Even though Business Intelligence applications have been out there for decades lots of people still struggle with “how do I get started with BI”. I constantly deal with clients who mistakenly start their BI journey by selecting a BI platform or not thinking about the data architecture. I know it’s a HUGE oversimplification but in a nutshell here’s a simple roadmap (for a more complete roadmap please see the Roadmap document in Forrester BI Playbook) that will ensure that your BI strategy is aligned with your business strategy and you will hit the road running. The best way to start, IMHO, is from the performance management point of view:
- Catalog your organization business units and departments
- For each business unit /department ask questions about their business strategy and objectives
- Then ask about what goals do they set for themselves in order achieve the objectives
- Next ask what metrics and indicators do they use to track where they are against their goals and objectives. Good rule of thumb: no business area, department needs to track more than 20 to 30 metrics. More than that is unmanageable.
- Then ask questions how they would like to slice/dice these metrics (by time period, by region, by business unit, by customer segment, etc)
- Last but not least ask how do they envision tracking these metrics: printed reports, interactive visual dashboards, which metrics go onto which dashboard, which metrics are related to each other and should be tracked together on the same dashboard, etc. Keep in mind that only about 20% of such requirements can be envisioned up front, the other 80% will come only after you deliver the first version of an application. Follow Forrester agile BI best practices to ensure success.
- Build a simple BI benefits business case (increased revenue, lower margins, cost savings, etc)
- Then ask questions (and these you'll probably know) what operational, transaction and other source applications contain raw data to populate these metrics. Investigate the complexity of building data extracts and data integration processes. See if you actually need to build new OLTP applications to capture and maintain transactions or reference data. These will all contribute to the cost part of your business case.
Create a matrix all of your collected requirements ranked by business value and cost/complexity. Biggest business value at the lowest/cost complexity is where you want to start. Build momentum and good rapport with your business clients to ensure approval for the next, more complex, more costly projects.