Requirements gathering is a critical phase in any IT project, both technical and non-technical. It is understood that requirements always come before design. Requirements focus on "building the right system"; design lays emphasis on "building the system right". Considering the vast scope that BI offers, gathering requirements for a BI system is a challenging proposition. A business intelligence system should deliver an environment that is business focused rather than an application that performs specific functions. BI should be embedded into the business process that can be used to improve business processes.
Gathering requirements is a three-step process which is commonly known as elicit, specify, and test (EST). Requirement eliciting is the practice of collecting requirements from the stakeholders through interviews, questionnaires, brain storming and prototyping. The most important part of this phase is to identify what the users 'need' (that can be successfully built) by asking the right questions.
The complete description of each requirement along with non-functional requirements of the system is captured in requirement specification. Each requirement in the specification describes a property or characteristic which the BI system should possess in order to add value to the business users.
A requirement specification includes:
WHAT Describes the requirement along with features of individual attributes. The requirement traceability matrix is also included which enables tracking of requirements.
WHY Provides the rationale for the BI system, describing the value to be achieved
WHO Describes the source of requirement and the business users who will benefit from the system
No matter how well a BI system is developed, it will only do what the requirements ask it to do. Requirement testing is a very important step in the process. As part of requirement testing, each requirement must be validated to ensure that it is clear, unambiguous and feasible.
On a lighter note, a comic strip to summarize the importance of requirement gathering in a project life cycle:
More to follow in my next blog!
References:
http://www.miproconsulting.com/blog/2009/06/bi-requirements-gathering
http://www.kimballgroup.com/data-warehouse-business-intelligence-consulting/business-intelligence-requirements/
https://cours.etsmtl.ca/mti820/public_docs/lectures/TenMistakesToAvoidWhenGatheringBIRequirements.pdf
http://tdwi.org/research/2008/09/ten-mistakes-to-avoid-when-gathering-bi-requirements.aspx
Very realistic, Pradeep. Enjoyed the article and the comic strip, too!.
ReplyDeleteAdata warehouse is a repository of data used for the purpose of analytics. A data warehouse is a centralized, subject-oriented, time-variant, non-volatile collection of data integrated from operational data stores (ODS). A data warehouse stores data to support more complex analytical queries than a traditional collection of operational data stores. This is accomplished in a data warehouse by integrating data from multiple heterogeneous sources into a single database environment.
ReplyDelete