Self Service BI or Shadow IT?

Companies are increasingly recognising the inevitability of Self Service BI (SSBI), but not infrequently they fail to implement it and therefore cannot fully exploit its potential.

With this contribution we want to answer the following question: Which factors must be fulfilled so that the cooperation between the central BI organisation and users in the specialist areas - the so-called "power users" - works and both sides benefit from each other? Or formulated differently: When are the activities of the power users to be regarded as positive SSBI and when are they to be regarded as negative shadow IT?

In this article, we would like to highlight four aspects for a successful implementation of SSBI: In addition to a checklist of so-called hygiene factors, we would like to focus on the human factor and its personal mindset, as well as cultural aspects and finally technology. For it is the holistic view of these different areas and their interaction that forms the breeding ground for SSBI to work. A plea for more culture in the context of technology.

Definition of terms

"Business Intelligence" (BI) includes the collection, preparation, analysis and presentation of data in electronic form and thus the process of extracting information from data for decision support in companies. Self Service BI" (SSBI) is when this process is wholly or partly the responsibility of non-technical users in the business units and they are not dependent on the help of technical users in the central BI organisation. SSBI is not limited to the independent creation of analyses and reports, but also includes the necessary enrichment, collection and preparation of data.

"Shadow IT" describes information technology processes and systems that are located in non-technical departments and the central IT organisation has no knowledge of them or does not approve of them. This usually leads to major risks in terms of security and compliance.

The terms SSBI and shadow IT are therefore like two sides of the same coin: SSBI is the positive term for the greater autonomy in specialist departments, which has become indispensable due to the ever-increasing importance of data and data-driven decisions. Shadow IT is a negative term that focuses on the risks of this greater autonomy.

So is it more like SSBI or shadow IT when a non-technical user collects, prepares, analyses and presents data? Ultimately, both sides of the coin must be considered. However, the concrete answer to this question depends on whether the person asked has a positive or negative view of the respective use case - often even very subjective. The following examples are intended to illustrate this.

Negative examples: How does shadow IT expose itself?

In our BI projects, we often experience situations that are a breeding ground for shadow IT:

• A BI expert who, when a power user explains the current solution created with Excel, inwardly shuts down and thinks to himself: "What nonsense?! We'll do it differently and better when we implement it anyway...".
• A power user who stubbornly insists that his server-based solution on a decommissioned computer under his desk can never be mapped with another technology (better suited to the central BI strategy) anyway and fights against centralisation.
• A BI expert who is annoyed by constantly changing professional requirements, therefore avoids every conversation, only gets to work after a 15-page specification from the department and tries to frantically maintain the status quo.
• A power user who is reluctant to share his professional knowledge, wants to preserve his head monopoly, cannot get involved in new processes, technologies and responsibilities and sabotages them with artificial "show stoppers".

Of course, these are exaggerated examples. However, anyone who has worked in the field of data and analysis - whether as a BI expert or as a power user - will be familiar with similar situations. What is remarkable is that the breeding ground for shadow IT reveals itself directly in interpersonal interactions and the human factor plays a central role in this.

The following section discusses the success factors that need to be taken into account in order to pave the way for a successful implementation of SSBI.

Hygiene factors for successful SSBI

Of course, the so-called hygiene factors are relevant first, which have proven themselves in practice, can be objectively assessed and are unavoidable from our point of view:

• A clearly defined data, BI and data governance strategy
  • SSBI is taken into account organisationally and technologically
  • and which is formulated and well communicated in such a way that it can also be applied in the departments.
• A "sufficient" level of knowledge of the central BI organisation with regard to the SSBI activities of power users in specialist areas.
  • which makes it possible to offer assistance to power users,
  • to exert a corrective influence on the BI strategy in case of doubt,
  • and not to hinder, but to see themselves as enablers.
• Central data governance requirements (e.g. sensitivity of data, DSGVO, access security, marking of data with confidentiality labels, encryption, etc.).
  • which are formulated in such a comprehensible way and communicated well that they can also be applied in the specialist areas.
• Use of uniform software for SSBI that fits the company's BI strategy
  • in order to avoid a company-wide "tool zoo" which, in the worst case, ranges from local, file-based applications in Excel to independently operated servers in the specialist departments.
• Sufficient, formally defined, organisational structures and communication platforms
  • which promote and demand an exchange among power users and with the central BI organisation - e.g. BI Competence Centre, knowledge management, training courses/workshops/fireplace evenings, joint projects, etc.

Human factors for successful SSBI

All hygiene factors must be observed, but we are convinced that the true breeding ground for an optimal design of modern SSBI lies in a certain mentality or attitude of those involved - both the power users in specialist areas and their colleagues from the central BI organisation.

