There was a time when data models were only touched by experts in white coats wearing soft cotton gloves while carrying them around. The data models were kept in a vault that Fort Knox would have been jealous of. Business users could validate the models and comment on them, but the IT specialists were responsible for them. Complex and rigid procedures existed to maintain and change existing data models.
This time is over, because so much has changed. One of the key changes is the introduction of self-service BI tools, such as Qlikview, Tableau, and Spotfire. With these tools users can access any kind of data source, they can integrate data sources, and they can even transform and analyze the data themselves. All this can be done with minimal or possibly involvement of IT specialists.
To develop their own reports, users must develop their own integration logic when data from multiple data sources is combined, they must prepare their own data, and they must deal with the cryptic and technical table and column names assigned by IT. A problem is that for business users it’s not obvious that, for example, a table called DMW_SAL_23 contains sales data for the North-American stores. They need access to the data models that explain to them what all the data means. It’s their Rosetta Stone. For them a data model is an active source of information. Parts of a data model may be investigated by hundreds of users every day.
Without access to a data model, the following business problems may arise:
- Because users have access to so much data distributed over so many data sources, they may not even be aware that certain types of data are available. This may lead to missed and unexploited business opportunities. Users do not and cannot analyze data they have no knowledge of.
- Users may interpret the data incorrectly, and thus develop the wrong integration logic and reporting logic. This can lead to erroneous report results.
- If data is “hidden” in a large set of tables, it may take users a long time to locate the right tables for their analysis, and a long time to convince themselves that they are really using the right tables. In the process, they are losing valuable business time.
- Data structures may be so complex that users can’t locate the right data, although they know it exists.
- Users may not be sure of what particular data values represent.
The time has passed that data models can be locked away in the ivory tower (IT). IT and the business users have equal rights to the data models. Because it’s their data, the data models are theirs as well.
Self-service BI is a key aspect that forces IT departments to rethink how data models are developed, maintained, and managed. Data models must become active sources of information that are continuously used by business users and not only by the IT elite. Consequently, data models are now pivotal for an organization. Everything revolves around the data model. Without data models, organizations are data-rich, but still information-poor. For more information see this whitepaper.