Is it worth using analysis services for a business with 1000-5000 sales
transactions per month?

I can't imagine that it is but I have been asked to look into it. The
point of OLAP is speed and I can't see any performance problem with
such a small amount of data.

What is the threshhold number of transactions per month where OLAP
might be worthwhile?

thanks,

Karen

look at the Project "REAL"
http://www.microsoft.com/technet/pro.../projreal.mspx
(there is more then 8billions of rows)

or the "old" T3 project (created in 2001)
http://www.microsoft.com/sql/evaluat...rabytecube.asp

these project talk about thousands records by day (or by hour)
a total of 8 billions of rows

I work on a "small" solution with between 5000 to 15000 rows / day without
any problem.
its a small hardware, we use a dynamic distinct count measure on this cube
and the performance is good.
another project consume 2 millions of rows by month, the hardware is "small
to medium" (database on a 4proc; cubes on a 2proc) and again, there is no
problems.

if you start to play with 1 billion of rows / month, then you have a huge
solution and you'll have performance issues ;-)


<karensundq (AT) yahoo (DOT) com> wrote

But the question was - do I have to few records...

You have clearly a point - OLAP is good with lots of records but what do you
want to do with the data
matters as well.

If you tell me that you have about 10-15 dimensions that you want to drill
down in certain locked hierchies or just combine and get information out of
and you have users that are able to use these analytical tools then OLAp
could be a way to instead of doing "report explosion" where you create 30
reports showing the same dimensions just in different setups.

Hope that helps.


<karensundq (AT) yahoo (DOT) com> wrote

:-)

I think I need to wake up :-)

well..
we have created some projects with small volume of data like yours (30 000
sales by year!!!).
the users are very happy.
playing with a cube is a great feature for these users.
these companies are small one with small and simple accounting software not
capable to provide any analysis capabilities and the reports available don't
provide any flexibility.
So OLAP cubes are a good solution.


"Michael Vardinghus" <michaelvardinghus (AT) hotmail (DOT) com> wrote

It's not just about the number of records but the number of cuts
(dimensions) and
whether you want users to be able to generate there own results from the
data or do you want to code t-sql rollup for the rest of your life?


dlr

<karensundq (AT) yahoo (DOT) com> wrote

some free software available for creating reports:

http://cam70.sta.uniroma1.it/TechnicalPreview/

can handle any number of records up to the computer memory
to hold all the report cells
Can create any layout (cross, vertical, mixed,...)