25 January 2006

Apple Thoughts and Myths/Realities Follow-up

I've been working on a number of pieces since the MacWorld article, and the Apple financials, but they've usually been usurped by something else I've read which is much better (and usually more concise) than I can possibly write. So, instead, I'm going to provide a few links to the more interesting comments, and also summarise my current thinking as regards Apple's challenges ahead. This is still going to be a very long article, but hopefully will cover a lot of ground.

I felt leading up to MacWorld that expectations had got far ahead of what could reasonably be delivered. I analysed views relating to Timing, Pricing, Performance and Functionality, and felt that, on balance, there could only be disappointment. And I think in general among the Mac Faithful, there has been an unusual amount of negative postings (example: see O'Grady Powerpage posting today). Many of these are so at odds with each other, that it perhaps highlights that Apple probably got it about right!

Let's talk about the new machines first. As I mentioned at the time, most forecasts were for the lower end machines to get upgraded to Intel first. While I hedged a bit here (expecting Apple to only use a dual core chip in some up-market new product), I succumbed to their overall views. In the end, Apple surprised (most of) us by releasing the mid-range higher margin machines first - iMac and MacBook Pro. In my myth/reality series, I had highlighted the problems of introducing the low-end machines first. So, business-wise, Apple took the right decision, though more on the risks, later. I was hoping to have some informed insight on the new machines, and some comments on what I felt were some silly objections. But instead, I'm going to refer you to the excellent Mr John Gruber's Daring Fireball blog (currently the first article) where he has provided a really good commentary on the new machines. Please also read his comments on the ridiculous iSuppli analysis on cost of an iMac that has been given over-coverage by some of the press.

There have been all sorts of comments on the price of the new machines, their performance, etc and I want to cover some of these a bit below - as they relate to the myths/realities I posted before.

On balance I think Apple provided a very solid engineering solution way ahead of the time they had originally outlined, and indeed brought forward the migration across the whole range. They used current designs for the most part which some feel is very unambitious. But in doing so, that also provides a comfort factor to many of the faithful. And I think many people truly underestimate how much work was involved in the migration and what compromises had to be made. Changing the form factors would have been extremely risky, and probably unnecessary. The iMac and MacBook Pro (ex-Powerbook) form factors are generally best-of-breed still and are relatively well-tested in the field. The changes to the MacBook Pro are generally forward looking rather than mean - at the end of the day there's more in this machine than there is today, without accounting for the performance leap.

So, let's take a look at the myths/realities in more detail.

1. Timing. I had bet we would have Intel Macs and we do. And, as I mentioned above, financially-speaking, these are the right machines to do first. Apple has also announced the whole transition will be completed by the end of this calendar year - about 12 months earlier than originally planned. This is great - it reduced the period of uncertainty for many and for Apple as a company. But it will also place great strain on the resources of the company this year, and the effects of the uncertainty may be magnified, even as the timing is reduced. As I will cover in the financials review later, this is going to be a rocky year.

2. Pricing.
I have seen so much poorly written stuff on pricing that it is hard to know where to begin. I had written that an advantage of the move in the end would be transparency with the competition, and I'd hoped this would paint Apple in a more rosy picture. But for everyone who can do a price comparison, there's several who can't. There is little to compare to an iMac, save a desktop all-in-one Sony Vaio. But for the MacBook Pro, there are already several PC competitors out there, including models from Dell, Acer, Lenovo, HP and Gateway. I have looked in detail at many of these offerings, and I have to say that Apple has priced the new MacBook Pro positively and reasonably. As I originally mentioned for every machine an Apple compares favourably to in the PC World, there will be many others seemingly priced better. You want a 15" windows laptop - there are hundreds to choose from. You want a 15" Apple laptop, and there is just one. But if you look at the other models and what they offer, the Mac does NOT carry a price premium anymore than a good Dell (and less than a Dell XPS?) or a good Thinkpad. Of course, if you didn't want that level of finish/features, then it might be out of your price bracket. But please don't call it expensive, when it's not for what it includes and relative to similarly-equipped other models. A key observation I made was that the MacBook appears to be 1lb lighter - or about 20% - than any other 15" widescreen model I saw. (and a word to UK observers, please, please please price in VAT when you do comparisons. How many times do I have to tell you!).

3. Performance.
This is an area where I've been equally disappointed with what I've read. While I had suggested that performance wouldn't be earth-shattering, I had written this on the basis that it would be low end machines first. Instead, Apple has taken the mid-range machines and turned them into potentially PowerMac killers. I say "potentially" because there are several caveats I will come to. So, let's see what some people have said:

Let's go to The Register first to see how it covered a report produced by MacWorld.

This is a really poor report and they have gone down significantly in my estimation on this article alone. For instance, they quote 4 times as the benchmark expectation. That is ridiculous. Apple claimed that the iMac was 2-3 times faster on certain benchmarks demonstrating the theoretical performance of the insides. At the same time it showed other stats demonstrating that real world (native) applications could be as much as 1.9 - 2.6 times faster. SJ also pointedly noted at the keynote that real world speeds would be affected by many other circumstances such as RAM, hard disk speed, application type etc. Steve also deliberately showed Photoshop working under Rosetta and demonstrated that it was usable but on balance slower than on a G5. So, for the iMac, Apple's position has reasonably been to expect some aspects of performance to be about 2x better, and overtime for that performance improvement to apply to other applications. So, let's look at the other part of the Register story - that it is in fact just 1/4 faster or 25%. (Incidentally, a figure that is not in the least embarrassing given that the iMac already used a fast G5 chip, and is priced at the same price as it's predecessor). To look at the 1/4 claim, let's go to the MacWorld story itself.

Jason Snell, MacWorld editor has in fact done quite a commendable job and he has suffered a lot of abuse over the article. Some of it's his own problem by failing to comprehend the concept of percentages. He has fortunately corrected that in the article you will read, but not all coverage reflects that. Currently headline figures show performance improvements of zero to 1.84 times (with one slightly negative figure). So, Andrew Orlowski's headline figure is just that - a generalisation he has pulled out of the air to prove a cyncial point. While Jason's article is more detailed than many and he has tried to offset some issues, I think the problems are clear if you can be bothered to wade through the comments left by some observers. Let me summarise some though:

a. Dual core tests. Almost none of MacWorld's tests will have tested the dual core performance of the machine, on which a lot of Apple's claims rest. Perhaps the one test to do this is the boot up test, as we assume Apple has managed to optimise some of Mac OS 10.4.4 to use a dual core when on offer. In this regards a 1.84x improvement is pretty decent - especially considering disk and RAM will play a part in this too. While some may claim dual performance isn't reasonable, I would say it very much is. Increasingly we are really using our machines to do many things - some in the background. Converting a video to MPEG 4 for iPod while browsing and having an iChat are good examples of things we are all doing more and more. I'm surprised that MacWorld didn't design the tests to show how things work better (assuming they do) when we're doing more than one thing - which is when we probably most want the performance improvement.

b. RAM. While it may of course be fair to use the stock RAM (512MB) in both test machines, the level of RAM chosen seems strange given the design of many of the tests. Many media tests and certainly Photoshop tests will be RAM constrained on a 512MB system. I would imagine Rosetta would only serve to squeeze that more (though it would be unreasonable to compare a 1GB Intel iMac with a 512MB G5 iMac). But given that SJ said other factors came into it, I would have thought MacWorld could have designed the tests a little better to isolate the effect of RAM on such tests.

c. Scenarios used. The MacWorld tests were limited to a few scenarios which are overly focused on just a few applications, and do not focus on what people do much of the time. For instance, why not cover application loading time? Why not cover page rendering time for websites? Another way to look at performance would have been to see the free CPU available when running tests. I have seen reports of HD replaying that are very positive indeed - putting a MacBook Pro (pre-production) easily on a par with a dual core G5 PowerMac, for instance (40% overall CPU usage vs 55% for full frame performance). While graphics card changes might help with this test, these are important numbers - especially going forward.
On another review of the iMac, David Pogue
at the NYTimes
quotes a massive speed improvement - certainly of the order of 2x for page rendering and application start-up. While I'm not sure of the scientific nature of these tests, these strike me as more meaningful than the MacWorld tests.

d. Last but not least, on performance, I am really surprised that we have not seen some performance comparisons between Windows machines and the iMac. There are some Yonah-based laptops around that could be compared with an iMac. Or use a Pentium D or something similar in price. iTunes would be a good application to compare against. Boot up times, wake from sleep, times to duplicate files etc, would all be interesting benchmarks, as would HD playback or other Quicktime tests. In the end, these are the tests that will determine whether the Mac succeeds or not.

