Does anybody have any suggestions as to how Business Objects would be
able to handle the following scenario:

Our users want to analyze the data kept in a SQL Server dimensional
model in a completely flexible way to determine trends e.g. look at
product sales within a time range, for a subset of products, for two
or three regions, etc. until they find a subset of records in the fact
table that have certain trends that they are interested in. If the
first analysis didn't show any interesting trends they would then
continue to choose different criteria (perhaps a different date range,
look at sales by customer demographic characteristics, etc.) until an
interesting trend was noticed and then choose this subset of records.

They then want to take this subset of records and analyze them further
to find out the reason for the noticed trends. This analysis also
needs to be completly flexible, e.g. they may want to look at this
subset of sales records for a different time range, based on product
shelf location, based on product packaging, etc.

This analysis may lead to a further subset of records being
identified, which would go through the same process. This can occur n
number of times.

The question I have is: what Business Objects products should be used
to achieve this very flexible, iterative approach. Note that the
criteria defining the initial subset, and subsequent subsets of
records is unknown at the start.

Any suggestions as to what system architecture would be used to best
use the BO tools to provide this subset selection and further analysis
capability and that makes the whole process seemless to the user (i.e.
they should not have to go through the untidy process of exporting the
subset to save it and then having to re-import it into their next
analysis session.

If someone knew of a way to keep track of the variables that they used
to define the subset so that this could be made into a repeatable
process it would be a real bonus!