Complexity of data model – Which solution?

PLM data model is usualy a huge topic. In order to have an idea of the situation, take an average user, and explain him why we should think about where to put a property, on the object or on the connection between another object. Or ask him the question of cardinality. Usually you get a big question mark from the end user, on which you should not spend too much time to interpret…

The point is there: while the PLM data model usually represents very well the business process, it is quite impossible to expose the data model itself to the users. The bad news is that our PLM solutions, built on standard applications, are by default presenting this data model. An important exception here is SAP in my view.

Consequence of that situation, is when you want to report data to the end users, or build KPI, to provide for example the same rendering as in excel sheets, you have to deal with the data model, which usually leads to complex queries, browsing a network of objects. A big network. At that time, you realise the complexity of your data model, having, as a simple example, to use the last released version, or the last version, or the current one, or…

Have you experience this situation, and if yes, how you deal with that situation? Is there another view to bring, which may be disrupting, but solve the situation?

I feel that another data layer is required, to provide a user view of the situation, implementing another data model, more suited to provide a more user oriented data representation. It is for example a view built in search engines and BI applications.

I am as well wondering about the solutions implemented for big network of objects, for example social networking application. Never in facebook the network is exposed to the user, while for sure it is behind.

What do you think?