Jamie Thomson

Thoughts, about stuff

Gatineau, Live Spaces, Analysis Services et al…

leave a comment »

Today I took 10 minutes to watch a video on Gatineau, Microsoft’s web analytics offering that is in some ways a competitor to Google Analytics. You can see the video for yourself here: http://visitmix.com/Blogs/Joshua/gatineau-analytics-walkthrough-and-details/

I’m particularly interested in this because in a former life I was heavily engaged in web analytics work and was responsible for developing and maintaining a web analytics solution for one of the UK’s leading high street banks. One of the main things that my clients wanted at that time was demographic segmentation of their visitors but that was difficult to achieve because it would usually require people to login and on most of the sites I worked on logging in was not a necessity. Gatineau though is different. It uses a visitor’s Live ID if they have one (I assume it picks up the Live ID from a cookie on the client) to access demographic information for that site visitor which is then aggregated and made available for Gatineau clients to browse.

Gatineau metrics

The video above made mention of Gatineau tracking individual page views. That’s fairly useful information but in my experience site owners care more about browsing sessions and the metrics they measure browsing by are:

  • session length
  • number of page views in a session
  • session entry point
  • session exit point

There was no mention of session tracking in the video so I would hope that that was an oversight and session tracking is fully supported. I have applied for a Gatineau subscription so hopefully I shall find out.

Live Spaces integration

I talked recently about how Microsoft will look to leverage the Windows Live ID in as many possible ways that they can think of. I see no reason why Gatineau metrics shouldn’t be made available to people with an active Live Spaces site. Its fair to assume that the percentage of Live Space users with an active Live ID will be significantly higher than for other sites and I would love to log onto my Spaces home page and instantly see a demographic breakdown of all my site visitors. That is definitely attainable with Gatineau so hopefully the Live Spaces guys are listening.

Leveraging SQL Server

Aside from the Gatineau features I am also very interested in the technology infrastructure that underpins it. Those that know me from prior to beginning this blog will know that I am a SQL Server pro by trade and specifically I deal in SQL Server’s BI technologies like SQL Server Analysis Services (SSAS) and SQL Server Integration Services (SSIS). When I saw the video it was immediately obvious to me (I hope I’m right) that SSAS is at the heart of Gatineau’s analytical prowess and I wouldn’t mind betting that SSIS is being used to do the heavy lifting of data behind the scenes. Once Gatineau has been going for a few months I wouldn’t mind seeing a case study produced that talks about the sort of data volumes they are pushing because if Gatineau picks up (which it will) those data volumes are going to be unfathomable. If the Gatineau team want a SSIS pro to come and help them manage that data then I’m ready and waiting ;).

The use of SSAS raises some interesting questions for myself and I’m sure for people like Chris Webb as well.

  • How much data have they got pushed through their cube?
  • What is the granularity of that data?
  • How many members are in the Gatineau customer dimension?
  • Actually, never mind Gatineau customers, how many members are in the browser dimension? Last I heard there were 450m active Live IDs although the video does state that the presented data is aggregated. That doesn’t necessarily mean that the data isn’t stored at the granularity of each individual browser though.
  • How do they ensure that Gatineau customers can only see their own data and no-one else’s?
  • What hierarchies are they building on browsers and on Gatineau customers?
  • How do they partition the data?
  • What’s the ratio of known browsers to unknown browsers? This will be an interesting indication of Live ID distribution.
  • Are they managing the browser dimension as a type 2 dimension? If they are then the data is going to explode even more.
  • What tweaks have been made to the SSAS engine to enable serving up huge data volumes like these? I have heard on the grapevine that some serious changes have been made to the SSAS engine purely because of the work going on in Microsoft AdCenter.

These questions might not interest most people reading this but to me they are fascinating. I’m going to keep digging until I find some answers.

In the video the guy from the Gatineau team talked about enabling customers to export data to CSV files. Hmm..well..yeah..that’s errr….BORING! If this data is stored in an Analysis Services cube then I want to fire up my favourite MDX client and start firing my own queries against it. Better still, let me download to a local cube file so that I can slice-and-dice it to my heart’s content.

Moreover, will Gatineau customers be given access to the raw logs that Gatineau collects? There will be a wealth of information in those logs that Gatineau simply won’t be able to present because they have to cater for all customers. Access to that raw data would be a wonderful proposition for customers; let me try and explain why. Imagine Amazon were a Gatineau user, here is a URI taken from a link on Amazon’s homepage:


There is an abundance of information in that URI that means a lot to Amazon (particularly after the ‘?’) but that Gatineau simply will not report because they don’t understand what the different parts of the URI mean. This may be Gatineau’s Achilles Heel.


I’m looking forward to using Gatineau and as an when I have more information about it I shall talk more about it here. if you want to know more check out the video linked to at the top of this blog entry and check out the blog of Ian Thomas at http://www.liesdamnedlies.com/. Ian is one of the main guys working on Gatineau and he gives the 411 on Gatineau here. if you want to understand just how significant Gatineau is to Microsoft’s Ad business (the answer is "incredibly so") then read Ian’s blog entry Welcome to the world of Microsoft APS.


UPDATE: I was contacted by Thierry D’Hers from the Analysis Services team today. He tells me that in fact Analysis Services is NOT being used by Gatineau but it will be in the future. So good and bad news then.


Written by Jamiet

October 23, 2007 at 3:31 am

Posted in Live Spaces

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: