Requirements specification of a BI solution is an independent discipline for a BI team and it often takes a lot of practice to capture all parts of the business' needs and thus deliver a fully functioning BI solution.
In Kapacity, we have often met companies where BI is synonymous with data extraction. These can be made by a BI department and delivered at fixed intervals to the users, or they can be prepared by the users themselves, who then have direct access to make extracts from the relevant source systems.
In relation to the task that users have to solve with the use of data, there are very often several steps that data must go through from the raw data extracts and until the users' task is solved. In headings, there are basically three areas where a real BI solution differs from a data extract:
-
Other data groupings - I have to divide the data into the groupings I need
-
Other data compilations - I have to make sure that the different data extracts can be compiled
-
Other visual display of data - I have to transform the data into the message that the management need
Other data groupings
Users often need to analyze or report in groups other than the divisions in which the source data extract is delivered. This can e.g. be order data, which is delivered as a long list of transactions, where the user needs to analyze how data is distributed by product groups, sales regions or customer segments. Data must therefore be divided and summed up in the relevant groupings before the user is done with the analysis.
Other data compilations
The classic example of data compilations that users need to make before they can reach their final reporting is compilation with data from previous periods. If the extracts you get delivered or make yourself are snapshots, then it may be necessary for the users to save history from previous periods, which they can then use in connection with the preparation of a trend for a development in e.g. order entry. Another example that Kapacity also often encounters is users who have to combine data from one system with data from another to create the reporting logic they need.
Other visual display of data
Many analysts provide input to specific report packages, which are provided to management e.g. on a monthly basis. This is typically in a completely fixed visual format and many people repeat the process of transforming a data extract into this fixed format every period, e.g. in PowerPoint or PDF, which is then sent out. If you count in a larger company how much time is spent on repetitive tasks like this, the result is often frightening.
When a BI team makes solutions for business users, there are therefore at least three good reasons to ensure that the needs of the user are properly covered: If the user is left with an experience that they have to complete the work themselves, there is reasons for dissatisfaction and for many companies also a really bad business case in terms of time spent on tasks that do not in themselves create value.
As a BI team, we must therefore always ensure that we spend time understanding the situation in which users must use data:
-
Ask about groupings, divisions and the typical filters in which data must be analyzed
-
Make sure that all the data needed to get the whole analysis in place is built into the solution. This also applies to history in data and history in master data
-
Also ask for the finished product to ensure that you go the last mile to give users the visual presentation they need