Oracle Business Intelligence Applications (OBIA) is a complete, pre-built business intelligence solution with sample dashboards and reports, self-service framework, dimensional data model, and data integration adapters all available out-of-the-box. With a well-defined path for installation, pre-install, and post-install configurations, a complete data warehouse solution can be made up and running in matter of days.
Based on industry-leading tools like Oracle Business Intelligence Enterprise Edition (OBIEE), Oracle Data Integrator (ODI), and Oracle Database, OBIA provides a foundation for enterprise data warehouse and reporting solutions. Additionally, with pre-built adapters for most ERP applications like Oracle EBS, PeopleSoft, JD Edwards, Siebel, and several other third-party applications, OBIA contains a number of analytical applications like financial, HR, marketing, supply chain, and several more.
Being a canned, out-of-the-box solution has its challenges, though. Several OBIA implementations fail due to lack of adoption from end users. This usually is due to several reasons – improper implementation, data quality, vague requirements, user training, etc. This blog post mentions a few of these common issues and provides potential solutions for successful implementations of Oracle BI applications.
Oracle’s BI — a quick history lesson
From its earlier days of embedded data warehouse (EDW) and business intelligence system (BIS), the business intelligence offering has evolved significantly, but still remains closely tied around Oracle ERP (EBS) applications. During EBS 11i days, the PMF and PMV provided the initial BI framework. Given the customization challenges of this framework, Discoverer, with its integration to EBS and non-EBS data, was a valuable proposition for 11i reporting. Slowly, PMF/PMV evolved into DBI, or daily business intelligence, with its powerful drill-to-detail capabilities and materialized, view-based architecture. Around this time, XML Publisher tool was created to cater to high-definition reporting and with several other rich functionalities, leveraging the separation of XML for data and XSL for presentation.
While these products were natively developed, Oracle began to acquire companies like Siebel, Hyperion, PeopleSoft, JD Edwards, and many more. These companies brought with them many more advanced reporting tools and solutions like Siebel Analytics, Hyperion Essbase, and PeopleSoft EPM. Evolution of each of these within their parent company had its own history. However, after merging, Oracle bundled several of these tools together and came up with its initial offering of OBIEE (+), OBISE, OBISE1 (for Standard Edition), etc. These tools were bundled into different combinations, for different customers, with different pricing. Meanwhile, Siebel Analytics was transformed into OBIA with OBIEE as its reporting layer and Informatica and DAC for data integration. Finally, ODI replaced Informatica to bring out the latest version of OBIA. And now, the move towards Cloud – BICS and OAC.
It’s important to understand this product history and to point out that being “embedded,” “pre-packaged,” and “integrated” (to ERP), had been the DNA of BI solutions historically developed by Oracle. At the same time, Oracle’s effort was to add the latest and greatest, like the strength of multi-dimensional DB of Essbase, ELT of ODI and now the Cloud-based offerings.
So clearly, the biggest advantage of using OBIA is when your primary ERP applications are from Oracle, and you require a foundational data warehouse as a rapid graduation towards enterprise BI. Most customers go into OBIA implementations wanting to leverage these pre-built reports instead of custom development.
Reporting versus data warehouse
Challenges pertaining to BI implementations significantly vary based on the size of companies. On one hand, a smaller sized organization may have the culture of using BI applications for data dump into their personal data repository, while a larger sized organization may be using it purely for Oracle EBS reports with complete disconnect from its enterprise data warehouse. But, usually the primary purpose of selecting OBIA is a reporting solution on Oracle ERP and not as much as a data warehouse application. Please note that we discuss reporting and data warehouse solutions differently in this context.
As the name indicates, a reporting solution comprises of several methods with which different types of reports can be generated. A data warehouse solution is more about how the historical data is archived and stored, albeit for reporting purposes as well. Likewise, an analytical application is intended for use towards data analytics using all the available tools, while business intelligence can be any method by which a business organization derives information from its own business data. Academicians may differ on various definitions, but the key point here is that every requirement is approached by an organization differently.
