My experience with data warehousing for site analytics is that it is very expensive. You are either going to spend a lot of money using a pre-packaged solution, or you are going to spend money (in the form of time or hiring) building onone yourself.
If you are just starting out, forego the expense of custom analytics. Use free off the shelf products. Most websites these days make do with Google Analytics. You can also get a lot of mileage out of the raw log files. Especially if you are familiar with grep
, sort
, uniq
, and other Unix command line tools.
I've worked with big websites that use data warehousing for analytics. None of them started out with a data warehouse solution. They all implemented it once they could afford it (and the insights they could get from it would make them more money than they were spending on itthe amount spent.)