Filed under: Hardware, iBook, PowerBook
Apple notebook reliability study, courtesy of Macintouch
MacInTouch has
released a comprehensive, in-depth report on the
reliability of Apple's portables throughout the years. They apparently logged over 10,000 notebooks from 41 models,
covering the gamut from iBook G3's to the latest Aluminum PowerBook family (basically anything that can handle OS X), as
well as thousands of reader comments. Their survey included what years a notebook had to be repaired, how many times or
if something broke down and the owner simply decided to deal with the situation, such as an optical drive they no
longer need.Their results are very well summarized and broken down into groups such as most/fewest repairs and most common failures from specific models. MacInTouch also tackled the ever-prevelant topic of Applecare, asking whether those surveyed purchased it, if they felt it was worth the money or if they went with an outside insurance plan (like myself) that covers hardware defects as well as accidents.
I'd like to spoil the results, but I honestly feel they did one heckuva job here and would hate to ruin all their hard work. Swing on by and check out everything they put together, and scroll to the bottom for the cliff notes.
[via MacSlash]
![TUAW [Cafepress]](http://www.blogsmithmedia.com/www.tuaw.com/media/tuaw-cafepress-promo.png)
Reader Comments (Page 1 of 1)
Neil said 9:14PM on 1-03-2006
It seems like they've printed off the results in a spreadsheet then scanned it in - what the hell was the point?!
Reply
Derek said 10:31PM on 1-03-2006
I totally agree that everything past the introduction of the aluminum powerbooks are by far the best out there. I wouldn't get a portable until I first met the aluminum model. There are no PC equivalents based on the core design alone. I can't envision Apple making a better case than the aluminum powerbooks.
Reply
Andrew Montgomery said 11:36PM on 1-03-2006
I'm not trying to flame you. That ire is reserved for Ric Ford, et al. So please, mister messenger, don't feel like I've shot you.
This is an undergraduate course in bastardized uses for Excel, not a comprehensive analysis. Just because someone poured through "over 10,000 laptop" entries doesn't make it comprehensive or in-depth. It just makes it tedious.
There is a lot of horrible mathematics involved in this analysis. But many of their conclusions are comically uninformed.
This article claims that one of the possible reasons for the reduced repair rates in every subsequent release of the PowerBook G4 15" Aluminum series is that "Apple engineers were able to learn from early problems and fix them." Really? Maybe it's because two of the four models are less than a year old. Two of the four models don't have any data older than 11 months, while the other two go back at least twice that far. This is clear evidence that younger laptops require fewer repairs. Thanks, Macintouch, for clearing that up. The only laptop series he mentions here that isn't currently in production is the PowerBook G3 Series. Look at the repair rates for that; except for the initial bump in reliability (50% down to 42%, but who knows what the error is )the repair rates didn't change at all for the entire life of the series. Since no error is quoted, we don't even know if that initial drop in repair rates is even statistically significant.
My favorite conclusion, however, is the one where he says practically in one breath that the "drop percentage correlated very loosely with hard drive and optical drive repairs", then reminds us "that correlation is not causation", but then immediately precedes to speculate on why Apple engineers may have introduced 'sudden motion sensor' hard drive protection included in recent PowerBooks. Jee, I wonder why Apple engineers introduced 'sudden motion sensor' hard drive protection? If your analysis of dropped hard drives vs. damaged hard drives shows anything but a strong correlation, your data are suspect.
There's a reason experimental scientists have their PhD: it's really really hard. The problem is, there's no such thing as close. You either do it right and get meaningful data, or you do it half-asked and end up with meaningful sounding numbers that are devoid of anything substantial. Like this.
Reply
Mojo said 12:47AM on 1-04-2006
I would also view the data's reliability carefully given that respondants were more likely to have had problems and that respondents with problems who take the time to complain would also be more liekley to lie or omit pertinent info such as abuse. I would probably conservatively halve about all the percentages and feel this was closer to the truth.
Reply
brandon said 4:57PM on 1-04-2006
I've had a few Apple laptops running OS X since about 2001. I've been pretty happy with the machines even though I have had some problems with them. I had the issue with the dead iBook motherboards twice and the hard drive of my Powerbook died. Overall I really enjoy traveling with them and will look forward to the Intel models.
Reply
Moquiti said 9:45PM on 1-04-2006
The survey gives a general qualitative sense, and that may be all that it's designers intended. Past that, it has questionable value. Who would care about a few percentage point difference in a survey of this scope? It seems to have greater value within each product release date than it does within the product model. Since we're looking at an innovation-driven product, I see no value in comparing older product lines with current ones. For example, how many users trade up to newer model before the end of the survey period (4+ years)?
It's qualitative conclusions could have easily been gleaned from user forum reports, where users are eager to vent their frustrations.
It compares rather fragile, portable products with a relatively short design life span, that see a range of user environments and demands, are often user-maintained and repaired, and whose OEM as well as replacement & upgrade parts are made by a variety of sub-contractors.
Consumer Reports does a similar spreadsheet for cars & light truck reliability that I've used when shopping for a good used vehicle model. The data is more useful as the vehicle model ages, and you can look up older issues of the surveys. A good qualitative sense can be gained for each model. For a product that has a long useful life, this can be valuable information. For a computer, with a comparatively short life, the first 12 to 18 months are probably the most important (if not the first week!) and a company's reputation is built more by features & usability during this time span. You're definitely not going to get it rebuilt down the line to get another 6 years out of it!
Reply