Although it is far from being a recent concept, Self-Service BI continues to be a difficult challenge to overcome in many companies, either because of the technical or functional difficulties it poses, or because of the organization's own inability to adapt and the scalability of this business intelligence solution.
How can an organization maximize the use of its models and get the most value out of the investment made in building a Data Warehouse / Data Lake and all surrounding systems? Here are some tips...
Technical and functional training on the models and tools
One of the main reasons for Self-Service resistance in a company is obviously the lack of training, not only technical, but also functional.
An efficient training strategy in the area should not only contemplate all the technical features of the selected tools, but also, and most importantly, the functional specifications of the provided models, demonstrating what are the main concepts, relations and goals of each model and providing sample analysis so that users can acquire a higher familiarity with the models.
At this stage, the company may decide to move on with the creation of a centralized Data Dictionary, which allows any user to check the functional specifications and understand which concepts, calculations and goals exist in which model and how should they be used.
Another success strategy at this stage can be the creation of analytic start-up kits with pre-compiled information and analysis that might be used as a starting point for the casual users, eliminating the fear of the blank-page.
Give users the ability to navigate and develop
One of the most hidden Self-Service BI traps is precisely the restriction on how Self-Service can Self-Service BI be.
At this stage, it is important to understand that Self-Service BI is much more than interactive analysis and should have a strong focus on free model navigation, as long as it obeys all security and confidentiality requirements.
The organization role at this stage should not be focused on restricting users to a set of interactive analysis and dashboards, but on the promotion of free model navigation, data visualization and representation with technical and functional support.
Each user should have not only all the information he / she can access, but also a range of tools and techniques to analyse it, from a simple “slice & dice” to advanced visuals and / or predictive models included in the tools.
Insight sharing promotion in an intra or inter departmental basis
The next step on making the most of the Self-Service capabilities will be the promotion of insight sharing in the organization.
It is not unusual that Self-Service BI ends up being nothing more than a set of several isolated knowledge-islands, where each has a similar set of analysis and tasks, which never get to be known in the other islands.
At this stage, the organization can move on with the creation of a Self-Service Analytics portal, where all the users / departments can navigate, view and share their insights, both on a departmental and organizational level.
This kind of sharing can help disseminate ideas, analysis and techniques to tackle organizational problems and will promote the evolution of the Analytics area as a whole. It might even work as a showroom where the organization gets aware of all the work done in the several areas.
Besides, this information sharing will increase the global awareness of the existing models and analysis, making the Analytics experience an interactive component between all departments.
Improve the existing information with customized data
For technical-prone users, the simple navigation and visualization might not be enough and it becomes necessary, as a supplement to its function, to complement the existing corporate Data Warehouse / Data Lake information with external or customized data.
When not addressed by the organization, this kind of situations frequently flow into the creation of a shadow-BI layer where entire Data Marts are created by sophisticated schemas of data enhancement and custom data improvements.
The company can address this problem in two ways:
• First, by creating a transversal master-data layer in the whole organization whose responsibility will fall upon the main key-users in the several business areas. This layer will keep the crucial information updated and will allow the centralization of all forms of data qualification and categorization.
• In second place, by allowing the users to access the corporate model metadata, so that they can connect the model directly with the custom data they need.
This kind of evolution will allow the company, in a first stage, to enhance the existing information and answer the most urgent needs to qualify, categorize and add information. Then, in a second stage, to enhance the integration layer with the keys-users input by promoting their customizations to the corporate Data Warehouse / Data Lake and making the key-users as part of the company’s BI strategy instead of simple internal customers.
Governance and platform control
Finally, as consequence of all the previous evolution and growth, the company should feel the necessity of a governance strategy for the Self-Service platform.
With the growth of the Self-Service BI layer in the company, it is inevitable that there is a need to coordinate data sources, review contents, audit performance as well as document and audit all the Self-Service processes, especially those that are candidates to corporate BI promotion.
At this stage, there is a frequent need to control all the data flows in the existent Self-Service processes with source-to-destination mappings in order to realize what are the main sources and the main analysis in the processes. This might be an indicator that helps the company devise its BI strategy.
On the other hand, the master data layer might slowly extend beyond a data-centered vision and help the organization to create master data on its own contents, allowing users to document their analysis by explaining the need, technique, source data, etc.
On a technical perspective, there might be a need to audit certain user contents, especially when it comes to performance and data management situations or critical information.
A company should take Self-Service BI as a complementary layer of its corporate Data Warehouse / Data Lake, in which it will provide its users, especially the power users, exploration, navigation and representation capabilities that did not exist until then.
This layer should never replace the existing corporate Data Warehouse / Data Lake, but as a mantle to provide the greatest possible flexibility to the users and help the organization growth in the Analytics area.
As so, all the growth of the Self-Service BI area should be seen as a responsibility of the whole organization not only on technical and functional perspectives, but also as the management of the whole evolution of the BI platform.