I can give an unbiased view of these two tools, having used both of them quite a lot in the last 2 years. I am a consultant, with no axe to grind, nor any partnership or encumbrance with either one.
[ul]
Business Objects Advantages:
[li]User interface for intermediate to power users is easier to use. Building simple reports is pretty easy in both, but with BO if you know SQL you can understand what you are doing. That goes for the universe design too.[/li]
[li]Llinked universes is very powerful, especially if you are selling SW that uses BO in the front end. It allows the customer to customize, and you to upgrade your core product.[/li]
[/ul]
[ul]
BO disadvantages:
[li]With SQL Server - bad code. It generates joins in the where clause, which in SQL Server 2000 will take a major performance hit. This will change with the new SQL Server, or so I've heard.[/li]
[li]The vendor. This is a major problem when choosing SW - all "best practices" talk about choosing a good partner for a long term relationship. The customer support and tech support is pretty awful. And let me warn you about negotiating a price. Negotiate your heart out. They promise the world, and will start with a very high price. You have been warned.[/li]
[li]I don't like the fact that BO manages each environment (development, test, production) in one metadata repository. This is bad practice. When I tried to manage them in separate databases on separate servers (much better methodology) it was fight to keep them synchronized, and we had many problems migrating reports and universes. If you have to choose, do it the BO way, and mitigate the risks of having dev and production running out of the same metadata database in other ways.[/li]
[/ul]
[ul]
Microstrategy Advantages:
[li]Performance and technical infrastructure. Multi-pass SQL will generally be faster than BO's nested SQL statements, or parallel data providers running independent queries. Also, since Microstrategy was built from the ground up in object oriented model, there is nice flexibility: you can embed filter in prompts, searches in filters, attributes in metrics and metrics or attributes or filter or prompts in other filters or prompts. It allows you to build a report where the user decides what to put in the rows (from your predetermined list), what in the columns, which metrics, and which filters.[/li]
[/ul]
[ul]
Microstrategy Disadvantages:
[li]There is more of a learning curve for the report developer. The results may be better, but the road to get there is not intuitive. Complex reports use "tool features", not complex SQL. You need to learn how to use them (with help from tek-tips).[/li]
[li]Cost. Microstrategy, from what I've seen, can be quite pricey, and too much for some mid-size businesses, though you can't take my word for it. Pricing models change over time, and my info could easily be out of date.[/li]
[li]At least in Desktop, I have found the single-query thread (albeit multi-pass) constraining in terms of building reports with multiple tables of charts. I think this is possible in MSTR Web, but I haven't used it. In BO, it is just another data provider.
[/li]
[/ul]
That's what I can come up with off the top of my head. This is not a scientific analysis, nor even a methodical one. To do this right (and it's not hard) use a full K-T matrix. That's Kepner-Tregoe Analysis. Search for it on the web, you'll find plenty of info on how to do this. It's really the only way to do it right, and the decision is too big to do it wrong.
That's my take, and I'd love to see the K-T matrix when you are done! Post it!
Dave
