Thursday, November 30, 2006
NFL, P&G, Counting Surfers, That Excel Bug is Elsewhere, Bad Data, Other Stuff
Monday's column has been submitted... the common wisdom about postal rate increases affecting print shipments significantly to the downside is debunked. I know I'll get letters. The data actually suggest the opposite.
Saw this coming a mile away... NFL.com is going to start broadcasting games on the Internet http://today.reuters.com/news/articlenews.aspx?type=internetNews&storyID=2006-11-29T202925Z_01_N29339251_RTRUKOC_0_US-MEDIA-SUMMIT-NFL-VERIZON.xml
Major League Baseball did the same this year on MLB.TV
P&G is allocating more money to mobile marketing... it's now a legitimate spending category
http://adage.com/digital/article?article_id=113514
They're still having trouble counting the number of people on the Internet, and Newsweek has documented the chase and its problems. And there's a terminology problem... in our industry UV is a way of drying ink, but in the Internet business, it's "unique visitors." How can we keep this straight... even when we say "go to the web," printers think of a press, but their kids think it means to go surfing.
http://www.msnbc.msn.com/id/15789991/site/newsweek/
Each of the major measuring firms is eager to point out its rivals' flaws. Omniture's measures are based on technology that counts computers, not people—an approach that can lead to double-counting of UVs (for example, when the same user contacts a Web site from work and then from home). On the other hand, since panel data by definition are extrapolations, the counts produced by Net-Ratings and comScore can be substantially off, Web-site publishers say. "Where you get into craziness is taking 10 different numbers and trying to figure out which is correct," says Peter Daboll, Yahoo's chief insight officer. "What we're looking for is consistency within each vendor." Industry groups like the Interactive Advertising Bureau and the Media Research Council have been aggressively trying to establish measurement guidelines. As in counting couch potatoes, totaling up UVs will require wanna-be Nielsens in cyberspace to keep an eye peeled for the details.
The other day I posted an Excel file that Excel gets wrong because it has a bug. http://www.yousendit.com/transfer.php?action=download&ufid=6C2C711C618CA347
Other products have failed the test as well.
WordPerfect Office Quattro Pro = PASSED
Google online spreadsheet = PASSED
Thinkfree online spreadsheet = PASSED
Zoho online spreadsheet = PASSED
GNUmeric open source spreadsheet = FAILED
Evermore Office spreadsheet = FAILED
Ability Office = FAILED
602PC Suite = PASSED
goBeProductive Suite = PASSED
Softmaker Planmaker = FAILED
For the ones that failed, it almost makes you wonder if they've pirated MSFT's buggy code! I've written to Tony Bove (author of "Just Say No to Microsoft") to see if he's had results from yet other suites. I don't have EasyOffice or LotusSmartSuite handy.
I had to transfer about 500GB of data to a 1 terabyte external drive and found that my Windows systems could not "see" it. I tried it on 3 of them. My two Linux computers could see it, but could not perform any file functions on it. It turns out it was formatted as a Mac drive. Stumped, I Internet searched my way to software called "MacDrive" which allowed a 5-day free trial. The product worked, and did so very well. Info is at http://www.mediafour.com/products/macdrive6/bootcamp.asp and the free download is at http://www.mediafour.com/products/macdrive6/freetrial.asp
I'm always on the lookout for articles that describe "bad data" and how they got that way. The WSJ "Numbers Guy" column is always good in that regard. Here he tackles the misinterpreted number that 1 in 166 kids suffers from autism. One thing he does not mention is the incentive school districts have to identify kids with a variety of maladies to qualify for extra Federal dollars. Here, statistics with monetary incentives tend to create greater incidences. The result is that solving the right problem with the right approaches becomes difficult. In the case of the article, the 1 in 166 is justified based on new definitions and only in its proper context, but it rolls up so many categories that one has to wonder if it does anyone any good other than the shock value that has proved itself lucrative in fundraising.
http://online.wsj.com/public/article/numbers_guy.html?mod=djemnumbers
Speaking of cutting through junk data, I have always loved the site www.quackwatch.com that gets to the heart of the matter when it comes to medical and various homeopathic things. It's amazing how few things that people assume are true, especially homeopathic approaches, actually survive clinical trials; this site documents them. There's also my favorite science site www.junkscience.com. And as far as urban legends go, www.snopes.com is quite good, especially in checking out some of those chain letters about viruses that new computer users send about with such paranoia and panic. A quick visit to Snopes is all it takes to see if it's real or not.
