Often the importance of following due process when designing Business Intelligence systems is overlooked, the most critical step being the requirements gathering process. This may sound obvious but it is frequently overlooked or given too little attention. It really is the old proverb “failing to plan is the same as planning to fail”.
There is a lot of discussion in technology news currently about how a large percentage of data warehouses fail to deliver on business expectations, and my experience shows that if you do not have a documented, agreed version of the Business Requirements, you will inevitably end up producing something that looks completely different to what the business is expecting.
A colleague often uses the analogy:
“Building a Data Warehouse is like building a house”
Would you build a house without discussing your requirements with the Architect and having the plans drawn up first? Probably not, so why do so many organisations embark on building Business Intelligence systems this way?
So – yes, building a Data Warehouse is very similar to building a house. And guess what! Doing the analysis properly and having a documented and agreed Business Requirement Specification is the equivalent of creating the plans for your ‘house’. Having plans means that you are now focussed on delivering what the business wants, and just as importantly you can show these plans to the business and executive sponsors to garner their feedback, just as an architect would walk you through the plans they had created for your house.
Identifying organisational requirements is by no means an easy task….
It is essential that you have identified all of the stakeholders across the organisation and that you have discussed requirements with ALL of them. Stakeholders should preferably be at the Management and the Executive level.
Strategic KPI’s will come from the Executive members whilst Operational KPI’s will come from the Managers. Do not assume that they want to report on the same things or by the same views. It is possible that they may want to report by completely different reporting structures or dimensions. They may require data that is not in your pre-defined list of data sources.
It is very possible that some of the data that is provided to senior management and executives is derived from excel spreadsheets and hierarchies that have been managed and stored on desktops across the organisation. A typical example would be a Product Hierarchy – a financial services organisation may want to see products or services by organisational view vs. financial view vs. customer view.
Many initial implementations of data are focused on a single business unit or function within an organisation, eg. sales. This may be a great start but don’t forget the importance of having a high level view of how the organisation works across all areas.
Having workshops with all of the stakeholders will minimise the risks of ending up with Silo architecture. Not doing this may result in the need to rework a fair bit of what you have already built as Conformed Dimensions are the building blocks that are going to connect all the areas of your data warehouse and need to be built involving all the areas of the business. http://merolimited.wpengine.com/blog/important-business-requirements-and-seeing-the-full-picture/You need to ensure enough time and planning is put into developing and understanding how these are going to look and work otherwise you will end up with a house with no hallway to connect the rooms, or a data warehouse where we cannot easily determine a customer and all their interactions with our business.
Another way to increase success is to use a Business Analyst that is familiar with delivering Business Intelligence projects as their experience can help provide higher quality requirements in a shorter time period. Use these skilled staff and get them to mentor your internal staff too.
Remember – the business would like to see data as the Business works – Not according to how the Source Systems or Data Sources are structured.
“By wisdom a house is built, and by understanding it is established”