How to make Self-Service BI successful in your organization?

The BI industry has been evolving over the last five years to place more emphasis on self-service, end-user-driven analytical reporting. In response to market needs, many vendors have been broadening their portfolios to offer a wide range of tools that promise to deliver advanced reporting capabilities quickly and without much effort for business users. There are many flavors of Self-Service BI of course; in the strict sense it stands for fully IT independent BI capabilities for business users.

However, in real life, self-service BI users often run into data challenges. These challenges, when not addressed properly, can cause users and business managers to lose faith in their ability to take advantage of Self-Service BI. Understanding why these challenges emerge and how to solve or prevent them is critical when planning a BI strategy that integrates full self-service BI capabilities for business users.

To that extent, let’s divide business users into three broad groups:

• Report Consumers: those whose main interest is in viewing pre-developed reports, dashboards etc…. without much effort required of themselves to get the insights needed.
• Power Users: Those whose main interest is slicing/dicing/interacting with data with the explicit goal of creating compelling analytical views.
• Data Modelers: Those whose main interest is to organize data sets and enrich existing data sets to reduce the amount of redundant work required to create reports and to explore new opportunities by combining data which is not always generated in the organization itself.

The first group, Report Consumers, is a common one among business users in each organization. This group is not interested in creating ‘own’ reports or dashboards. They just want their insights as fast and pre-defined as possible. Some user flexibility is often granted by the usage of parameter selections in reports to get different angles for looking at the data but the ‘drill path’ is pre-defined by the author of the report. This type of users is best served with the traditional BI set-up were the reports are being created by the IT staff or BI unit and are certainly not target users for Self-Service BI.

The second group, Power Users, are interested in creating their own reports. It is important for them to create compelling ways of representing data, for example - animated scatter charts in Power View or 3D map visualizations in Power Map. This is understandable given their reports are likely to attain high visibility among executives, and compelling interfaces are desired or even required. In some cases, Self-Service BI users are the executives themselves.

To the majority of users in the Power User category, Self Service BI is all about pivoting, exploring, and visualizing data. Any issues that may arise from joining multiple data sources, handling data quality issues, and producing complex business calculations in order to prepare the dataset may not be something they want to or have time to address directly.

On the other hand, the business of analyzing data has evolved to the point that now users can also model their datasets independently. Although not something that many users want to engage in, it is difficult and sometimes impossible to produce compelling and accurate business reporting without a properly designed data model to support the effort.

A data model can be defined as the physical representation of business concepts, typically in the form of tables of data, the relationships among those tables, as well as business logic to augment analysis such as calculations, hierarchies, and so on. The most successful and far reaching Self-Service BI implementations commonly have well-defined data models as a supporting infrastructure. In the Microsoft world, a data model is a requirement for implementing Self-Service BI and can be created by business users using Power Pivot for Excel, or by IT/BI technical staff using the SQL Server Analysis Services OLAP technology.

Developing well-structured Self-Service BI models is critical to the process of ensuring calculation accuracy and re-usability, as well as categorization of data structures in business terms that users can understand. Without a sound underlying data model, the analyst’s efforts with reporting and visualization can produce undesired consequences in the form of multiple definitions for a single business metric and/or redundant efforts to prepare data for business reporting.

As important as it is, modeling data is a skill only a small percent of business users are willing to learn and use. A third group of users (the data modelers) are more technically skilled and are likely to have a good deal of business expertise. Not surprisingly, they are an attractive asset to any organization. Their modeling activities can be tedious, and require an understanding of applied database fundamentals that other business users do not normally have or even want to have.

Many organizations are not willing, or able, to staff their business units with at least one data modeler. As a consequence, the ability to develop a robust vision around generating business insight can be severely compromised. Given that these advanced users with data modeling skills tend to be a scarce asset, knowing from the start what your options are could be of help.

Sometimes the traditional BI set-up with a Data Warehouse and BI front-end solution maintained and nurtured by IT might be the best option depending on the skills of the business users.

The most successful BI set-ups we see in our customer base are those where we can combine the traditional BI set-up with a full Self-Service capability for advanced Power Users, this set-up will unleash the full potential of BI in your organization.  

The following flowchart can be of assistance when determining where to focus your efforts when designing a Self-Service BI strategy that can achieve high relevance for the business:

Understanding stakeholder expectations and your team’s current capabilities is key in order to implement the flavor of Self-Service BI that is appropriate for your organization.