The squandered G5

If you just looked at the positive press going out on the G5, you could be forgiven for believing the G5 to be one blockbuster-selling piece of kit for Apple. However, this isn't really the case.

While it's gotten great press, when you actually look at the figures, it's pretty clear that it's been more than a little lackluster. Which isn't to dismiss the positive press, it's been pretty spectacular for a company with Apples' history. The idea of an "Apple Supercomputer" has been something of an oxymoron for a long time, especially given some of their earlier G4 advertisements. At the moment, it's not, which is super cool.

There have been some negatives, like basically being told by governments that their advertising was lying not quite truthful, or quiet articles basically saying that the G5 is not hitting expectations... but on the whole things like the Virginia Tech supercomputer have been fantastic for slowly shifting the perception of some of Apple's kit, and where it can be used.

If you'd have made a bet with most people as to whether or not Apple's towers would be powering one of the most powerful supercomputers on the planet in a year, it's a bet most people would have lost (including myself, and, if some of the interviews from the VT cluster's architect are to be believed, Apple also...). Some kudo's are in order there: Apple used to have one hell of a share in the scientific market, which slowly eroded over time, but it's now making some serious inroads back into. Credit where credit is due, you can at least use the words Apple & performance in the same paragraph without feeling queasy. Credit where credit is due.

But that's being squandered. Now, some of you are going to go "Microsoft shill! The G5 is selling tons more units than the G4! How much is microsoft paying you!", and that's fine. But let's take a look at the unit numbers by quarters for their towers, in thousands of units:

  • 2002Q1 213
  • 2002Q2 211
  • 2002Q3 167
  • 2002Q4 176
  • 2003Q1 157
  • 2003Q2 157
  • 2003Q3 133
  • 2003Q4 221 (G5 intro)
  • 2004Q1 206

Woo hoo! When the G5 was intro'd, there was an instant 60% jump. How can that not be good? Well, it is good. However, you have to realize just how poorly the current G4 towers were selling. It isn't hyperbole to say they were selling incredibly poorly. The G5 has basically brought them up to 2002 G4 numbers, which aren't considered to be good numbers at all, even when compared to the heyday of the 2000-esque advertising boom.

It can't be stressed how poor a level sales of the G4 towers reached. We're not talking cube-esque numbers here, but it's obvious the only people buying them were people who had to have a mac tower, and couldn't wait. Most just weren't interested, no matter how much mac users talked about the "Megahertz Myth", it's fairly obvious that the price performance ratio was way out of whack. But, but the G5 is the fastest personal computer on the planet, right? So that should be taken care of... which is part of the problem: The G5 has not only failed analyst's expectations, it's fallen far short of Apple's own estimations which it gave to the analysts.

Apple's big line to the analysts was that people weren't buying G4's because of a perceived performance problem, the lack of native software for some of their core markets, and the huge advertising slump. They basically said the 1st two were fixed, and the 3rd was turning around bigtime, so bring on the upgrade churn as they believed there is a large chunk of their base who were waiting with baited breath to move on, and that the G5 was going to sell in fantastic numbers due to this backlog.

It just hasn't happened en-mass as they'd hoped, at least not yet. Some of it is dependent on 3rd parties, some of it is unavoidable, but some assuredly rests on Apple's shoulders, and some of it, while it may not be 'their fault', is still very much their problem when it comes to 3rd parties.

Now we don't know how they are doing this quarter, but we can infer that they aren't doing very well. They've been running promotion after promotion with them lately, and the G5 hasn't left Apple's front page with lots of life-altering quotes about it from various bigwigs... sure, since it's been 9 months since a real refresh, chances are they trying to clear out inventory before a refresh soon. But that usually involves some price cuts and perhaps a quick deal or two, but they are trying really damn hard to move units. Really hard.