10 Common Reasons for “not-so-successful” Implementations
So while there are several very good reasons as to why companies pick OBIA for implementation, unfortunately, there are several factors and issues that cause the implementation to fail.
- Installation is not implementation: Given that OBIA is said to be pre-packaged and out-of-the-box, the message is often misconstrued as if nothing needs to be done after the tools are installed and data is loaded by the ETL run. There are definite functional setups and post-install configurations to be completed even for an out-of-the-box implementation. Several solution vendors may take you for a ride with just an installation of the product, rendering it useless for the end users.
- ERP Reporting vs DW Solution: As explained earlier, often the actual need is that of an ERP reporting solution. When a company has a DW solution instead, several aspects are different. For example, the field names are different than your source, and the data is not a straight dump of ERP tables but functional extraction. This often becomes very challenging for customers, as the IT team continues to develop and build on a solution that does not address what you need in the first place.
- Performance: While a DW solution is expected to improve overall performance, a wrong implementation may have a complete reverse effect. The load time and run time performance needs to be handled appropriately. Often, a lack of understanding here can get very frustrating for both end users and IT support teams.
- Not a process application: The advantage with implementation projects with a focus on process is that once the process is successfully defined, the end user will use the OBIA solutions leading to a successful project. If the process is not defined correctly in analytical projects, even though the analytics may be right, it leads to a lack of use of the new application, causing project failure.
- Shortcuts: Often organizations tend to take shortcuts in using OLTP reporting as an alternative and in the beginning of the BI project. While this may work in the short-term, it starts to fail in a few years with the growth of data. By that time, the momentum for OBIA is lost and instead of deep analytics, there remains just conventional reporting.
- Lack of information (or reporting) strategy: A wrong tool, data source, or architecture can result in a completely dysfunctional reporting strategy. With the enormous choice available in selection of tools, the information delivery model needs to be properly defined. For example – OBIEE is not just BI answers or dashboards. It is a suite of tools, with other components like BI Publisher, Agents, Scheduler, etc. A proper strategy is critical for optimal use of these tools.
- Data quality & inadequate testing: The key to success of a BI application is the reliability of its data. Often dirty data from source applications flows into the data warehouse, causing incorrect information and eventual lack of confidence in the system. OBIA itself does not come with any DQM process out-of-the-box. This is usually left upon the customers to implement. Furthermore, lack of testing or lack of complete understanding of the metrics and calculations results in incorrect understanding of the reports output.
- No project champion: Several IT projects, including BI, fail when there is no executive-level support and oversight to the project. The value of analytics and the subsequent improved efficiency within the organization can only be achieved with top-level support. Often, the BI projects are started as technical projects and the businesses fail to adapt them. A C-level champion is always the most important factor for success.
- Big Data, AI…really?: Given the fast pace of technology change, often companies move over to a new solution without having implemented the existing solution properly. Many times, the cause of the issue is unrelated to the technology itself. For example, almost every company, small or large, has several siloed data repositories, and because end users mostly treat analytical application for data dumps into their personal spreadsheets, etc., big data, artificial intelligence, the Cloud, etc. is not going to change this. Any technology, new or old, needs to be implemented with the right strategy and plan to make it work.
- Training, training, training: Most of the end users are still evolving from mindset of reports as the form of information. Continuously working with end users, empowering them with information, and training them on advance use of analytics is critical for a successful OBIA implementation overall. This requires a significant change process from management and IT.
At Infolob, we are passionate about shaping our customers’ success. Data and analytics in particular is a core area of focus to achieve that objective. The BI team consists of highly motivated, top-notch, experienced consultants who have successful BI implementations in their track records. Our game plan ensures that BI implementations maximize ROI for organizations and creates delight for end users.
About the Author: Amit Bhatnagar has spent the last 25 years helping organizations get the right value from their business intelligence and data science initiatives. Being personally involved in over 50 BI implementations, he has closely observed and understood what makes a BI project successful and where it fails. He can be reached at email@example.com