Comments on: Web Analytics Tracking on a Facebook Page

By: tgwilson_php

tgwilson_php — Fri, 05 Mar 2010 19:48:16 +0000

Thanks, Gary. My read of that article is that it is much more geared towards “Facebook media analytics” rather than “Facebook fan page analytics.”

By: Gary McDaniel

Gary McDaniel — Fri, 05 Mar 2010 17:33:35 +0000

You were wondering about the status of Omniture with regards to Facebook. Here’s your answer:

http://www.mediapost.com/publications/?fa=Articles.showArticle&art_aid=123533

By: tgwilson_php

tgwilson_php — Thu, 04 Mar 2010 13:45:07 +0000

Justin — I realize that apps can require the “do you want to allow” question, but not all apps require that. For instance, an FBML app does not. I’m assuming Webtrends is using apps in a way that necessitate this for legitimate reasons — but it’s still an unpleasant blip in the user experience, IMHO, especially if it’s not clear to the user why that access is needed based on the content they’re trying to access.

On the WebDigi front, that’s a good call-out, and it’s why I suggest using an approach that includes the generation of a quasi-random number as GA does with the utmn parameter value generated normally. I am curious if using .njs approach in Webtrends with a gratuitous random number-based parameter tacked on the end would reduce caching issues with that approach. Would it?

Chris — I’m not sure. Maybe Justin will chime back in and clarify on the Webtrends front. Certainly the Google Analytics hack — the WebDigi one and the one described here — are not geared towards this. I would expect that doing this *would* require some level of permission from the user (as the Webtrends application-based approach does), and you’d want to be *very* familiar with the Facebook terms of service to make sure that any use of that data is appropriate. There would certainly be some juicy information there — the ability to segment traffic by age, gender, and even geography (especially if there are a lot of local events-type content) would be great, including linking that back to on-site activity.

By: Chris Grant

Chris Grant — Wed, 03 Mar 2010 14:50:18 +0000

I was thrilled to find this post. Thank you for doing it! I’ve read the source material (WebDigi and WebTrends) but this is what makes it start to click.

Am I correct in thinking that neither of these methods handles cookies well enough to allow a visitor’s FB data to be joined with their GA or WT data, i.e. merged into a single session?

By: Justin Kistner

Justin Kistner — Mon, 01 Mar 2010 21:27:00 +0000

Great write up, Tim. I wanted to clarify that apps are the object that asks the “Do you want to allow” question, not our tracking. You can also track tabs, ads, and links in addition to the apps. Our tab tracking does not use the no script version of our tag, it uses a new method that we’re patenting.

It’s also worth pointing out that the GA hack WebDigi developed relies on an image call, which Facebook and your browser caches, which prevents complete data collection. That image caching hurdle is why we developed a new tagging method.

We do use our no script tagging method on Flashbook on tabs, but I think you already knew that