While self-service BI may appear as a panacea for the problems outlined above, organizations need to ensure that they are equipped for successful implementations. The four pillars for successful self-service BI adoption are technology enablers, process maturity, best practices and the right mindset (see Figure 2 at top left).
Technology Enablers
When selecting the proper tool, businesses should have functionality and scalability in mind; without these attributes, self-service BI may not yield desirable results. A tool should have rich and intuitive functionality that businesspeople can leverage and use to build what they need. A tool that does not provide strong base functionality requires a lot of effort to produce desirable reports/dashboards and will not serve the cause. Some high level key features are:-
- Templates for reports and dashboards,
- Intuitive features for report creation, e.g., wizard based, prompts, good online help,
- Rich visualization features,
- Search capabilities,
- Data federation to support data analysis from multiple systems,
- Integration of unstructured data,
- Personalized view,
- Collaboration features.
In-memory databases will support large analysis and rapid query response. As the design of the underlying data models may not have been suited for the analysis that users will perform, an in-memory solution and integration will help meet the performance needs.
It is also important to have a rich semantic layer (metadata documentation) to enable the business. The success of self-service BI depends on how you enable businesspeople to leverage what is already present in BI. Thus, a rich metadata layer that is business-user centric and user friendly is required.
Process Maturity
Self-service BI will not be successful if organizations do not have the right processes defined to leverage it. The key points to note are:
- Well-defined roles and responsibilities between IT and business: This helps clearly set expectations upfront. The service levels from IT must be defined as well.
- Data stewardship: Data governance becomes especially important because the business will be increasingly enabled to use the data. Ongoing data quality checks,, preferably automated, become useful. Additionally, data security processes must be identified to ensure that data is controlled with clearly defined ownership and that there is a process to make it available to as needed.
- Chargeback models: IT will still invest time and money once the solution is launched. Hence, chargeback models for cost must be defined to ensure that appropriate business units pay for the services they are getting.
- Clearly defined areas for report enablement: It is necessary to define the difference between IT-developed reports/dashboards and business-enabled self-service before launching the self-service strategy. Typically, it is recommended that IT should enable reports that are widely used across organizations, thus standardized. Also, reports for statutory and compliance purposes should be enabled by IT.
- Standardization of KPIs and a dictionary to ensure single version of truth: An undesirable fallback of businesspeople producing reports is that it can lead to data manipulation at the hands of users, and standard KPIs produced by a different group may show different values as a result. To avoid this, the corporate standard KPIs and dictionary should be defined.
- Governance and change management: There must be joint governance of BI by business and IT. As part of the governance model, organizations should define change management processes and also address compliance needs.
Following Industry Best Practices
BI organizations must continuously evolve, striving to follow the best practices in this area. These actions can include:
- Driving a culture of continuous improvement,
- Sharing best practices in usage by driving a culture of collaboration and knowledge sharing
- Providing the right level of business documentation
- Establishing a strong power-user community
Change in Mindset
- Drive culture of self-help within business teams
- IT needs to act as an enabler
- Provide the proper training with right level of details
- Pro-active support with data enablement
- Enable best of breed technological solutions and continuous technological improvements
Methodology for Self-Service BI
We recommend the following methodology for implementation (see Figure 3).
See related graphic 3: Methodology of Self-Service BI
Design Roadmap
Organizations need to assess if they are already prepared to undertake the journey or if they need to streamline processes, implementing enablers before they can start. In this phase, you will define a plan for setting processes, enabling the proper technology, performing prototype for a business group/region and then rolling it out to the entire business. It is important to have clear goals and objectives in mind before you plan. You must involve the business, including senior executives, in this exercise to understand expectations and roles throughout the process. Without business involvement, the initiative will not succeed. Also you need to define indicators for measuring success, which can be used to monitor and quantify improvements..
Define Processes
Self-service BI requires robust processes and governance. In this phase, you need to define processes that will be the foundation for successful implementation. The key areas include governance structure, roles and responsibilities between IT and business, data stewardship processes, support processes with acceptable service levels, cost chargeback model, security policies and processes.
Implement Enabler for Self-Service BI
Enablers include the proper technologies and tools to support self-service. Here, you need to select and implement tools that will support desired functionality and features. You will also need additional things such business metadata layer, security and data quality frameworks, including a master data framework.
