• On TechRepublic: Five super-secret features in Windows 7
May 20, 2009 1:23 PM PDT

Acronis miscalculates data on users' bad backup habits

by Dong Ngo

The survey section where the percentages were taken out--the numbers just don't justify what Acronis said in its report.

(Credit: Acronis)

Acronis, a major vendor of backup software, released a report earlier this week stating that about 87 percent of computer users back up their data once every two or three months--way less frequently than recommended. This suggests that most of us live dangerously when it comes to backing up.

The survey was widely reported in the press, but it turns out the numbers didn't seem to go well with Acronis' report.

The percentages released to the media were taken from the question "How often do you back up your hard drive or files?" Results include: 48 percent for "once a week (or more)"; 55 percent for "2-3 times a month"; 81.5 percent for "once a month"; 86.8 percent for "every 2-3 months", 91.4 percent for "2-3 times a year", 94.6 percent for "once a year or less"; and 25.5 percent for "never."

It may be that I'm Asian and extra good with math, but I couldn't help but notice that when added up, the numbers total around 500 percent. Other ways to interpret the chart didn't justify the reported 87 percent, either. So maybe you're not as bad at backing up as some media reports told you you were.

Marc Mombourquette, Acronis' product marketing manager, confirmed Wednesday morning that there was an error in the data calculation. He added that "in the future, we'll (be) restructuring the survey to allow only one answer and double-checking the total number of answers with the number of reported responses." In the meantime, the company is working on a revised report on the survey and the possibility of re-doing it entirely.

Preliminary recalculations by Acronis show that 64 percent of some 6,100 participants who answered the survey said they back up once every two or three months; 36 percent say they back up once a week.

Even with these new, more encouraging numbers, I still believe most of us don't take backing up seriously enough. Generally, if you use a computer frequently, it's important to back it up at least once a week.

Apart from True Image, a product by Acronis, there are also quite a few free and powerful backup products I recommend, such as GFi Backup Home Edition.

Remember, it's better to be safe than sorry, and you can never be too careful when it comes to important data (unless, of course, you're happy being a statistic).

Originally posted at Crave
Dong Ngo is a CNET editor who covers networking and network storage, and writes about anything else he finds interesting. You can also listen to his podcast at insidecnetlabs.cnet.com. E-mail Dong.
Recent posts from Inside CNET Labs Podcast
HP to connect objects and people, sensitively
Photos: The Dell 2410 arrives
Windows Mobile 6.5 upgrade available for Samsung Jack
HP's LP3065: A typical 30-incher, but that's (mostly) a good thing
Western Digital enters enterprise-storage market
Inside CNET Labs 67: Kimbo Slice stinks
Atheros launches new energy-efficient WLAN chip for mobile devices
G-Tech refreshes high-speed, high-capacity external storage lineup
Add a Comment (Log in or register) (13 Comments)
  • prev
  • 1
  • next
by whizkid454 May 20, 2009 1:38 PM PDT
Those are probably consecutive percentages. For example, if you back up once a week, you also backup at least once per month, at least once per year, at least twice per year, etc. So the higher percentages should be higher than previous time frames...
Reply to this comment
by ngodong May 20, 2009 1:40 PM PDT
Yes, I thought about that too but you can't have 94.6% backup once a year AND 25.5% never back up.
by cyclingkitties May 20, 2009 2:20 PM PDT
@ngodong

You are not reading the questions carefully. 94.9% back up once a year OR LESS, which includes Never. The 5.1% left over are the 124 people who skipped the question.

There is nothing wrong with the chart or numbers at all.

I'm not Asian, but I did go to MIT, which may be why I get it...
by ngodong May 20, 2009 2:31 PM PDT
@ cyclingkitties, then the survey should say that 100% back up once a year or less, in which case what's the point of a survey?
by cyclingkitties May 20, 2009 2:46 PM PDT
As I said, some people skipped the question, and they are not counting those in the 94.9%. This is standard practice in multiple-question surveys. 75% of the respondents may answer one question, 95% answer another, and so on. It would be wrong to show each question adding up to 100%.
by cyclingkitties May 20, 2009 2:51 PM PDT
PS the better focus of this blog entry would have been the nearly 50% who back up at least once a week. In my experience, that's a pretty incredibly high number.

Your interpretation of "about 87 percent of computer users back up their data once every two or three months--way less frequently than recommended." is wrong also. What the chart shows is that about 87% of computer users back up AT LEAST once every two or three months, including the 48% who back up at least once a week or more.

