Newegg 1TB, 1.5TB, and 2TB Hard Drive Ratings – What is the best drive manufacturer

This weekend I woke up one morning and decided that I wanted to know if retail packaged drives had a lower DOA rate than OEM drives from Newegg. In all fairness, I think I was just trying to put off a 5am Saturday morning gym trip for a few hours. I ended up filtering Newegg’s hard drive category by internal drives of 1TB, 1.5TB, and 2TB in capacity. I then went through each result and recorded the quantity of reviews for each of Newegg’s awesome egg-scale along with a few other parameters. It turns out that Newegg did not have as much information on retail packaged hard drives as I had wanted, but I found some interesting results nonetheless. Anyone can do this survey, but hopefully this saves some time. For the most up-to-date information see Newegg.

The total number of reviews I tallied this weekend (5 June 2010) was 9591 and the average egg rating was 3.97 across the population. I only included Hitachi, Seagate, Samsung, and Western Digital in my tally, and only considered drives that had egg ratings. This yielded 36 drive models between 1TB, 1.5TB, and 2TB in size from the four major manufacturers.

The 3.97 was a bit higher than my initial hypothesis that the average Newegg rating would be 3.25 or so. I am an avid Newegg user (when they have the best price) and therefore am a bit ashamed that I was off by this much.

The Newegg egg rating system

While I was gathering data, I noticed a few patterns in hard drive rankings on Newegg. This is by no means an exhaustive list as I did not read every review and track results in a highly organized fashion. Instead it is a representation of my impressions of how users utilized the relative Newegg ratings.

5-egg

Generally this rating was given to drives for solid performance and non-DOA/ early failure reliability. Low heat and low noise seemed to also be factors that ranked favorably in 5-egg drives. Realistically though, this seemed to be the default rating for working drives that had no major issues.

4-egg

Perhaps one of the easiest ratings to describe, a 4-egg review seemed to have one factor or another enumerated in the 5-egg category working against it. For example, a drive that is otherwise great but perhaps generated too much noise for the users liking.

3-egg

The three egg category seemed to be a fairly eclectic mix of reviews. I saw some DOA or later RMA issues generating a 3-egg rating, especially with multiple drive purchases. Other reviews had things such as a great drive but expensive. I think price should not be a factor in the rating of a drive since obviously, the buyer made the economic decision to purchase the drive. Another driver for a 3-egg rating seemed to be multiple factors such as hard drive heat, noise, and performance acting in unison to lower the score below a 4-egg rating. One can note that the 3-egg and 2-egg ratings were used less, combined, than any one of the other three ratings.

2-egg

The 2-egg category in my casual browsing often included two cases. First, drive failures between DOA and failure within a few months. Consumers unhappy about drive failures that chose not to give drives a 1-egg rating seemed to pick 2-eggs. Second I saw quite a few 2-egg reviews where users purchased multiple drives and had one drive fail. For example a user purchased three drives and has two working great and one required an RMA. This is an understandable emotionally driven viewpoint, however it probably is a factor in lowering the average ratings. If those three drives were purchased separately there would be two positive reviews (say 5-egg) and one negative (say 1-egg). The average rating would be 3.67 and have the weight of three reviews. When reviewed together and given a two, those three drives would be given the weight of one review at 2.

1-egg

Usually these were DOA drives or drives that died within a few weeks of purchase. I did see some outliers like one reviewer that gave a Hitachi drive a poor review and 1-egg because of the drive being loud. I found that fairly amusing especially after using 15k rpm 3.5″ SAS drives, and the fact that I have many of those Hitachi drives running in various servers. While they may be a bit loud, if a drive works and performs acceptably then there is little cause to rate it below 3-eggs. Then again, that was an outlier. Multi-RMA drive purchases seemed to fall in this category.

Results

After pulling the data together, there were a few clear trends and a few that were less than clear. I did limit the sample to 1TB, 1.5TB, and 2TB drives with Newegg ratings which skew the results for manufacturer data, especially if a manufacturer is strong in lower capacity SATA drives, enterprise SAS drives, or notebook drives.

What is better OEM versus Retail drives?

This was my original question. I have been under the belief that retail packaged drives have lower DOA and early failure rates than OEM packed drives.

Taking the total population of drives, it seems as though the retail drives do fare slightly better egg rating wise than their OEM counterparts by about 4.5%. On the other hand, the sample size was very small for the retail packaged drives versus the OEM packaged drives. My next step was to remove Samsung from the population as Newegg only sells OEM Samsung drives on its site. Then I looked at the average egg rating by manufacturer.

