Comments for Dodgy Web and Blog Stats: Redux

Dodgy Web and Blog Stats: Redux

2007-01-04T22:27:18Z

As we all know, web statistics are an unreliable thing at the best of times. I posted on this issue back in March 2006, noting that Urchin and web server stats can be used by people to grossly exaggerate their statistics. Well now Nik Cubrilovic has posted on a similar story, this time involving Feedburner subscriber figures. The blog that Nik refers to has been a concern for quite some time, as it claims to have hundreds of thousands of RSS and email subscribers - and (imho) a similarly inflated page view count. These exaggerated web stats are helping to sell ads on that particular blog, which of course puts honest blogs (like mine) in a bad position. And frankly, this kind of thing reflects badly on the web content industry as a whole.

Thankfully, as Nik points out, Feedburner addressed the issue and the RSS subscriber stats of the blog in question are now showing the correct figure. I can verify that I checked this myself, before said blog removed access to his Feedburner chicklet. The rightful figure is less than 30,000 - not quite "hundreds of thousands" I think you'll agree.

It's great that Feedburner fixed this issue quickly (and thank you to Pageflakes for also working it out), but really this kind of web stats exaggeration has to stop, for the good of the industry. While Google is trying its best to stamp out click fraud on CPC ads, there is a more subtle fraud that can be perpetrated by fudging your web stats - either through quoting outlandish Urchin or web server stats, or by other means.

Comment from Pete on 2007-01-04

2007-01-04T22:56:23Z

Hey Richard,

I'm gonna post a response on Nik's blog regarding feeds, but I'll say for now that I strongly resent your claims that our stats are in any way "fraudulent".

In particular, I resent your allegation that our pageview count is inaccurate - advertisers are welcome to get a login to our sitemeter stats and check those for themselves. At the current time, the claimed stats are in fact *less* than than the actual figures (I haven't updated the sidebar text for a few months). For now we've decided not to make them public to competitors, since they would allow you to figure out which posts are performing well for us, and change your strategy accordingly. It is these pageview stats that dictate pricing for advertisers, and we believe them to be accurate.

Comment from Carrie Meyers on 2007-01-04

2007-01-04T23:13:47Z

Mashable should be ashamed, not out attacking people who called him on his fraud.

Comment from Jeff on 2007-01-04

2007-01-04T23:28:16Z

what a dirtbag.

Comment from Richard MacManus on 2007-01-04

2007-01-05T00:27:01Z

Pete, I couldn't care less which of your posts are performing well - and it certainly won't make me change my content strategy! We write about different things and in a different way, for the most part.

What I object to is how you used a plainly ridiculous RSS/email subscriber figure for *months*, knowing full well that it was untrue and not based on reality. I also think your page view figures don't add up either - because as I said in my post, there are ways to fudge those as well. But because I have no proof, that is purely my opinion.

I'm not in the habit of writing posts like this, but I do think it's an issue which the web content industry needs to address if it's going to be taken seriously.

Comment from Emre Sokullu on 2007-01-04

2007-01-05T02:30:08Z

I agree with Richard, this was obvious

Comment from Kahal on 2007-01-04

2007-01-05T03:07:22Z

Uh, yeah. You've just been caught with your pants down, dude. Not the time to get all outraged. Time to apologize and clean up your very shifty act.

Comment from Mark Varquez on 2007-01-04

2007-01-05T05:14:31Z

I am not sure about the pageviews, but it is obvious that he did manipulate his feedcount. His sponsors paid for sponsorships on his site based on the numbers he reported - I wonder how they feel! I know if I was buying ad space on that site, I wouldn't be too please right now.

Comment from Juha on 2007-01-04

2007-01-05T05:40:54Z

The issue of more exact statistics for web sites is quite important. I don't think anyone's quite cracked it - for instance, there seem to be plenty of people out there browsing sans or with site-selective Javascript; Sitemeter and Statcounter don't pick those up.

As an journo who writes for the Traditional Media as well, I can't help drawing a parallel to that world. Circulation statistics are pretty hard to fudge but then you have readership figures. These are odd extrapolations based on surveys of apparently statistically sound samples (even though it means door knocking in residential areas during the day to ask about business publications) and sell ads.

I've seen readership figures 10, 15 and 20 times higher than the circulation ones... think about that for a while.

It would be good if there was a generally accepted standard for counting that everyone adhered to.

Comment from kahal on 2007-01-05

2007-01-05T08:38:39Z

There was a very good way to measure feed count - feedburner - until this guy created a fraud bot to game it.

Comment from Dan Grossman on 2007-01-05

2007-01-06T05:37:05Z

If the "blog in question" didn't also turn off AwarenessAPI as well as removing the chicklet (which they probably didn't!), you can always access their subscriber count and other information via Feedburner's API. I use it to snoop on blog subscriber info all the time.