In general, your entire article misunderstands and misinterprets the data and should be corrected. Not everyone reads comments.
by ngodong May 20, 2009 2:56 PM PDT
@ cyclingkitties you got it, all interpretations, including the 87% back up once every 2-3 months, were done by Acronis yes they were incorrect. That's the main reason for the blog. I think you should skip the numbers and read the text a bit :).
by cyclingkitties May 20, 2009 3:00 PM PDT
Ah, I didn't realize that first interpretation was theirs and not yours. Fine.

But you later say "I couldn't help but notice that when added up, the numbers total around 500 percent." which makes no sense. These are cumulative percentages and are not meant to be added up.
by ngodong May 20, 2009 3:06 PM PDT
@cyclingkitties , you are right, but that's just one (and most obvious) of the scenarios (it would take too long and be boring to bring up all of them -- not everyone like numbers like you :) ) to show how the chart might not make sense. I have discussed this multiple times with Acronis and we went through different ways to interpret it, none was flawless.
by upuaut May 20, 2009 2:12 PM PDT
There should be an option for 'once every more than a year' for those who back up once every 1.5,2,3,4, or 5 years, like i do.
Reply to this comment
by Jack K1 May 20, 2009 2:15 PM PDT
Back up? I back up nightly. I don't see any point in risking more than a day's worth of work. Seriously.
I image my drive monthly because I also don't have a day or two to waste rebuilding my machine. Drives fail. It's in their job description. It's like musical chairs. When the drive fails, the game is over, and whatever you had is gone, gone, gone.

Math skills? Marketing people aren't hired for their math skills. More's the pity.
Reply to this comment
by DJ_Lae May 20, 2009 2:28 PM PDT
They don't add up, but I can see how they did it. The Full and Incremental columns' percentages are calculated from the Response column. The way they're stacked is misleading, perhaps, but not incorrect. They won't add up, either, as some people will do full or incremental backups (or both).

Unless they really wanted to say how many people in each bracket performed (or didn't perform) a particular backup type, which seems weird.
Reply to this comment
by Banderasky June 2, 2009 11:10 PM PDT
I don't care about this arguement, it's just a promotion method of Acronis, you don't take notice in there is many link backing to Acronis. com, there more it's reported. the happier they are~~~
Reply to this comment
(13 Comments)
  • prev
  • 1
  • next
Subscribe to the Inside CNET Labs podcast

Subscribe to this podcast using an RSS reader other than iTunes

Subscribe to this podcast using iTunes

View all Inside CNET Labs episodes
advertisement

About Inside CNET Labs Podcast

"Inside CNET Labs" has two meanings. First, this podcast takes you behind the scenes of CNET's editorial process from a performance testing perspective. It will demystify CNET's performance testing process, allowing the listener an inside look at the good, the bad, and the ugly of performance testing. The second meaning gets equal attention--and sometimes more so--as we go inside the heads of CNET's San Francisco Labs staff, Eric Franklin and Dong Ngo, who will have opinionated discussions on subjects ranging from the insecurities of people to whether the existence of time can be proven. This is the stuff they've talked about every day for several years. Unfortunately for them (and fortunately for you, we hope), it's now being recorded.

Email us at insidecnetlabs@cnet.com.


Add this feed to your online news reader

Inside CNET Labs Podcast topics

More on Inside CNET Labs
CNET Labs posts in Crave blog
Dong Ngo's posts in Crave blog
How we test
The hosts of Inside CNET Labs
Eric Franklin Eric Franklin's colleagues once had the following to say to him: "Eric, you've been doing this performance testing thing for over 10 years now. How about you try something different?" To which Eric responded, "How about you shut the #%$@! up?" This candid attitude allows him to go toe-to-toe with the most extreme of personalities, including that of Dong Ngo. This bio was written by Eric Franklin.
Dong Ngo Dong Ngo is a knowledgeable, opinionated individual who wants to convince the world that he's just a normal person; but he hasn't had much success. According to him, this is because the world itself is abnormal. Dong loves traveling and is well-versed in several languages. He, unfortunately, is not so well-versed in English. Still, it's best to ask him questions. From networking and how to optimize your system, to turning a strange place into home or what the meaning of life is…most of the time, he has the answers. The question is: Will he make himself understood? Subscribe and find out!