Why Oracle BI Needs Time to Mature

BI – Thoughts (7) – Cadran publishes a series of articles providing insights on Oracle Business Intelligence in combination with Oracle JD Edwards. Each article offers perspectives that aid in making informed decisions on implementing and using these systems. Previous articles discussed the star schema, facts, dimensions, and their interconnections. This article focuses on the development cycle and the journey toward maturity.

Organizational growth 

Organizations develop, like people, in stages. From the beginning to full maturity, a company goes through various phases. For many smaller businesses, automation is not an end in itself but a valuable tool. Automation helps make processes more efficient and effective. With Oracle Business Intelligence, you can add value to your organization through proactive management and alerting, known as management by exception. But just as Rome wasn’t built in a day, it takes time to fully deploy and broadly implement Oracle BI, meaning an organization should give itself that time.

Implementing and adopting BI takes time, which is always scarce in daily operations. A good BI solution can also require more work and time upfront. BI may reveal bottlenecks in data quality and process optimization that need to be addressed before the benefits truly become apparent.

Small Steps, Big Impact

Oracle BI provides insights at all levels—operational, tactical, and strategic. However, trying to use all these insights at once can overwhelm an organization. Small, manageable steps are key to success.

This principle applies to any solution, even when starting with a comprehensive tool like Oracle Business Intelligence Applications for JD Edwards Finance. Instead of trying to use everything at once, start small; for instance, by analyzing Accounts Receivables in Finance to gain insight into Days Outstanding. This phased approach keeps adoption rates high within the organization.

A Proof of Concept also works well when kept simple and focused. A powerful example is margin reporting, where you start with an analysis of revenue and profit per period by business unit.

The secret lies in small development cycles. Start with a simple topic and use real production data. Once successful, add a small component at a time. This way, reports and analyses gradually expand, and BI becomes increasingly valuable. For instance, adding Purchasing might start with a basic report on Open Purchase Orders by Status. Working step by step allows BI to become visible throughout the organization, boosting adoption.

The image above shows how an organization progresses through increasingly mature BI stages—from a system that primarily provides historical operational reports to a tactical and strategic tool with predictive capabilities.

Maturity of IT

When an organization has been using Oracle JD Edwards for some time, initial issues, security concerns, and master data problems are usually resolved. This provides a major advantage when implementing Oracle Business Intelligence, whether that’s Oracle Business Intelligence Applications or Oracle Business Intelligence Enterprise Edition.

In some cases, Oracle BI and Oracle JD Edwards are implemented in parallel. This requires careful coordination between both projects, as definitions in Oracle BI depend on the setup of Oracle JD Edwards. The main challenges here are:

• Bridging two constantly changing systems
• The absence of representative and historical data
• Incompleteness in situations, facts, and dimensions
• Inability to fully test and thus accept the BI solution

Therefore, it’s advisable to implement Oracle BI only once sufficient data is available, usually once Oracle JD Edwards has been in production for a while.

After Oracle JD Edwards goes live, data is often converted from older systems, but historical data is usually not included. However, historical data can be valuable for comparisons in a BI implementation. If no historical data is available, trends and time analyses will only provide added value after some time. Here too, you can start small and gradually expand: within a few days of going live, you’ll see trends from the past few days; after a quarter, monthly overviews offer valuable insights.

Testing and Acceptance – A Continuous Cycle

As Oracle BI runs on a representative dataset, the product becomes increasingly valuable. Moreover, BI often quickly reveals weaknesses in data quality—both in master and transactional data. Oracle BI not only contributes to better business processes but also to improved administration. Testing and assessing BI remain ongoing processes, and it’s recommended to perform a full user acceptance test at least annually to detect data quality issues early.

When testing and accepting Oracle BI, the following questions should be asked:

• Is each dataset correct when queried without constraints and without dimensions?
• Are all dimensions or attributes correct when queried without constraints?
• Is each dataset accurate when combined with all relevant dimensions and attributes?

“Correct” here means that data is reliable, complete, consistent, integral, and accurate. It’s essential to be critical of the numbers: even a well-functioning analysis can show incorrect figures at some point if there are issues in the source data. When decisions are made based on these numbers, it’s crucial that they are indeed “correct.”

In the next article, we will discuss how these principles apply to the reporting capabilities within Oracle JD Edwards according to the Oracle BI Continuum—from small to large, from straightforward to multidimensional, from historical to predictive, and from operational to strategic.