In the previous articles, I discussed the different aspects of BI as a service. Now that you are familiar with the terms “business intelligence” and “services”, I will take the next step and discuss: How do we offer business intelligence capabilities as a service?

VisionWhat is the vision of your organization? What is the role of the business intelligence team? What does the business expect? Without answers to these questions it’s difficult to set up a road map to BI as a service. Your business intelligence department needs to function in a stable environment and must be trusted by the business community. BI as a service is certainly not the ultimate solution to solve organizational issues or to overcome a bad working business intelligence environment.

BI as a service doesn’t mean that you need to outsource your business intelligence work to external parties. Your internal BI team plays an important role in business intelligence as a service; they can deliver these services or outsource them.

Assessment

If your vision is clear, an assessment of your current BI environment is the next step. You need to examine the people, processes and technology involved.

The “people” part is a very important one. Does your team have the necessary skills to function in a service setup? Working with a service-oriented approach requires additional skills, certainly if you plan to ask services outside (outsource) your organization. In a traditional setup, organizations allow service providers and consultants to work with their own framework and/or technology. In some companies this approach is considered a differentiator. In a service-oriented approach, the business intelligence team is in the lead and will push for their way of working.

From a “process” perspective, you need to have a working BI process in place. Your end-to-end information chain must be under control and work effectively. Most companies are struggling with their business intelligence processes. The line between development and production is very thin, ad hoc decisions and changes to the BI processes are more standard than the exception. And even worse, the direct access to the production environment seems to be a privilege for business intelligence experts.

In the current business intelligence market it’s very difficult to cope with the fast changing technology market. Mergers and acquisitions are happening every week. Today you have SAP, tomorrow you need to integrate with Business Objects, what will be next? Therefore, the result of your assessment should also describe the vision and expected evolution on tools within your business intelligence environment.

One aspect you need to take into account as well is the evaluation of past projects. What are the lessons learned in these projects? These results are important input when defining your service level agreement (SLAs). SLAs are of major importance on guaranteeing the quality and availability of your business intelligence environment in a BI-as-a-service mode.

Stabilize Your BI Environment

A stable BI environment delivers system availability, system performance, data quality and security in a consistent way. Business and IT must have a clear understanding and mutual agreement on the expectations.

When moving to business intelligence as a service, data quality and privacy become an obstacle. These must be understood and accepted by the business otherwise you can’t define the service you want to be delivered.

System performance is one of the critical success factors for BI as a service. The system performance must be acceptable and agreed upon. Don’t expect that BI as a service will solve performance issues during month closing. These issues will put serious pressure on your SLA agreements.

What about system uptime? The requirements will be written in stone in the SLA agreement. When business intelligence is used as an instrument to run a company’s business, system uptime must be close to 100 percent.

A stable environment means also that standard procedures and policies must be in place. For example: How do we handle security? What about naming standards?

Service Level Agreements An effective service level agreement is absolutely vital and extremely important to any organization to ensure effective engagements for any service or product delivery. The SLA is important because it sets boundaries and expectations for both the BI team and the business. It clearly defines the level of services expected between two targets and ensures that these targets are met within an expected budget frame. An SLA must be realistic and measurable. A good SLA helps the business intelligence team (or partner) promise what is possible to deliver and deliver what is promised. A good SLA has five key aspects:

• What the provider is promising
• How the provider will deliver on those promises
• Who will measure delivery, and how
• What happens if the provider fails to deliver as promised
• How the SLA will change over time

The provider can either be your own business intelligence team or an external partner. Big BangMoving to BI as a service is like in every data warehouse project: “Think big, start small”. A step-by-step approach is the best way forward. In the first phase, you will stabilize your BI environment and create a basic foundation for the future. You need to define roles and responsibilities, agree upon BI processes and select the technology and partners used for your BI services.

To reach this situation, it is helpful to work with consultants who know your environment and can support you in reaching a stable environment in a minimum amount of time. Because of their knowledge of the BI environment, they are best placed to bridge the gap between the current environment and the expected environment.An easy topic to start with is an SLA agreement with the finance department. You can start with providing SLAs for the finance department. Make clear SLAs to deliver correct end- of-month results in a timely way. From this first SLA you can take the lessons learned and move to the next step in turning your BI approach into BI as a service.

After a few months you will be able to evaluate if BI as a service is a workable solution. If the result is positive you can start working with external partners to deliver all or part of your services.

Recent articles by Ivan Schotsmans

• BI-Day Hot Spots
• Business Intelligence as a Service: Service Level Agreements
• BI as a Service: ITSM for Business Intelligence
• Introduction to BI as a Service