John,
Here is a compilation of the latest (July 2003) PC Magazine service and reliability results.
Cheers,
gary
-----------------------------------------
DESKTOPS:
I had to call [Dell] tech support several times, but I eventually reached an extremely knowledgeable tech, who solved the problem quickly." —David Cantor
For the 12th time in 13 years, Dell earns an A or A+ for desktop PCs. This year, Dell has company at the top: Apple, not on the charts last year because of its modest share among survey takers, joins Dell with an A+. Sony moves up from a B to an A, and newcomer ABS debuts with an A.
We don't know whether the HP/Compaq merger was meant to benefit users as well as investors, but it hasn't—yet. HP and Compaq PCs scored D- and E overall last year, respectively. This year they're both E's, no matter how you slice the numbers, except for first-year owners, who give Compaq a D. For measures of technical support and repair experience (such as time waiting for help or getting a repair right the first time), both Compaq and HP are worse than average on 12 out of 15 items. By comparison, Dell is better than average on 7 measures, average on 7, and worse than average on 1 ("Tech understood my problem").
The other troubled vendor is eMachines, which receives its third straight E, with no indica-tion from first-year-of-ownership scores that respondents have noticed its moves to improve tech support and warranty handling. Acer also gets an E (it got a D in 2002). NEC, omitted last year because of a lack of responses, reappears with an E.
Increased participation in the 2003 Service and Reliability Survey has given us at least 50 responses (the minimum for inclusion in our resulted results) for 13 vendors, up from 9 last year. A change in our methodology renames the "locally built" desktop PC as a clone (also known as a white box) and adds the self-built category. Because these are different from typical PCs—how do you rate satisfaction with tech support on a computer you built?—their scores don't contribute to the category averages. That said, compared with mass-produced PCs, self-built systems get an A+, and clones get a B+.
Self-built PCs are used at home almost exclusively (93 percent), clones to a lesser extent (73 percent). Clones used in work settings have worse-than-average reliability, dropping to a D+.
Micron/ MPC, a B last year, falls to a C. Gateway has the most differentiated grades. It gets a C+ overall (versus B+ last year), D+ at home, E at work, and A for first year of ownership. Its C+ results from average scores for overall satisfaction and satisfaction with reliability, a worse-than-average score for the number needing repair in the past year, and a better-than-average likelihood of buying again.
How can an overall grade be higher than the separate home and work grades? In each slice of the data, the companies are rated against their peers in that group. Many of the vendors without enough responses to appear in the work category have poor overall scores. Their absence raises the bar for home and work machines.
Gateway's A for the first year of use results from average repair scores and better-than-average scores on the other three key components of the grade. Because Gateway PCs tend to be older (a sign of flat or falling sales), their approval falls off more precipitously than the others going from first-year to overall.
Sony's C for first-year systems, versus A's overall and for home PCs, is due to its four key scores (except repairs), which, while better for the first year, still do not match the competition.
---------------------------------------------------------
NOTEBOOKS:
"I can take my Toshiba laptop wherever I want— even backpacking—and I know it can handle being knocked around." —Mike Foster
For the most part, Service and Reliability grades for notebooks are the same as those in 2002. Apple, IBM, and Toshiba continue to receive A's, and five out of nine other notebook vendors' grades are also unchanged. Even the problems holding back other vendors are little changed.
Notably, Dell still has a high percentage of units needing repair in the past 12 months. As in 2002, only Dell and Compaq have worse-than-average scores in this category; in both cases, about one in four notebooks needed work. That keeps Dell at a B+; the company does so well on all the other measures that even an average score for repair frequency would have boosted Dell to an A, a score its notebooks last saw in 2001.
We don't have the space here to print repair records for specific product lines, but Dell's business-oriented Latitude line appears to need more repairs than the consumer-focused Inspirons; with Compaq notebooks, the numbers are more respectable for Evo business notebooks than consumer Presarios. Compaq's poor repair frequency record keeps it mired at an E. Sony remains a B-, Gateway a C, Micron/MPC a D.
