Sometimes I am approached by the information managers, with a grouse that functional heads are simply not ready to give away their strangle-hold on their functional level information structure. That is considered as a huge barrier to building an enterprise level data warehouse. My first question to them is 'why do you want to integrate?' OR 'what kind of information you need, which will use both business and financial information?' OR 'is there a pressing business need for the same?’.

I am not a heretic. A well integrated enterprise data-warehouse has significant benefits, and organizations should pursue the same relentlessly and with discipline. However, sometimes we fall so much in love with this end-objective, that we ignore dozens of milestones which need to be covered, before we achieve the same. The biggest challenge of an enterprise data-warehouse is the sheer cross-functional involvement needed both for business specs, design and testing.

Organization's today are trying to manage the conflict of de-centralization (for speed and localization) vs. having well coordinated integration for strategic alignment (for better expense management and consistency), in all spheres of their operations. The same conflict is well-manifested in the Data Warehouse world. The best way to manage it is to let the functional level BI agenda being fixed to a certain extent, before this is elevated to an enterprise data-warehouse level.

By letting the functional level BI to evolve, one can have the following benefits:

• Function feels the ownership of information management: The worst thing which you can do in pursuing your enterprise data warehouse dreams is to take-away the sense of business ownership for BI. Ask any manager, and he will name BI (or MIS in conventional terminology) as among the last things, he will give away to a central control. Information is key to managers and that's why they need their MIS team sitting next to them. A team, which not only generates the MIS but also, understands the business and day-to-day issues of that function. Therefore, one needs to build the ground for enterprise data-warehouse by letting functions take the ownership and control of their BI agenda, while providing a strong support, oversight and best practices. (Namely through BICC-Business Intelligence Competency Centre)

• Let the functions understand and appreciate the BI challenges: BI is not easy and functional heads soon realize that they need expertise and help. The best way to get the functional involvement in an enterprise data-warehouse is when one builds the credibility and knowledge-leadership. This does not mean that the central information management team will come-in only when they are asked for. As you give functional level flexibility for the business to pursue their BI agenda, one will need to have a strong central governance (for example, through BICC)

• Achieve the integrated BI environment at functional level: A function in medium to large sized organization can be an entity in itself. For example a products management function, could be having hundreds of products (or thousands of product variants), hundreds of pricing packages, IRR models etc...Similarly a large sales & distribution management function may have dozens of different channels, hundreds of different sales compensation schemes, thousands of locations etc...If one is able to create a robust functional level BI, which is built as per some defined enterprise standards, you would have done more than half the work for your enterprise BI agenda.

• Build the knowledge and skills: As a function pursues the BI agenda within their functional boundaries, super-users and super-designers emerge, who have solid business knowledge and they are battle-scarred veterans of building a BI from zero-base. As you start working on integrated the functional level BI to enterprise level BI, you will have cooked resources to drive the same.

The areas of central governance should ensure that business are deploying common tool-set (one needs to manage the licensing centrally), best practices for designing ETL, project management, data quality management practices. You can find a lot of detail on this subject in our chapter on Business Intelligence Management and Business Intelligence competency centre.