Companies have invested a lot of time and effort in building Business Intelligence systems and corporate data warehouses. Typically, a specific IT team is in charge and meticulously profiles and analyzes the data. They clean, transform and store it according to the requirements and build reports and dashboards to visualize the necessary information in a clear way.
Those teams typically have a backlog of changes to be executed. Also, doing things the right way takes time – it can take days, weeks (or even months) before the data you wanted gets added the system or the report you needed gets built.
Today, data is everywhere. It exists in different forms and sizes, from ERP and LOB transactions stored on premise to sensor, social and public data living in the cloud. The volume and speed at which data is being generated is astounding and for a lot of analyses, data from those disparate sources need to be combined. At this point, the corporate BI team just cannot cope anymore.
To cope with this trend and empower people to make their own analyses, you can introduce a self-service BI solution. Both major vendors like Microsoft (Power BI), SAP (Lumira) and SAS (Visual Analytics) and niche players like QlikView or Tableau offer specific solutions specifically built for this purpose.
Those solutions make it intuitive and easy for end users to create reports and visualizations based on a wide range of data sources inside and outside of the organization. They allow to combine data from those different sources and to add simple transformations and calculations. You can publish and share your results with colleagues that can use either a browser or a mobile device to consume the content.
There are some trade-offs though – to make it easy to use, part of the complexity of such a tool has been removed or hidden. You can create really stunning and interactive visualizations and configure them up to a certain level, but you do not have the same pixel perfect control that programmed/enterprise reporting solutions offer. In a similar way, some really complex data transformations might be out of reach.
Finally, different people will use different definitions of similar concepts, use raw data without cleaning it and produce similar reports with different numbers. Concepts like master data, data quality and ‘single version of the truth’ seem almost contradictory to the very nature of self-service. If not properly addressed, the risk of creating a big pile of reports nobody trusts or uses is lurking around the corner.
Self-service solutions are a great evolution and can really help to democratize access to data, you just have to deploy them in the correct way. You do need a clear plan around governance, prepare your IT team to handle a different kind of requests and really invest in training and supporting people to use these tools in an optimal way. As with all IT projects, the success does not only depend on picking the correct tool, but also on the way it is implemented and used.
Interested to learn more? We are glad to announce ‘Self-service or no service’, our Delaware Consulting D’Lunch, scheduled on May 28th. During this lunch, we will discuss best practices on making Self Service BI a success story, not a fairy tale. Subscribe now!