opcoded: (aaron stack)
opcoded ([personal profile] opcoded) wrote2007-09-19 12:08 am

WTF?

Thanks to [livejournal.com profile] kuangning for posting this.

Livejournal announces stat gathering changes

I'm curious as to other people's opinions on this. On the surface, at least to me, it looks somewhat harmless. However, there's not much known about the service they're using to gather these statistics, so I'm a bit apprehensive.

There is an opt-out option, thankfully. Just go to the Admin Console and type "set opt_exclude_stats 1"

I'm just wondering other people's thoughts on this. I realize a lot of people will likely opt out of the data gathering, but I'm curious as to what the overall concensus is (to kind of foster some sort of discussion, at least. I need brain fuel!).

[identity profile] shippo.livejournal.com 2007-09-19 08:11 am (UTC)(link)
A lot of the comments with privacy concerns are pretty out there. LJ's already been able to collect this kind of usage data since, like, forever via their HTTP access logs. They're probably contracting their analytics through a third party since it's more feasible to buy something off-the-shelf than it is to build their own system to crunch access data.

6A will probably end up using clickstream data for one (or both) of two things:
1.) To see how well changes in the site perform, and how changes affect the way that people use the site.
2.) To deliver better targeted advertising to users. If they're able to have an idea where a user's interests lie, they're capable of delivering ads that a user may actually be interested in. I'm not really sure whether 6A is intending for LJ to have more of an advertising-supported model, but analytics would play a useful role in that.

[identity profile] etherpunk.livejournal.com 2007-09-19 08:12 am (UTC)(link)
Yeah, that's where the discussion on this issue has been heading. I'm not worried about it. The fact they give you an opt-out option is a nice deal, but if it's just anonymous data, opting out would sort of skew the results a bit, right?

[identity profile] shippo.livejournal.com 2007-09-19 08:39 am (UTC)(link)
Yeah, opting out would skew the data a bit, but it's probably not enough to be statistically relevant.