cost increase on the Intel CPU shows up in the new products in the absence of improved functionality -particularly with respect to the screen, media drives, ports, and power management- and in the absence of price cuts. Indeed the current iMac and the forthcoming MacBook are the first new Macintosh series ever released not to have a price/performance advantage, when purchased as complete systems, over their Wintel competitors.

Look at the resulting product weakness from an Apple shareholder perspective and this might not seem to matter because Apple’s profits and growth are being driven by the entertainment products, notably the pods, the music store, and the forthcoming video service; not the computer business. Unfortunately that’s a very shallow perspective because sales in those areas are ultimately driven by the halo effect cast by the Apple brand -and that’s driven by widespread perception of the Macintosh as a premium product.

It’s generally true that Macintosh users genuinely don’t care what’s under the hood, they just want the product to meet their needs. Those needs, however, aren’t limited to running their software and extend, instead, into the realm of self affirmation.

Fit perceptions of the Mac to Maslow’s hierarchy of needs and what you see is that Mac user needs only start with the basic requirement that the machine process their work, and run their applications, at a reasonable rate. Beyond that, it is product differentiation as a premium brand based on better design, better performance, and more comprehensive functionality that drives customer self-affirmation and thus contributes to the community membership, fulfillment, and self actualisation elements at the top of the hierarchy.

Fundamentally it’s the Apple brand, not intrinsic value, that sells the stuff that makes the money.

With that in mind it’s possible to look at what "Intel Inside" really means for Apple.

First, the reversion to 32bit CPUs isn’t much of an issue for the laptops, iMac, and the entertainment products. It does, however, knock out the key business lines: the PowerMac and X-Serves, and therefore cripple Apple’s drive to maintain its market share in the high end publishing, photography, and video processing businesses. Fundamentally what’s going on with those lines is that each time Intel announces further delays in getting lower power, 64bit, CPUs or integrated multi-core processors out the door in volume, Apple’s options for this business line narrow and its credibility among key customer groups driving widespread downstream adoption decreases.

Secondly, the promised speed increases simply aren’t there. Many reviews have now been done of the Core Duo based products, and the results are virtually unanimous: on applications built specifically for the x86 architecture the dual core 2.Ghz Intel machine is in the range of 10 to 30% faster than the G5 it replaces while producing significantly less than 50% of the G5’s throughput on key user applications like Photoshop that have yet to be ported back to the x86 world.

Look at an average real world usage mix and a new iMac is considerably slower than an old one. Users will, of course, understand intellectually why that is, but the right to watch Microsoft Word run more than twice as fast on the other guy’s Windows/XP machine isn’t going to sell a lot of MacBooks or iMacs.

Third: the volume isn’t there. The proposed MacViiv apparently had to be postponed because the CPU isn’t yet available (and may be cancelled outright for performance and cost reasons); delays on the Macbook are traceable to Intel supply problems, and the next generation "Merom" and "Conroe" chips Apple needs have just been delayed again.

Fourth: the price decreases aren’t there. On the contrary, Intel CPUs cost considerably more and that gap has to be filled from reductions in product quality, in initial configuration, in long term hardware support, and in Apple’s plant gate margins.

Fifth: the attempt to shift the effort to build brand value from the total hardware/software package to just the software experience has already been derailed.

The problem is simple: because Apple can’t drop the open source OS underlying MacOS X, they can’t stop people from running the combined product for x86 on non Apple x86 hardware. For different reasons they won’t be able to stop people running Windows/XP on Apple’s x86 hardware either. Taken together these realities turn Apple’s hardware into a commodity and force Apple into direct competition with Microsoft exactly where Microsoft is strongest and Apple weakest: on the user eXPerience in its home grown entertainment applications.

Sixth: there is a potential problem with the trustworthiness of the MacOS X on x86 that has yet to appear but could be devastating to Apple. The problem is this: it’s a lot harder to exploit a code vulnerability like a buffer overflow or heap offset on PPC than on x86. Thus vulnerability announcements for Apple have rarely been accompanied by exploits and Mac users have generally considered the whole PC security mess a Wintel problem - and thus a key dimension on which their choice is smarter.

Unfortunately that may change with the move to x86: there are lots of people in the security arena who are highly motivated to attack Apple, and the x86 makes the second half of the process: exploiting a vulnerability once found, much easier than it used to be.

Bottom line? "Intel Inside" cheapens the brand, weakens the halo effect supporting Apple’s highly profitable entertainment products, raises Apple’s costs, results in reduced overall performance, and limits Apple’s ability to differentiate its products.

So what can be done? Stay tuned, tomorrow’s blog is about turning this disaster into opportunity.