Why you can't just do self-service BI

Comment by Christian Schneider, QuinScape Why you can't just do self-service BI

Author/editor: Christian Schneider/Nico Litzel

The benefit of self-service BI is undisputed. Nevertheless, its relevance in the BARC trend monitor fell from 2nd place in 2017 to 5th place in 2020. In view of the self-service BI projects that have failed at times and the hidden hurdles to success, this trend is not surprising.

Companies on the topic

Quinscape GmbHQUNIS GmbH

However, we must not see this as a failure, because the topic "Data Driven Culture", the transformation to a data-driven company, makes some disciplines from self-service BI all the more important. It is therefore worth taking a look at common mistakes, because we will encounter them again and again on the way to a data-driven culture:

Why you can't do self-service BI with every architecture

The smoother access to the required data is, the more likely it is that the accompanying training measures will arouse enthusiasm for using the advantages offered. At its core are two requirements that read very similarly, but are clearly different in nature:

These examples make it clear that self-service BI, like any data integration project, needs to be planned. It has to be part of your data strategy.

In the following we would like to present an excerpt from our checklist. This can be used to question where a company stands:

Why not just self-service BI

Trusted data - data warehouse (DWH)

The source of trustworthy data is still a DWH. For a DWH, all mechanisms such as historization or quality assurance have already been thought through and solved. In our current DWH modernization projects, we are experiencing a trend towards agility through the modeling of data vaults, which are the foundation for another important field of action in modern data architectures:

Doing everything at the same time is a mammoth project and you will fail. But all fragments of a good data strategy can be iteratively incorporated into day-to-day operations and shaped into best-in-class data management if you approach them in the right order.

This order is always individual and it takes a lot of experience, as there are also quite a few wrong orders and what was intended as a self-reinforcing mechanism.

Why you can't do self-service BI without a strategy.

In a study by Forrester and KPMG, only 35 percent of respondents said they trust their own company's analytics. The phenomenon of the "trust gap", which is also known in many other domains in addition to analytics, is one of the biggest obstacles to initiative, creativity and the acquisition of knowledge: If the correctness of the basics cannot be trusted, why should anyone invest time in working with it.

Trust is primarily created through commitment and transparency, i.e. primarily through how a data strategy is lived in a company. The cultural handling of data - similar to different management styles - is an important factor on the much-discussed path to the democratization of data. A data strategy must be actively managed, accounted for and lived by management.

Frequently, the following topics are not properly addressed in companies:

Anyone who tries to completely save on analytics and IT experts through a self-service approach has already failed at this point.

Self-service must be part of an overall data culture that must be established in the processes involved in handling data. The least you can do here is without experts who represent exactly the anchor of this trust.

Conclusion

The introduction of self-service BI can only succeed if professional end users, technical experts and a committed management team create a culture that has the necessary trust in and understanding of the data create. Ultimately, experience in introducing data-centric processes at all levels determines the success of self-service BI.

From our project experience, the technical challenges can be mastered very well with suitable architectures and platforms, but the composition of the technical possibilities with the clean orchestration of the necessary processes requires much more experience in order to be successful.

(ID:47515090)