Both Hitachi and Seagate see an increase in average egg rating for retail packaged drives versus bare drives while Western Digital’s retail drives receive higher bare drive average ratings. It should be noted that the proportion of Hitachi retail drives to OEM drives and Seagate or Western Digital OEM to retail drives is very disproportionate and the differences were not that large in terms of ratings. Between unscientific source data (user reviews) and varying population sizes (OEM v. Retail) I cannot say retail packaged drives are the clear winner. Yet the imperfect results speak for themselves. There is a higher proportion of positive (5-egg and 4-egg) reviews for retail drives. On the other hand, I was expecting to see something like an 8% difference so this was an interesting data point for me.

Highest rated hard drive manufacturer

Who is the highest Newegg rated 1TB+ hard drive manufacturer? Apparently, Samsung. Samsung being rated number one was probably the biggest surprise for me since I have been using Seagate, Western Digital, and Hitachi drives over the past two years but no Samsung drives.

Interestingly enough, the user review population fairly closely mimics the fact that Seagate and Western Digital have a combined 70-80% market share (depending on the quarter).

One factor that I think is very important with hard drives is the RMA process of the manufacturer (and not Newegg). Two years down the road, a good RMA process that provides an easy user experience and quick turn times becomes a major manufacturer ranking factor. If one is building a home server or a small business server, this is a factor outside of Newegg’s rankings that should be weighted very heavily.

Highest Rated Capacity between 1TB, 1.5TB, and 2TB

Are drives larger than 1TB less reliable than 1TB drives?

According to the Newegg rating system, the 1TB drives may be more reliable. Since anyone doing research into 1TB, 1.5TB, and 2TB drives has probably read a “watch out for Seagate 7200.11’s” post I decided to remove the Seagate 7200.11 1.5TB drive, that did have firmware issues a few months ago, from the population.

Here, two things become immediately obvious. First, there are over 2,000 reviews for that drive alone. That is more Newegg reviews for that one drive than all Samsung and Hitachi models in the population, combined. Pulling out that drive alone made the 1.5TB category jump in rating, but there are also about two and a half times as many ratings for that one drive than the rest of the 1.5TB population. I actually use well over a dozen of these drives and after the firmware fix, they have all been fine for fourteen to eighteen months.

Are the Newegg hard drive ratings accurate indicators of drive failure rates?

I personally do not believe so. Generally hard drive manufacturers sell drives to major OEMs at a discount in exchange for greatly reduced warranty coverage from the manufacturer. For example if one were to purchase a Hewlett-Packard computer with a Western Digital hard drive (just an example) and that drive were to fail, the consumer would generally ask HP for a replacement rather than Western Digital. To protect themselves, the major OEMs have contracts that specify a maximum annual failure rate or another reliability metric. This is a very common practice in the industry and is not exclusive to hard drive manufacturers. The reliability metrics are usually tied to some sort of financial recourse, often a penalty, if components fail at a higher rate. Generally these reliability metrics require manufacturers to produce drives with approximately 96% +/-3% 90 day failure rate. Terms of course vary quite a bit, especially between large tier-1 manufacturers and tier-2 manufacturers but it is fair to say that OEM’s expect over 90% of drives they receive to not be DOA.

Just to give an idea of the failure rates in Newegg ratings:

Given the general egg rating and associated review comments, one would guess that the vast majority of 2-egg and 1-egg reviews came from reviewers that had drives fail within three months. If this was indeed indicative of the broad population, every major manufacturer appears to see greater than 10% three month failure rates. This simply is not feasible given the industry margin and sales structure.

On the flip side, the Seagate 7200.11 example does show that users will respond to a drive issue with negative user reviews. In that instance, we have evidence that a firmware flaw caused failure and the Newegg user ratings were fairly low.

Conclusion

First off, please consider the source data, online user reviews. when reading this article. It is probably more indicative of the likelihood of users to post reviews than it is of hard drive reliability. Also, if I were to add a review for every drive I had, it would greatly sway the numbers. Just my Hitachi 2TB drive collection is equal to about 2.4% of the total Newegg Hitachi review population and about 1.4% of the total 2TB population. Frankly, hard drives probably need a minimum controlled sample size of 10,000 units to get meaningful reliability numbers. I woke up intending to ascertain if retail drives had lower DOA rates. I ended the morning convinced that the best use of Newegg hard drive reviews is as a starting point to see if there are any major issues such as the Seagate 7200.11 firmware bug (and for the record, the thousandth time over, the new firmware fixes the old bugs). If you have gotten this far, thanks for your time and hopefully you found this remotely useful.

For disclosure, I have purchased multiple Hitachi and Western Digital 2TB drives (Retail and OEM) respectively in the last few months from Newegg, all have worked flawlessly.