Average scores for reliability and units needing repair push HP from an A to a B. Fujitsu drops slightly, from a B to a C, as its record for frequency of repairs is now only average. WinBook returns to our survey with a D. Two E's last year, Acer and NEC, have improved to D and B, respectively.
Perhaps because notebook makers draw from the same pools of components, reliability is the category with the least variability. NEC, Sony, and Toshiba are the only vendors to score better-than-average on units needing repair. Overall satisfaction is the measure where makers are hardest pressed to be worse than average; there only Compaq has missed the boat.
Of the three A vendors, IBM has impressed users the most with its tech support and repair service (time on hold, repairs done right). IBM rates better than average on 14 of 15 measures, Apple on 6, and Toshiba on 1.
The three top vendors overall—Apple, IBM, and Toshiba—also receive A's for home use, as Apple and IBM do for work use. Toshiba's grade for work use is a D, where its overall satisfaction drops to worse than average; the other three key measures determining the letter grade are average.
First-year IBM notebook users give average scores on two satisfaction measures: frequency of repair and likelihood of buying again, leading to a C+. HP, a B overall, gets a D among first-year users due to worse-than-average scores for satisfaction with reliability and plans to buy again.
Deciding between a notebook and a desktop? Readers say notebooks require fewer repairs than desktops: 21.4 percent versus 24.3 percent (last year the gap was 6 percent). We ascribe that to the larger amount of tinkering users do to desktop PCs (although notebooks do get dropped occasionally).
-----------------------------------------------------------
SERVERS:
"My experience was complicated by an incompetent [Gateway] phone tech who appeared to know nothing more than to follow steps in a symptom/ resolution manual." —David Gore
In past years, we asked readers to rate their most frequently used servers. But large companies typically have one or more servers dedicated to file sharing, e-mail, database applications, and serving Web pages; for such organizations it's impossible to identify one most frequently used server. Therefore, this year we asked readers to rate whole server families, such as Compaq ProLiants and Dell PowerEdges. This gives a more accurate picture of server service and reliability.
Our survey results on servers have much in common with the desktop results: Dell and Apple are at the top with A's, IBM in the middle, and HP and Compaq at or near the bottom. Dell is the only vendor with better-than-average responses on all four key questions making up the letter grade: overall satisfaction, satisfaction with reliability, frequency of server failure, and likelihood of buying again.
As elsewhere, scores for clones and self-built machines are not factored into the category averages.
Sun, whose scores include ratings of its Cobalt appliance servers, receives a B, dropping from last year's A because it scores better than average only on frequency of server failure. Last year it was better than average on three of four categories.
Apple didn't get enough responses to appear in last year's survey; this year it scores better than average on three out of four measures (the exception: frequency of server failure).
IBM slips from B+ to C+. Compaq, D last year, is D+ this year. Only a worse-than-average score for likelihood of buying again keeps Compaq from a C; the plus comes from better-than-average scores for technical support and repair satisfaction. The Gateway and HP scores remain E's.
Self-built servers receive an A; clone servers, D, because of a worse-than-average score for likelihood of buying again.
The question that elicited the most varied responses involves the usefulness of automated-response phone systems. Compaq, IBM, and clone systems score better than average, meaning that fewer respondents encountered unhelpful automated systems; Dell, HP, and Sun are rated average; Gateway and self-built servers (where the phone response systems were likely those of component providers) have worse-than-average scores.
For servers in the first year of use, the pecking order changes considerably. Compaq improves to a B, aided by better-than-average satisfaction with reliability; HP rises to a C. Just-average scores for satisfaction with reliability and frequency of failure drop Dell to a B. A worse-than-average score for the server failure rate drags IBM to a D. Note that while the spread among the scores for server failure is statistically significant, the best and worst aren't far apart, separated by only a few tenths of a point on a 1-to-10 scale.
In tech support and repair service, we broke out scores to indicate whether systems are in or out of warranty (results not charted here). Again, there is more similarity than divergence among vendors.
Dell users who have called for phone support while their systems were under warranty rate their experience worse than average on the length of time it took to get questions answered, on having questions resolved in the first contact with tech support, and on getting tech people to understand the problems. No other manufacturer suffers any worse-than-average responses |
