You always hear about metadata management in business intelligence (BI).Your most important metadata may be your project metadata.This is Part 2 in the three-part series entitled Implementing the Business Intelligence collaboration (BI-C).Part 1 focused on setup.This article focuses on execution and covers project tracking, discovery, design, development and deployment activities as it relates to BI collaboration (BI-C).The intent of this article is not to rehash standard software development lifecycle methodology.Instead, it explores the details of executing a successful BI-C initiative and focuses on the data gathering and workflow needed at each stage of a project to be successful.

 

  

Tracking

 

Tracking a BI-C engagement must occur on a continuous basis to provide visibility, management and accountability for four key items:issues, tasks, status and BI project documents. Tracking all four of these is critical to the success of any BI-C engagement.

 

In order to track successfully, a mechanism or tool should be in place to allow for items to be captured, categorized, related, assigned and audited. Typically, project teams track items manually, through email or by piecemeal software tools.Accordingly, there is much reliance on individual accountability and ownership.

 

The following capabilities are required to facilitate successful tracking on any BI-C engagement:

 

  • Capture – Project team members must have the capability to capture issues, status, tasks and documents in a central location, with the ability to retrieve information, update and follow up as needed.All relevant data should be stored at the creation of the item, such as who, what, why, when, etc. Capturing this data is needed for effective workflow and project analytics/reporting.
  • Categories – Project items normally fall into logical groupings and should be tracked accordingly.Doing so provides a level of organization that facilitates oversight and resolution.For example, issues and/or tasks should be categorized related to project phase.At any given point in time, all issues and tasks specific to any one category can then be retrieved, reviewed and followed up on.This eliminates the granularity associated with reviewing tasks or issues one by one to determine who, how and what follow-up steps are necessary to proceed forward.
  • Relationships – All project items have built in relationships.There are issues related to tasks.There are documents that support the presence of an issue.Tasks are related to other tasks as dependencies.Tracking these data relationships is imperative to effectively track a BI-C engagement.
  • Assignment – The assignment of issues, tasks and status reports to project team members is another item that must be supported.Having the basic infrastructure to assign project items to team members and notify them when action is necessary will ensure that the entire team is aware of what to do and when to do it.
  • Auditing – An effective tracking tool will allow for auditing of all project artifacts.If an issue, document, task or status has changed, the tool will track this occurrence in a way that is easily accessible and reportable.

Establishing project tracking mechanisms will ensure all aspects of a BI-C initiative are visible, manageable and accountable.

 

Discovery

 

The discovery and analysis phase of a BI-C engagement is perhaps the most important.Without a thorough, organized and well-documented discovery phase, all other phases of the engagement will likely fail, or at a minimum be negatively impacted from a time, cost or quality standpoint.Discovery is the primary data gathering phase of the project and through an extensive interviewing process, requirements, features and releases will be defined and documented.

 

Interviews – Interviews are conducted with business users to gather and identify the requirements that will drive the BI-C initiative.Inconsistent interviewing procedures and guidelines can result in incomplete and inaccurate requirements.Successful BI-C interviews begin with creating key templates for interviewing business users and separately for interviewing technical users.These templates should be reviewed and signed off on by all impacted parties and stored in a central repository for common access/update.This will facilitate consistent and thorough interviews as your project team traverses the organization. 

As a part of the interview phase, all business, reporting, dashboard, scorecard, data management and technical information needs must be captured along with the person requesting the functionality. 

Requirements – A key aspect of defining and capturing requirements is defining requirements categories.These groupings should be distilled from cross-functional and departmental interviews.A user wish list, with all requests, should be documented.At this point in the BI-C engagement, project teams must avoid designing and simply focus on information gathering.Thinking ahead, for example, to determine technical feasibility or develop design solutions can not only derail requirements gathering activities, but also result in requirements censoring; possibly eliminating requirements needlessly or prematurely.

All requirements should be classified into buckets like reporting, technical,        dashboard or data.They should always be linked to the interview source and that relationship should be preserved and maintained throughout the engagement.Requirements should be prioritized.This is only after the requirements gathering process is complete.

 

Features – Once all requirements have been gathered, system features are defined, essentially transforming prioritized requirements into system functionality.The requirement feature link should be documented and maintained.Once complete, features must also be prioritized.

 

Releases – Based on interviews and requirements, releases represent BI features/functionality that will be implemented at the same time. For instance, release 1.0 may contain the following features: a data mart, reporting tool and ten canned reports that meet the highest priority needs of the sales managers.

 

The key part of a BI-C discovery phase is to store the genealogy of each of the items above, establishing a tangible link between business need and BI software releases.

 

Design

 

Key design activities that must be supported for a successful BI-C initiative include:

 

  • Identify, track, design data sources/targets;
  • Identify, track, design movement, mappings and transformations; and
  • Design report specifications and report mockups.

Designers need the ability to import source and target database definitions in order to design data mappings and transformations.They also need the ability to use the target data when designing end-user reports.By storing these items in a central repository that is accessible by project team members, work can be performed on a value-add basis.For instance, mappings will be built based off of the latest database definitions.Reports will contain the most current metrics that are in the database and can easily be mapped to their origin.Additionally, all items can be related to a task or issue so they are traceable.They can also be linked to their discovery counterparts in the form of requirements, features and releases.

 

Development

 

At a high level, development activities include building BI features and preparing the features for deployment. Many team-based integrated development tools exist to facilitate this activity.What is often neglected is establishing the link between the testing effort and the original business need.After ETL modules, reports, database code, etc. are deployed to a testing environment, it is important to not only successfully execute the test plans and test cases, but to track their progress back through the lifecycle.Fifty requirements, for instance, are currently being tested with twenty passing, 20 in progress and 10failed.By tying test plans/cases/executions to the discovery items, they can be tracked, and releases can be timed and managed more effectively. Leveraging this level of end-to-end integration may, for example, bring clarity to the fact that it makes sense to create a Release 1.1 for a particular engagement.This release may implement only 10 features (only a subset of those identified) that will meet the twenty high-priority requirements that are ready to go.

 

The other aspect of development phase collaboration is related to bug tracking.Having the capability to effectively manage bugs is well documented, thus the variety of one-off tools available that address this need.A successful BI-C initiative requires that bug tracking be integrated into the project lifecycle by establishing the relationships between bugs and test cases, features, requirements, etc.

 

Deployment

 

Deploying a release in a BI-C initiative is an important milestone, but only one step of an iterative cycle.New features should already be bundled and the next release planned.Successful BI-C initiatives allow for team collaboration and coordination, which in turn allows for quick, high-impact deployments that address the most important user needs in quick succession.

 

One of the positive side effects of having a central and cohesive project repository is that it becomes the foundation for your BI metadata.After a release is deployed, the BI-C project metadata becomes the main source for the BI metadata repository.

 

Standard project management best practices apply to any BI-C initiative.In addition, there are BI-C-specific activities, artifacts and best practices that should also be supported, including the capture of specific data points and workflow components discussed here.Many variables can impact project success – having the proper tools in place to support a sound BI-C implementation allows the project team to focus on the actual work at hand.Utilizing methodology, tools and project controls designed to support BI-C initiatives will undoubtedly help to increase the likelihood of success for your BI-C initiative.

 

The next article in this series Implementing the BI-C: Measurement, will focus on BI project reporting, metrics, measurement and evaluation of success.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access