Self-service BI is a trend with a somewhat vague definition. In the most general sense, self-service BI tasks are those that business users carry out themselves instead of passing them on to IT for fulfillment.
The aim is to give the users of BI tools more freedom and responsibility at the same time. At its heart lies the notion of user independence and self-sufficiency when it comes to the use of corporate information, which leads to a decentralization of BI in the organization.
Beyond this rather broad definition, self-service BI has many facets. Its meaning depends very much on the specific requirements of particular user roles. For each role, self-service BI can help users accomplish various tasks.
For example, casual BI users often only need to be able to filter and group data. In the very same environment, power users or business analysts might have to integrate local data from different sources on their own so they can quickly build or enhance existing reports. Therefore, the need for self-service within a BI environment varies according to user requirements.
Typical users can be assigned to three roles which, in reality, overlap and change according to the task at hand:
Casual or standard users make up roughly 70 percent of all BI users. Usually, they have a rather limited BI skillset which corresponds with their straightforward requirements. Therefore, analysis, dynamic reports and dashboards are sufficient to cover their self-service BI needs in most cases.
Power users make up around 25 percent of all users (typically much less in larger BI environments). They are skilled BI users who need a lot of flexibility and functionality for their daily work with data to answer their business problems. Suitable self-service tools allow them not only to analyze data but to change existing (or even create new) reports and dashboards from scratch.
About 1 to 5 percent of all BI users can be described as business analysts. These are the users with the most advanced BI skills and requirements. They have the highest demand for flexibility and functionality in their self-service BI solutions. For them, self-service must cover tasks like data exploration, modeling and deploying a sandbox environment for special use cases.
As described above, the requirements of business users for self-service are extremely diverse and range from free analysis and modification of reports to integration of local data and even to changing semantic models. These requirements can be categorized as follows:
Modification of Reports and Dashboards
The use of self-service business intelligence tools enables business users to modify reports and dashboards. Users can filter or produce reports visualizing their key indicators in the most meaningful way. They can independently create analyses specially addressing their particular needs and therefore derive new insight from relevant business processes. Furthermore, reports can easily be adapted. In this way, self-service functions provide users with a higher level of flexibility in the creation of analysis and reports.
Creation of Ad Hoc Reports and Dashboards
The creation of reports and dashboards no longer needs to be the sole task of IT. Thanks to intuitive tools and predefined report templates and dashboards objects, power users can create ad hoc reports and dashboards to support other end users (usually within their line of business) themselves.
Integration of Private, Local Data
A further requirement often associated with the self-service philosophy is the integration of private, local data into existing reports, analyses or data models. Such data can come from Excel documents, flat files or other external sources.
Self-service functions help business users to quickly integrate data into reports. Local data can be used to extend the information delivered by the data warehouse limiting the pressure and workload on data management.
Modification or Creation of Data Models
For power users or business analysts in certain environments, self-service must provide the possibility to modify or produce data models independently. Business users act as ‘data modelers’, adapting their semantic model to a business department’s needs without relying on IT or BICC involvement. Modeling can take place in a metadata layer, a database or a so-called ‘sandbox’ (a confined environment). Each company should define its own data management strategy to determine the best approach to be used.