The hygiene factors mentioned are almost worthless if the human factors do not fit. If there is no positive attitude or mentality for the common handling of data and SSBI, the acting actors lack the right SSBI mindset, so to speak. This is the part of the iceberg that lies beneath the surface. Successful or effective and efficient SSBI emerges and thrives from the fact that departments and the central BI organisation maintain regular and cooperative communication at eye level, mutual respect for each other and a balanced relationship of "give and take". Shadow IT always occurs when communication is reduced to a minimum or even stopped completely and cooperation is rigid, extremely formal and characterised by personal animosities.

In our opinion, a good SSBI mindset includes the following features:

• Helpfulness. Helping, empowering and moving someone else forward is at least as important in today's business world as pursuing and achieving one's individual goals.
• Trust and respect. Valuable communication is based on meeting at eye level, taking each other seriously and listening to each other properly. Empathy for the wishes, concerns and individual goals of one's colleagues is the basis for an optimal relationship of "give and take".
• Willingness and readiness to learn. You should have the desire to get to grips with new topics, technologies, approaches, solutions, etc. This is the only way to create innovation. This is the only way to create innovation. After completing a successful BI project, at best a subject matter expert understands the basics of dimensional models and a BI expert understands the basics of the mapped subject matter logics.
• Willingness to change. Fast-moving and highly dynamic also characterise BI systems. Changes are always associated with risks and often initially with rejection. Individual willingness to change is characterised by an openness towards new things and a focus on opportunities when evaluating a change.
• Personal responsibility and self-organisation. It should be easy for those involved to take on new areas of responsibility and, on the other hand, to hand over previous areas of responsibility, e.g. when power users in specialist areas take work off the hands of the central BI organisation or, conversely, when a power user's SSBI development is handed over to the hands of the central BI organisation.

However, these personal characteristics of employees cannot be directly influenced and are difficult to train. In order to provide a secure framework for promoting cooperative, appreciative, adaptive and change-ready behaviour among employees, there is a fundamental starting point: culture!

Cultural factors for successful SSBI

A good corporate culture is one of the most important foundations for the overall success of complex organisations or companies, especially in a time characterised by a shortage of skilled workers and VUCA[1] -world, but proves to be essential especially in the SSBI context.

In order to positively influence the culture of the company in general and the culture in dealing with data and SSBI in particular, there are concrete fields of action:

• Modern forms of organisation. We are convinced that the classic organisational division into "specialist department" and "IT" can no longer do justice to the ever faster speed of innovation cycles and technologies. In future, it should give way to a division into the primarily value-creating and market-oriented business units in the periphery vs. the meaningful core with central, service-oriented units for the periphery according to the principle of cell structure design ("peach model") [PfH14, PfH15, PfH20]. The peach model (cf.   Figure 1) is for us the counterpart to the rigid hierarchical thinking of the pyramid. The primarily value-creating and market-oriented business units should be organised cross-functionally based on customer needs and act as networked cells with decentralised responsibility in the sense of self-organised teams. The former hierarchies, on the other hand, become service providers, organisational architects and value guardians who support, empower and provide orientation to the business units from within, i.e. from the core of the peach. Here in the meaningful core, the central BI organisation should also be located and anchored as the "data DNA" of the company and primarily act in the role of enabler rather than implementation partner.

 

• Leadership Agility. In order to do justice to this modern organisational structure, we believe that the understanding of leadership must also change fundamentally. The role of managers and their personal behaviour is immensely important for the development of the desired SSBI mindset. While managers used to see themselves more as "architects of IT systems", they will increasingly transition to the role of "architects of social systems". This is because decentrally networked, largely self-organised teams not only need a new set of competencies and methods, but also a meaningful framework and an attitude that are diametrically opposed to classic leadership and control thinking in many places [Küs14]. With his Leadership Agility Model, Bill Joiner provides a suitable framework for how the leadership role can be developed from the tactical, problem-solving-oriented "expert" to the strategic, goal-oriented "archiever" to the visionary, development-oriented "catalyst" [Joi06]. We are convinced that the breeding ground for SSBI mindsets is strongly influenced by the agility levels of the shaping leader.
• Trust and error culture. Courage to change and innovate is strengthened when employees feel trust towards the company, their managers and colleagues. In particular, this also includes a good culture of mistakes, in which failures are allowed (or even expected) as long as one learns from them [Rot21].
• Change Management. With changes in organisation, responsibilities, technologies, etc., it is always important to focus on the people affected with the help of professional change management and to ultimately pick them up or bring them along, because successful change requires both the technical as well as the human page. [Pro20]

Technological factors for successful SSBI

In addition to the aspects discussed above, it should not be neglected that the successful implementation of SSBI is not least made possible by the use of suitable technologies.