Saw this coming a mile away... NFL.com is going to start broadcasting games on the Internet http://today.reuters.com/news/articlenews.aspx?type=internetNews&storyID=2006-11-29T202925Z_01_N29339251_RTRUKOC_0_US-MEDIA-SUMMIT-NFL-VERIZON.xml
Major League Baseball did the same this year on MLB.TV
P&G is allocating more money to mobile marketing... it's now a legitimate spending category
http://adage.com/digital/article?article_id=113514
They're still having trouble counting the number of people on the Internet, and Newsweek has documented the chase and its problems. And there's a terminology problem... in our industry UV is a way of drying ink, but in the Internet business, it's "unique visitors." How can we keep this straight... even when we say "go to the web," printers think of a press, but their kids think it means to go surfing.
http://www.msnbc.msn.com/id/15789991/site/newsweek/
Each of the major measuring firms is eager to point out its rivals' flaws. Omniture's measures are based on technology that counts computers, not people—an approach that can lead to double-counting of UVs (for example, when the same user contacts a Web site from work and then from home). On the other hand, since panel data by definition are extrapolations, the counts produced by Net-Ratings and comScore can be substantially off, Web-site publishers say. "Where you get into craziness is taking 10 different numbers and trying to figure out which is correct," says Peter Daboll, Yahoo's chief insight officer. "What we're looking for is consistency within each vendor." Industry groups like the Interactive Advertising Bureau and the Media Research Council have been aggressively trying to establish measurement guidelines. As in counting couch potatoes, totaling up UVs will require wanna-be Nielsens in cyberspace to keep an eye peeled for the details.
The other day I posted an Excel file that Excel gets wrong because it has a bug. http://www.yousendit.com/transfer.php?action=download&ufid=6C2C711C618CA347
Other products have failed the test as well.
WordPerfect Office Quattro Pro = PASSED
Google online spreadsheet = PASSED
Thinkfree online spreadsheet = PASSED
Zoho online spreadsheet = PASSED
GNUmeric open source spreadsheet = FAILED
Evermore Office spreadsheet = FAILED
Ability Office = FAILED
602PC Suite = PASSED
goBeProductive Suite = PASSED
Softmaker Planmaker = FAILED
For the ones that failed, it almost makes you wonder if they've pirated MSFT's buggy code! I've written to Tony Bove (author of "Just Say No to Microsoft") to see if he's had results from yet other suites. I don't have EasyOffice or LotusSmartSuite handy.
I had to transfer about 500GB of data to a 1 terabyte external drive and found that my Windows systems could not "see" it. I tried it on 3 of them. My two Linux computers could see it, but could not perform any file functions on it. It turns out it was formatted as a Mac drive. Stumped, I Internet searched my way to software called "MacDrive" which allowed a 5-day free trial. The product worked, and did so very well. Info is at http://www.mediafour.com/products/macdrive6/bootcamp.asp and the free download is at http://www.mediafour.com/products/macdrive6/freetrial.asp
I'm always on the lookout for articles that describe "bad data" and how they got that way. The WSJ "Numbers Guy" column is always good in that regard. Here he tackles the misinterpreted number that 1 in 166 kids suffers from autism. One thing he does not mention is the incentive school districts have to identify kids with a variety of maladies to qualify for extra Federal dollars. Here, statistics with monetary incentives tend to create greater incidences. The result is that solving the right problem with the right approaches becomes difficult. In the case of the article, the 1 in 166 is justified based on new definitions and only in its proper context, but it rolls up so many categories that one has to wonder if it does anyone any good other than the shock value that has proved itself lucrative in fundraising.
http://online.wsj.com/public/article/numbers_guy.html?mod=djemnumbers
Speaking of cutting through junk data, I have always loved the site www.quackwatch.com that gets to the heart of the matter when it comes to medical and various homeopathic things. It's amazing how few things that people assume are true, especially homeopathic approaches, actually survive clinical trials; this site documents them. There's also my favorite science site www.junkscience.com. And as far as urban legends go, www.snopes.com is quite good, especially in checking out some of those chain letters about viruses that new computer users send about with such paranoia and panic. A quick visit to Snopes is all it takes to see if it's real or not.
Comments:
Post a Comment
<< Home
<< Home
