Here’s an example of a business intelligence project plan. It’s easy to resort to little shortcuts, but every project should have a separate definition and execution phase with a reflection period in between.

Before: Requirement Definitions

Here’s a starter task list for the definition phase:

• Gather requirement definitions
• Draft conceptual data model (unless familiar domain)
• Gain access to data
• Do initial tests / investigate data details
• Plan preliminary data model (dimensions & facts)
• Create better estimates
• Pause & Reflect
• Decide if the benefits are worth the effort

Project Implementation

Aim to run something through full pipeline with as little work as possible so keep it lean and agile.

Here’s a starter task list for the implementation phase:

• Launch the project & create initial documentation
• Hereafter always create the documentation when you complete a task
• Enable access and right privileges to solution environments
• Create infrastructure CI/CD pipeline
• Build infra
• Add solution to CI/CD pipeline
• Do basic solution setup / template
• Do reports and integrations (details below)

Do Reports and Integrations

Here’s a starter task list for each integration / report combination:

• Integrate from system to solution (e.g., from ERP to Azure Data Lake)
• Integrate from solution staging area to refine area (e.g., from Data Lake to SQL Server)
• Create initial data model (dimensions & facts)
• Build report version 0.9 based on users input in the definition phase
• Reconciliate data
• Leave time for the client to test the report
• Fix report bugs and complete version 1.0
• Leave time for the client to accept solutions
• Repeat previous steps for each data source / report

After: Monitoring & Onboarding

Here’s a list of tasks to do after actual implementation phase:

• Configure monitoring
• Configure report sharing and access levels
• Deploy and onboard users
• Agree on support model

What About Visual Design Work?

I know some people like to create a design of the final report using illustrating software, but I don’t think it’s worth the effort when working with Power BI. It’s just as fast or even faster to do the design directly in Power BI. Using another software for illustration would just double the work.

Quick vs. Difficult

With every project there are two different approaches:

1. Create something that is quick to do but valuable (low hanging fruit)
2. Start with the hardest or most uncertain task

If you need to present results fast, go with 1.

If you need to figure something out for the project to even make sense, go with 2.

First option creates momentum. Second option might kill momentum, but it’s a good way to make sure you’re not wasting money.

End

This is just a starting point for the project. There can be some variance depending on the details. Also, worth to note, that this skeleton has nothing to do with project method. That is a story for another day.

Have fun building stuff!