The notion of an open source web analytics / business intelligence package is alluring.
The first is standards. While the WAA is making good progress in defining standards, and the vendors are starting to make some progress towards adopting those standards, I think we’re a long ways away from calling apple an apple and the apple being an apple. And not a Banapple.
The second is fault tolerance. Much of the existing codebase and tagging technology in web analytics is highly fault intolerant. Open source code, by way of repeated cleansing and improvement, would become less fault intolerant. In fact, if recent discourse around analytics has taught us anything it’s that simplicity wins and fault tolerance wins.
The third is scalability. Presently, there is a gap in the market between when a big-4 implementation (Webtrends, Omniture, Xiti, Coremetrics) makes sense, and when a company grows too big for Google Analytics. I invite people to come out and demonstrate why this is not the case, but I can’t make the figures work. There’s a step function at work here, and there’s no bridge in between. It’s right in that gap where I see the most innovation coming out of.
What’s a possible route to this divine scenario?
I can haz Pentaho?
Pentaho is an open source BI platform. It works well and it has a good ecosystem around it. It apparently even has a web analytics plugin. If the linkage between WA and BI is to happen anywhere, maybe it makes the most sense for it happen in that Pentaho ecosystem?