And, in the end, this story is only going to improve from here. We'll have more native applications, and more optimisation for Intel via improved compilers and code improvements as people become familiar. And later this year, we'll be moving towards the first 64-bit dual core chips. Finally, remember Steve's reason for moving to Intel was performance per watt, not absolute performance. By all accounts the new iMacs are very quiet which implies lower heat even with 2 cores versus single core G5.

As an aside here as well, it's worth commenting on Rosetta. The reports in so far seem to indicate that this has been indispensable technology to Apple in achieving this switch. Performance numbers combined with excellent application compatibility are really good stories and have been frequently overlooked.

So, I can summarise this by stating that I believe the performance numbers are looking pretty good already, and this performance combined with lower power consumption will provide Apple a springboard to introduce the great new leaps forward in functionality that some expected would come immediately with the shift to Intel. Don't believe all that you read unless you really dig deep (read the comments if they're available). And in the end, don't compare an old Mac to a new Mac - compare it to a Windows box as that's what is going to count long term. I don't know how they would stack up, but I can't believe that Steve made this move without believing that Mac OS X shows up pretty well against Windows.

4. Functionality. - the final one of the four myths/realities. I think many were disappointed that MacWorld did not herald a great leap in functionality - into the living room for instance, or perhaps a well-rounded iWork suite including a decent spreadsheet. We got very little other than an improved iWork and iLife. I have covered iLife a little here already and maintain a generally positive view on it (though feel iWeb is still very much a v0.9 application). I am using iWork but not in anger yet, so can't comment too much. With the hole left by Appleworks not being made available on the Intel Macs, I think Apple has a problem in providing a complete Home Office package competing with all-in-one packages available on the Windows platform and often included in the low end machines. Hopefully it has something up its sleeve to fill this void. But more importantly, what happened to the convergence functionality being expected? Some people expected an Apple DVR. Some (me included) expected a video-capable Airport Express. Others expected FrontRow to be made available across the range of Macs (though I think this is just a slightly cynical marketing move).

There have been reports that the MacWorld keynote was changed at the last minute with some key announcements dropped. It's difficult to know how true this is - especially as the keynote was already significantly longer than previous ones. In the end, I think Apple's MacWorld was indeed conservative but partly because it had to be. SJ made a deliberate statement about it being MACWorld, not iPodWorld. He focused his major announcements on the Mac platform, and hopefully re-assured the faithful that Apple's strategy is about Macs and Mac OS more than it is about the iPod which could be a transient (though, of course, huge) phenomenon. The conservative introductions (form-factor-wise of the new Macs) showed it was taking the migration seriously - not flippantly. It is also possible that some things were just not ready for primetime. The video airport express for instance might be dependent upon 802.11n - the next generation wireless specification. It had been expected this would be finalised last year, but in fact, it is only in the last week that this has happened. While Apple has always been at the forefront of wireless adoption, I think it could not afford to announce products at that time - and probably didn't need to. With the finalisation of this spec, I think we can expect an Apple announcement about using it within the next 3 months. I'm sure 802.11n is a vital part of Apple's whole-house strategy. But, if so, it also belongs with announcements about affordable home Macs - including iBook and Mac mini updates. The pieces are appearing - remote controls for new Macs, FrontRow, iSight built-in, Photocasting directly, more display connections (in the iMac)... The big missing piece though remains Apple's ability to incorporate broadcast/cable TV into the platform. I did not expect an Apple DVR this time round, and I don't underestimate the difficulties of doing this across the many technologies of terrestrial, cable and satellite in multiple world markets. But eventually, Apple must bite the bullet on this and make this an integral part of the Mac-in-the-home experience.


Financials

Finally, I'd like to wrap this analysis up with some commentary on the financials. Apple's numbers for the last quarter were staggering. Over 14million iPods sold is an incredible achievement, especially considering previous problems with inventory on new models (eg iPod mini). (As an aside, I believe Sony Ericsson shifted about 14million phones in the quarter, so that demonstrates the scale of this achievement) And the Mac numbers too were on balance very good. Year-on-year Mac shipments up 19%, and Mac revenue up 7.5%, with record Mac sales. Over 2 years, Mac shipments are now up over 51%. Most PC companies would kill for such numbers. But, in reality the growth was much reduced from the previous quarter and started to show that people were indeed holding off. Did Apple see this coming and cause it to concentrate on the higher end machines first? I'm not sure, as looking back, it is obvious to see that it is the dual core variants of Yonah that are in supply first and that these were always going to find their way into high end machines.