So, if we can say that the we've established that the G5 sales are "underperforming", we should probably look at why that might be? Besides some of the obvious, the G5 is a very... perplexing machine in some ways:

  1. Odd initial launch lineup
    They launched with single 1.6GHz, 1.8GHz & dually 2GHz. That was really weird, and arguably way out of whack. The top end was outselling the low end configs combined, by a huge margin, which is a pretty big clue stick that something is really wrong in the lineup. This has happened before. There was a time when Apple was releasing their high-end as a single proc system, and the rest were duals... which didn't exactly go over so well either. In the days of OS9 there was some justification for it, but not really with OSX: most people now know that duallies have massively better ROI, especially with with an OS that does its best to take full advantage of them. In this case, for several hundred dollars more, you just got a huge amount of extra power. The 1.6's & 1.8's weren't selling well, especially the 1.8's, they were just in this no-man's land of pricing. Apple stepped in and made the 1.8's duallies during a refresh, but now you have all these single 1.8's still in the channel (check any catalog or mac site).
  2. Forget what we told you you wanted before...
    It's a slightly larger case than the G4's that came before it, yet has drastically reduced expansion options when it comes to drives. There is no way a single serial ATA drive is going to saturate that bus. I don't think two even comes close. These cry out for RAID5, which would need a minimum of 3 drives. All you can basically get is software RAID 1 & 0, which for a machine in the G5's class is just kind of odd. To get RAID5 or other hardware raids you're essentially looking at firewire, which really adds to your costs and creates its own issues or you're looking at a very pricey external ultrascsi solution. Egh. Having only two drives is just very out of place in a machine of this caliber, and is perplexing. When you realize Apple tried to sell recent revisions on increasing drive bays, as well as dual optical, it gets more perplexing. Spending $3k on a large tower with two drive bays is just odd.
  3. Saggy GPUs
    Consumer-level, and dare i say close-to-cheapo-pc-level video cards in a workstation-class machine. $50 upgrade to something reasonable, $350 upgrade to a competitive card. 64megs of VRAM in a $3k computer that may be driving 2 displays unless you pay $350? Wtf? That's perplexing, especially for who these systems are geared to. Yeah it's fine if Apple has given up on the game market, but there are other reasons to want a more future-proof GPU and lots of VRAM... especially considering Quartz.
  4. Product overlap
    It's kinda weird how Apple went through so much trouble wiping out the deadzones from its product line, and has lately been blurring the lines all over the place. Part of this was out of desperation ("we can't compete on g4 towers, but we can build a decent laptop with their lower thermals- year of the notebook!"), but there's really a lot of overlap. Apple is talking up powermac users using powerbooks instead, but I don't really know about that. While I've seen it, and I'm sure there's some of that, but for their core content creation market the capabilities between a 1.25GHz G4 powerbook with a single slower drive and a dually G5 is pretty drastic. But I have seen a lot of people picking up a 17" or 20" iMac instead of a low end G5 w/apple display.
  5. MHz stagnation relapse
    When they did refresh the line (and surprisingly quickly) it was to go to 1.6 & dually 1.8 & 2GHz, which again, was kind of weird. To keep the MHz momentum up you would have thought there would have been some progression, if nothing else to help differentiate the low end from the iMacs. As it is, the G5's have been riding on 2GHz for over 9 months, meaning you can only coast for so long.
  6. 1st gen issues haven't been fixed
    The G5 hardware is still relatively immature, and hasn't yet had a good revision b or c. Some of the pro's out there were already pulling out their hair with the later G4 models (windtunnels, etc) but were a fairly well known quantity. The G5's have lots of new stuff, and there have been problems with heat, the fans, power supplies, logic boards, line noise... it's not to say that they're a POS, but they do have their first-gen-apple-problems. It's architecture is also very new, with 10.3.3 even bringing a big speed boost, and I'd have to be other apps will see revisions that will bring big gains... there are a number of people who don't want to buy 1st gen Apple hardware anymore, and Apple hasn't given them a revision to buy into.
  7. Regaining trust is hard
    This was something I had raised to me on a list, but makes a lot of sense... There was a ton of damage done to Apple's pro lines by their price/performance problems, even going back to when Microsoft was pushing hard into the content creation markets when NT/2000 hit the scene and Apple's next-gen OS strategy was starting to look DOA. Apple was able to rebuff that fairly well, both with lots of hype around rhapsody, showing they were going to have something to compete with NT/2000, and the fact that the media creation market, especially where print & advertising is concerned, is notoriously slow to move to anything. But the MHz problems dragged on for so long that a ton of people have jumped ship, and now they just aren't buying PowerMacs. They can make headway in bringing these people back, but it's going to take a sustained push as the leader, not one press release. Show these people where the momentum is, especially with AMD's new kit hitting.

