This is too long, and it took me too long to write, probably well over an hour (I had to write it in two sittings, though). But I don’t know what I could cut out. A lot of the problem was that the knowledge was slow to come into my memory, so I had to write a while to get it to come. In that case, it’s good I wrote it.

I should emphasize that OLAP is about ad-hoc retrieval. I think I confuse the issue by saying that data marts are designed around business goals, specifically around specific business questions or types of questions. Perhaps that’s how you start, but eventually you will need to support ad-hoc data retrieval and analysis. I need to include that.

The loss of attention time is actually 5 seconds, not 10 – 15 seconds. From HCI research. I didn’t include the evolution of data warehousing within an enterprise, but I hit everything else pretty much according to my notes.

If I want to take the Amazon example of leveraging data in data warehouses to create future business options further, I can also talk about changing the nature of competition. Anyone who competes with Amazon has to compete at least in price and in recommender quality, and with used book vendor/buyer support and population (rating system). You can talk about changing the nature of supplier relationships in the context again of suppliers because being able to funnel people looking for particular things to similar things that a supplier might also carry is a big plus for the supplier and encourages them to be more dependent on Amazon.

I did throw in some general themes of research this time; I’m assuming that this is enough. Simply due to space considerations, I don’t know how much more detail I could go into.