Cognos modeling questions

[ColumbiaDiver]

Hi All,

We have just completed an upgrade from ReportNet 1.1 to Cognos 8.4 and we have a variety of systems that we would like to use with Cognos which thus requires that we create models for them.

In the classes I’ve taken so far my understanding of a data model in a nutshell is that it is for the purpose of interpreting the data from a system in such a way that it can be utilized by Cognos for the creation of reports, dashboard etc. Correct me if I am wrong.

As I am the only person currently assigned to our Cognos project we are thinking about bringing in a consultant to assist with the modeling. We have requested that they provide us with an estimate to model our various systems and in speaking to him he told me that before he can do that:

1. He needs to verify that our data is clean.

2. He needs to know what we are currently getting from the various systems, (for example we might have a user that takes data from a report from our accounting application and combines it with data from our staffing application to show trends in labor costs).

3. Needs to know what we are looking to accomplish with Cognos

4. Needs to understand our environment

5. Needs to interview our staff who utilize the various applications that we want modeled etc.

This is where our confusion comes in as it sounds like major overkill to us. Perhaps my understanding of the purpose of a data model is overly simplistic?

I am having a difficult time understanding what relevance these questions have to building a model for system X. Our feelings are hey we just want Cognos to be able to talk to the application and have the tables and fields defined in Cognos so that we can access that data with Cognos and as far as how we intend to utilize that data that is something we will decide when the system and Cognos can successfully talk to one another.

i would really appreciate any information that you can pass along that relates to your modeling experience and if the level of details we are being asked to provide is "truly" necessary as we want to get the most bang for our buck and not spend money on areas that don’t directly relate to the modeling of the systems we want to use Cognos with.

Thanks

Gordon

 

[blom0344]

Looks to me your consultant indeed knows how to handle this! This is not overkill, I can think of extending the list even a bit further!
Modelling is not about importing your database structures into the Cognos modelling tool (FM) and then writing your reports. To achieve proper results you need to isolate the parts of your application databases that are of interest for reporting and remodel accordingly.
I will not go into detail, but a LOT has been put on paper/web on this subject.

Ties Blom

 

[drlex]

I agree with Ties - whilst you may think it sounds overly involved, it is only a sensible list, given the extent one can make of BI tools. Either get a second consultant's opinion or ask the first one if you can talk to his last two clients - I'm betting this will assist your deliberations.

(The first point is a doozy - it's no longer a surprise when one finds that the statement 'the data sources are fine and complete' should be read as 'the data is from a variety of sources, some on a deprecated DB system, manually and automatically updated; oh and often amended upto a year later when we find a boo-hoo' )

soi là, soi carré

 

[ColumbiaDiver]

Thanks for the feedback guys. Good to know that we are on the right track.

Cosultants are expensive and as we have limited manpower, (were a pretty small shop), my CIO wants to make sure any consultant we hire is focusing on the job and hand and not asking questions in the hopes that he will later be asked to help with reports/dashboard/scorecard/metrics creation as we do not need any outside help for that.

Cheers,

Gordon