Self-Service BI – what does it mean for different users?
Many BI departments are in a dilemma where the pressure for Self-Service BI is big and the resources to support it are limited.
28. June 2018
Self-service (self-service) as a concept has in recent years been a major focus area for many BI departments. There is a great desire to make users self-sufficient and thus give them faster and more agile access to exactly the analyzes and data visualizations that they demand.
However, it is not always easy to make a successful user adoption with that strategy; the people you already know are super good and interested in working with data, they would probably welcome the opportunity for self-service, but many organizations struggle to get everyone to join that approach.
What is Self-Service BI? It depends on who you ask…
It is Kapacity's experience that the request for self-service often is about the wish to have reports where they can change a filter with a period, product or e.g. seller - whereas for others it is a matter of full freedom to work with arbitrary data to find patterns and connections that the BI department has not already thought of.
An offer of Self-Service to the users in the organization must therefore be based on the data maturity and analysis tool the users have and from there the BI department can select the options that will be made available within self-service.
Below, we will review the different definitions of Self-Service.
Dynamic filters in reports
A dynamic version of classic PDF reports
Suitable for repetitive reporting with a need to filter on e.g. periods, categories or account
Filters and drills in dynamic dashboards
A more visual presentation of data with built-in history and given options for filtering and navigating to details via graphics
Suitable for users who prefer graphic forms of presentation
Build dashboards from predefined components
Ad hoc modeling of dashboards with key figures and dimensions from the predefined cube
Suitable for prototyping
Build analyzes in Excel based on predefined components
Excel as front-end for a predefined cube. Same user experience as PowerPivot but with server performance
Suitable for users who prefer Excel
Model data based on data files or databases
Modeling data in Excel/PowerPivot/R from one or more data files
Suitable for prototyping and sandbox analysis. Limited on performance and traceability. Not suitable for recycling
Search in data in a data lake to find connections and patterns
Data discovery in unknown data using Advanced Analytics techniques/tools to find unknown connections
Suitable for prototyping
The above reviews the continuum that the Self-Service definition can extend over: At the top is the very simple definition where the user instead of a PDF sent by e-mail has access to a dynamic setup, where one e.g. can change periods and at the bottom is the ability to do data discovery in a data lake using Advanced Analytics/Machine Learning tools.
Fixed dashboards, which the BI department has made for the users are also a kind of Self-Service within a set framework. Here, the data history is told visually but it is still possible for the users to get the filters and cuts in data they need.
Self-Service BI is also for the advanced users
For the more advanced users, self-service can take the form of the opportunity to build new dashboards or analyzes in Excel. In both cases based on data that has already been prepared by the BI department in e.g. cubes so you do not have to handle the modeling and business logic yourself.
Finally, some very advanced users may need to work with new data that has not yet been modeled - either to test whether there are analytical benefits in new data sources before deciding to include them in the organization's Data Warehouse or as dissemination of wishes in the form of a prototype for later development. This can be done with Excel as the engine or e.g. in the form of a desktop version of tools like Microsoft Power BI.
The different approaches to Self-Service have different pros and cons. Some of them appear from the above, and they place different demands on the competencies of the users and also present different challenges in relation to governance, which is often a natural concern when it comes to self-service. The BI departments must therefore make the choice based on an overall assessment of the users' maturity and wishes as well as their resources to support the users with education and support, which is essential for successful adoption.
Learn how we can help you with Business Intelligence.
Other blog posts you might find interesting
BI teamwork: How do you get a good collaboration between business and BI/IT?
A known problem regarding BI projects is the lack of cooperation between the business and the BI team/IT. Read our recipe for good BI teamwork here. [...]
IFRS 16 – Are you ready for system support?
From January 2019, major changes will take place for companies that submit accounts in accordance with IFRS. IFRS 16 arrives, and this has great significance for the company's key figures. [...]
From data to BI solution – does the BI team deliver what users dream of?
The dialogue between the data users and the BI department can be difficult at times, and it is not always that the users feel that they have the BI solution they need. Here we give you the recipe [...]