But, one does wonder whether Apple is responding to this dip rather than leading it. Furthermore, the next quarter's predictions - while a Q2 record if they're hit - are not awe-inspiring. Personally, I think they may struggle to hit even these numbers. iPod shipments in the final quarter benefitted hugely from the holiday season and the release almost simultaneously of major improvements in the product. As the market matures the buyers become less "influencers" and more "influenced". The latter category are more likely to concentrate their purchases around events such as Christmas or birthdays. I know many people who bought one or more iPods last quarter. I know of no one who has bought one or who is contemplating one this quarter.

So, I think we'll see a return to more normal levels (if 6 million a quarter can be described as normal). Mac sales I'm particularly worried about. Apple cannot begin shipping MacBook Pros until at least mid-quarter. I know many people who will hold off from the first release (though I've seen anecdotal evidence of many switchers stating their intent to buy). As for the iMac, it's not clear whether they'll either have enough, or whether the first release will attract more people than it loses due to initial compatibility and overall FUD factor. iBooks, and minis will barely shift, and even Powermac sales may be disrupted now people know what the Intel Macs can do, and know that new Intel-based Powermacs will be available starting this year - not the end of next year.

Of course, Apple will have achieved its transition to Intel completely by the time Microsoft is shipping Vista, and the importance of that as people make buying decisions cannot be underestimated. I also think we'll see some block-buster products this year - with new form factors and functionality following on behind successful product transitions. Apple had a window (sorry) in which to do all of this, and so far, they're doing really well on that. But the halo effect can be easily dented and much harder to re-light. And I find it slightly ironic that just as Apple is executing better than ever (eg iPod fulfillment, first Intel macs ahead of expectations) there are rumblings among the Mac faithful (perhaps encouraged by others who'd like to see Apple fail) whose expectations seemed to reach beyond what even Steve's infamous Reality Distortion Field (RDF) could conjure up this time round.

Here's the reality folks, and you should be enjoying and revelling in it: Apple is delivering well on products; it is transitioning faster and with the products needed; the performance numbers ARE good and even more so looking ahead; the compatibility backwards is also good; and it is maintaining and even growing its leadership in digital music. All-in-all what a strange time for Steve's RDF to evaporate!

6 comments:

Paul Guinnessy said...

The more I think of it, the more I'm convinced that the ibooks will be single core chips (because of pricing issues for Apple, and if the ibooks are too competitive to the Macbook Pro's, it might limit their orders). They may offer a dual core option that will be about $1499, but I bet the idea will be to go with the single core model. It would explain why no ibook was launched at the expo.

While a signel core would still lead to a 2x improvement of speed over the current ibook models, I'm really looking for a bigger speed boost myself. Hence I've just placed my order for a top of the range Macbook Pro, something I've held off from doing for a few weeks. According to amazon, my laptop won't ship until March 29, so on March 30, when the new single core chips are available, I suspect Apple will release a decent replacement for the 12 inch powerbook.

Matt Stowell said...

Hey there. I just bought a video iPod about a week ago. I had some left-over money that was a Christmas gift, and figured now would be a good time to finally buy an iPod. It handles music, pictures, video, and also stores my address book and calendars/to-do information. What else can they really add to it? Of course, any day now Apple will probably release the new iPod that's the size of a grain of rice and comes pre-loaded with every song ever recorded in human history....but oh well.

mark said...

