There are definitely people using BI/OLAP. I don't dispute that and I'm sure more companies are implementing these technologies all the time.

I guess my feeling is that star schemas and warehouses as complicated for most DBAs. They're hard to understand and it's a different way of looking at data. I think many DBAs could learn it, but it's a leap from the relational model and it's very simple method of putting data in tables. Getting to a more complex star schema, understanding meta data, tracking it, and successfully getting ETL jobs to run regularly is hard. And if you can't afford dedicated people to do it, you probably won't implement it or do it poorly enough that it won't be used.

The next leap is to cubes/OLAP technologies. To me, I think this is extremely complicated and beyond most people working as DBAs today. It's a complete reference change and not only does it require a great many more technical skills, it's a whole new way of working with the business. There's another great chasm of explanations that you have to bridge between how we technically view data, build measures, and actually understand what the mining models mean and what the business people are looking for. I'm not sure how many business people are great at spotting patterns or working with them.

I just feel that the current models, credit card fraud, sales trends, etc. are the edge of OLAP. They're like us automating general ledgers and sales entry. It's the bare beginnings of what can be done. I'm just not sure there are enough really intelligent people with the time to spend on digging into models.

I could be completely wrong and I'm looking forward to debating stuff at the conference and getting proven wrong. If I am, then it's great for SQL Server, which is the low cost platform in this area.