All the above are reasons why the G5 is a bit of a perplexing machine, at least to myself, for various markets... but at this point there is another very serious problem they have to contend with: the 3GHz gap.

Lots of people are trying to spin things a bit, and say "Jobs said by the end of the summer, which is technically x... they have time" or similar things. Bullox. Jobs may have said the end of the summer, later, but that's not what he said at the introduction in and what was fed to the press.

The screen capture to the right should make it fairly obvious: Jobs said 3GHz within 12 months, on June 30th I believe. Which leaves a hair under 3 months for them to go from 2GHz to 3GHz. That is one pretty damn drastic jump in power in 3 months time, especially when you consider that the G5 has been out for 9+ months.

So, we were promised 3GHz by June 30th 2004, and there's no way of getting around that. Apple is either going to keep that promise, and lurch from 2GHz to 3GHz in 3 months (chances are it will be 'announced' but not ship for several months...) or they are going to break their promise. Again. Which would be... uncool.

Lots of people are saying things along the lines of "Saying they would hit 3GHz in a specific time frame was so stupid, now people aren't buying. They never should have said that". I'm going to have to respectfully disagree with that mindset. Apple making that statement was the coolest thing they'd done in awhile, and make no mistake they said it for a reason: it was a mindshare ploy. And a needed one, and if they had kept up the momentum it would have been a great one.

Apple's performance problems had become deep-rooted, and really a part of the computer world's psyche. Oh sure, they'd have a jump here and there... and then another long plateau. 2GHz, while a big jump from where they were at, wasn't that huge in the grand scheme of things... they were practically in danger of their MHz being lapped before that. They wanted to have people talking about the G5 as a huge break with what came before, that this was the big break with their past problems of under delivering. They didn't want people thinking about the G5 as being at 2GHz, but on it's way to 3GHz. That it had momentum.

Which is a big deal, and, one could argue, a drastic change from their long-standing modus operandi which while great for hype, is really off-putting to a lot of the people they are trying to make inroads with. Actually keeping their word, and making their ship dates is another big thing with these people. It hasn't played out that way, and now we're stuck hearing pondering on what might be causing the holdup instead of how the G5 is coming on strong.

This isn't just a tower thing: take the server market. People were/are extremely leery of buying into something that won't see good support, and would be dropped at the first sign of trouble. Apple had some very, very nasty mindshare with people when it came to servers due to their past history. People have this image of them going on product flights of fancy, then dropping them dead when it doesn't seem to be working out. This is... disconcerting, and something that will take years to overcome.

Which goes back to the deep-rooted performance/price ratio Apple has been saddled with for ages. The G5 could have been a break from that image in more ways than one, as it had gotten that bad: we're more likely to believe someone has recovered when they have hit rock bottom. Instead, we're seeing signs of a relapse with the G5.

Now of course we're going to see a 3GHz G5. But here's the thing: people are already expecting them to announce it at 2004's WWDC, but actually ship it several months later. That makes me a little sad, as it feels as though an opportunity to really start to change people's perception of Apple has been squandered.

yummy alcohol posted button Posted by drunkenbatman
    April 09, 2004, at 08:53 PM


Comments (99)




Post a comment



Anonymous comments are allowed, but please enter something for a name.

And do endeavor to appear sane.









Remember personal info?