Ian, this is the first time I've read your blog, sent here via MacSurfer. I've seen your good comments, though, on various other sites. Just wanted to say you've written an excellent critique. And toss in a few thoughts for discussion:
- I think Apple held off on its living room solution because it believes a reliable wireless connection to send high-quality video to the TV was absolutely essential. UWB was the leading contender but IEEE-approved 802.11n has taken over. Broadcom and Marvell said last week they're sampling chipsets with expectation that products from partners will ship this quarter or next. Thus, I think Apple will hold a special anniversary event (Tues before or after Apr 1) to introduce its new Airport HD Express, alongside an upgraded iMac with 802.11n. Movies at the iTMS will debut as well. And maybe a new larger-screen $499 iPod video (could be the rumored iPod boombox with a ... 7" screen!). This event will mitigate any shortfalls in Apple's Q2 results announced two weeks later.
- Note that Gateway will be selling 1.66GHz dual-core laptops for 1099 and 1299. Granted that they are stripped down... Regardless, I think the iBook and Mac mini await single core chips.
- Q3 will be an incredible Apple quarter. The living room stuff, iBooks, and Mac minis (all with 802.11n) should begin shipping, while the 12"/17" Powerbooks might be unveiled the week before or during WWDC.
- If iPod sales flag during this quarter, Apple could give it a small bump: move 1Gb shuffle to $99, 2GB nano to $169, 4GB nano to $199, and a new 6GB nano for $249. Flash prices seem to be able to support such an adjustment now.

Anonymous said...

Real-world tests running multiple-applications at the same time shows that the iMac is almost as fast as a dual-core G5 machine!

http://macspeedzone.com/html/hardware/machine/performance_in_the_raw/06/1_23.shtml

MacWorld had pretty dumb tests running only single applications.

Ian Hobson said...

Paul
I got the HighRes 15" Powerbook G4 in November thinking it'd be a while before Intel PB's would come out! I'm pretty happy with it really - my 6th Powerbook since the 180(?). Though there is a nagging audio problem that's now getting some web coverage which I'm sure will be resolved (and only affects me in Skype and iChat). The pro machines really are very good and last well. I sold my Pismo (admittedly upgraded to a G4 with memory, disk etc) for £370 in November! You will not regret this machine I'm sure.

Matt:
Have you seen "Steve Jobs" on Saturday Night live late last year announcing the iPod invisa? Here's a link (not sure it works - try googling) but hilarious
http://vpwpartners.blogs.com/viewpoint_west_partners/2005/11/steve_jobs_on_s.html
the 5G iPod is great, and again you won't regret it.

Mark:
Thanks for the compliments, and it's nice to be picked up for a wider audience. I'd like to get some informed, intelligent Mac comment here and you've certainly done that.
I agree with your thoughts too. I wonder whether we'll see a few additions in the laptop range like a MacBook Light using the Low-voltage core duos (when they're available) for a smaller package/longer battery life? What do you think?

The Gateways you mention are exactly the sort of machines that will always be used to reinforce the message that Apple is expensive. But when you deconstruct them, you get what you pay for as you say. Given the pricing of the Acers, Dells and Lenovos, I don't see the MacBook Pro expensive for a high-specced machine.

Anonymous:
Thanks for that link - a much better test indeed, and showing how the new iMacs can indeed be superb performers. Don't be shy with your identity next time!

mark said...

Informed, intelligent discussion is what I'm looking for. And altho in reading your past articles, I don't always agree with your economic/business points, it looks like that's what I'll find here.

Regarding new form factors, I think Apple would love to release a tablet - not sure if it would be a Mac or an iPod. Big question: Will the architecture for Macs and iPods converge via Intel chips? Both pros and cons so not sure. In any case, Macs do it all (info/content creation, delivery, consumption), so it would need touch-screen/handwriting or voice input. iPods do consumption today, so a tablet (basically much larger screen) with clickwheel is easy-to-do. iPods could might even do delivery this year with the low-powered version of 802.11n.

In the longer term, as NAND flash capacity growth continues, wireless access becomes less-power-hungry, and access becomes ubiquitous (maybe via WiMax/3g cell), we should see Mac (and iPod) products with 10-20 GB NAND and continual access to remote storage (Mobile Me maybe?). Can't imagine what they'll look like but they'll be real flat and have 10 hr battery life. Maybe in 2007.

Separately, the MacSpeedZone tests really point out the flaws of current benchmark schemes for comparing single, dual-core, and dual dual-core (quad) processor systems. We need consumer and pro scenarios that use multiple pieces of software at once (from the user point of view). Like launch four apps rapidly in succession. Have one foreground operation (viewing video) and three background operations (encoding audio, searching drive, updating mail/feeds) simultaneously. Then we'll see the true speed gains we are getting.