Software as a service is already widely recognized as an accepatable delivery mechanism and implemented at many organizations. It is a way of delivering software which is widely accepted. On the contrary, business intelligence (BI) as a service seems a distant goal, and companies are reluctant to take this path. 

The business case for implementing BI as a service should be simple:

• Decrease costs
• Quicker time to market
• Service reliability
• Data security

There are several reasons why BI as a service is not widely adopted:

• Strategic information for a company
• Security issues
• Only feasible for smaller companies

Maybe this was the case for early adopters of BI as a service. We can take Salesforce.com as an example; they store and provide strategic information (customer) in a secured way. I see other reasons why BI as a service isn’t as successful as SaaS:

• Status of the business intelligence environment
• Cultural issues
• Lack of trust between business and IT
• Traditional vendors are in a constant merger scenario

The current economic situation could have a positive impact on the market share of BI as a service. Cultural attitudes will disappear when people realize there is a legitimate cheaper, stable alternative. And, if some analysts are right, the growing adoption of open source BI will result in an increase of BI as a service.

One of the main reasons for slow adoption, certainly for our local (Belgian) market, is the status of our business intelligence environment. We are not yet into BI as a service, open source, unstructured data … We are still struggling in delivering business intelligence projects on time and within budget. The foundation or the status of our data warehouse is not stable. We don’t have one version of the truth, no go-live process, no trust between business and IT. To take the path of BI as a service, we certainly need a stable data warehouse environment.

I have learned from working on global projects that successful business intelligence environments are those where business and IT work closely together in a stable environment. No uncertainty on project timeframes, deliverables or the correctness of our data.

How do We Get to a Stable Business Intelligence Environment?

Working with several service providers or consultancy companies, it is funny to see that each of them has his own business intelligence framework or methodology. Most organizations don’t insist on their own methodology and are adapting these different frameworks. Although these service providers and consultancy companies have their framework, they start each project from scratch. Project templates, naming conventions, roles and responsibilities … are reinvented.

I believe that BI as a service can bring us back to the basics. Why don’t we apply service management rules to our BI environment? IT service management (ITSM) is a process-based practice intended to align the delivery of information technology (IT) services with the needs of the enterprise, emphasizing benefits to customers.

The keywords lie in “process-based practice.” We need to study our way of working in the business intelligence environment and turn the different activities into processes that can be measured. If we can measure, we can improve, define pain points and, last but not least, make clear agreements with the business.

We can apply ITSM in a pragmatic manner into our business intelligence processes. The processes we need are in place and documented. It is one way of working for our internal organization but also for service providers or consultancy companies. They need to adapt our existing framework.

One of the first advantages of measuring our BI processes is to get rid of misperception in our deliverables. For example: queries that run less then a second in 99% of the cases, but run more then 10 seconds in 1% of the cases will be measured as bad performance. On the contrary queries that always run for 15 minutes will be seen as good performance (because the user is used to this performance).

Another important advantage is the relationship with different internal or external suppliers. Our processes and measurement systems are already in place and will be the basis for negotiations on deliverables.

Although it seems to be a simple approach, it’s not that easy. One of the biggest risks is getting stuck in the processes and forgetting the results. Therefore, we can use simple BI principles: think big, start small. First, we need to identify the pain points of our business intelligence environment. This should happen in cooperation with the business as one of our goals is to bridge the gap between business and IT. Once we have the pain points identified we can start prioritize and implement step by step.

Next month, I will discuss the different processes in business intelligence from a service management perspective. In the meantime, I welcome your opinion on BI as a service.

SOURCE: Introduction to BI as a Service

Recent articles by Ivan Schotsmans

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