Our focus here is on the Microsoft Data Platform and, in the following, on Power BI in particular. In the context of SSBI, we see two outstanding functionalities in the current developments to optimally support the interaction of Power Users and the central BI organisation:

1. A unified platform for power users and central BI organisation

Power BI makes it possible to work with different tools on one platform: On the one hand, Power BI Desktop can be used simply and intuitively to create data models, and on the other hand, these models can then be further processed via XMLA endpoints [Mic23-1] with significantly more powerful, but also more complex tools such as Tabular Editor and, in the future, Visual Studio Code [ThM23]. The underlying technologies, modelling considerations and measure definitions (DAX) are identical in all cases. Thus, power users and the central BI organisation speak a common language. In addition, the transition in the life cycle of a BI application developed by a power user but then migrating to the central BI organisation (or vice versa) is simplified.

2. Deep extensibility of existing data models

A particularly outstanding SSBI functionality of Power BI is Composite Models [Mic23-2]. These enable the extension of complex, centrally provided analysis models. Here, the central models can be enriched with all their predefined logics (relationships, measure definitions, etc.) and on the basis of the detailed data granularity they contain - e.g. with an individual grouping of customers and products or with plan, forecast and benchmark values from any source.

In our view, this functionality is a real game changer for SSBI [Nie23] and other professional colleagues also see it as a milestone [Rus20]. It offers highly interesting architectural and organisational possibilities: Until now, centrally provided analysis models could normally only be used in a read-only manner in specialist departments and in personal SSBI for analyses and reports. The combinations with own data sources required high manual effort or limited the user to simple calculations only. Composite models, on the other hand, now enable simple individual data enrichment of the central models with all interfaces supported by Power BI. Redundancies with regard to data storage and model definition as well as the export of data are thus mostly no longer necessary.

Changes and extensions to the central models are automatically integrated into the composite models because they are only linked but not replicated. Composite models enable nesting, so that data democratisation is promoted by allowing individuals to carry out their own evaluations on datasets provided centrally or from specialist departments (cf. Figure 2). This can be used to support a datamesh concept in Power BI.

Conclusion and implications for the successful implementation of SSBI

As shown in the last section, composite models in Power BI thus pursue the same goal in a technological way as the cultural aspects described earlier: Above all, it is about bringing all people working with data closer together. Only the interaction of what is technologically feasible with a fertile breeding ground of cultural and human-enabling aspects forms the necessary prerequisite for SSBI to work. We are convinced that this offers the opportunity to dissolve historical divides and create value together in order to derive the greatest possible benefit from data.

 

 

Literature

[Mic23-1] Microsoft: Dataset connectivity with the XMLA endpoint, 6.2.2023, https://learn.microsoft.com/en-us/power-bi/enterprise/service-premium-connect-tools, abgerufen am 25.3.2023 

[Mic23-2] Microsoft: Using DirectQuery for Power BI datasets and Analysis Services (preview), 7.3.2023, https://learn.microsoft.com/en-us/power-bi/connect-data/desktop-directquery-datasets-azure-analysis-services, abgerufen am 25.3.2023 

[Nie23] Nielinger M.: Wie Unternehmen Composite Models in Power BI als Game-Changer für Self Service Business Intelligence einsetzen können, 27.3.2023, https://www.noventum.de/de/nc360-artikel/wie-unternehmen-composite-models-in-power-bi-als-game-changer-fuer-self-service-business-intelligence-einsetzen-koennen.html, abgerufen am 27.3.2023 

[PfH15] Pfläging, N.: Komplexithoden: Clevere Wege zur (Wieder)Belebung von Unternehmen und Arbeit in Komplexität, Redline Verlag 2015 

[PfH14] Pfläging, N.: Organisation für Komplexität: Wie Arbeit wieder lebendig wird – und Höchstleistung entsteht, Redline Verlag 2014 

[PfH20] Pfläging, N./Herman S.: Zellstrukturdesign: Eine neue Sozialtechnologie, die unternehmerischer Wertschöpfung Flügel verleiht, Vahlen 2020 

[Joi06] Joiner, B., Leadership Agility: five levels of mastery for anticipating and initiating change, Jossey-Bass 2006 

[Küs14] Küster, H., Leadership Agility – die Führungsherausforderung in der IT. In: Lang, M.: CIO 3.0: Die neue Rolle des IT Managers. Symposium Publishing 2014 

[Rot21] Rotermund, U.: Ausbruch aus der Komplexitätsfalle: Leitfaden zum selbstorganisierten Zusammenarbeitsmodell für Manager und Macher, Springer Gabler, 2021 

[Pro20] Karen, B. et al.: Best Practices in Change Management 11th Edition: Prosci Benchmarking Report 1863 Change Leaders Share Lessons and Best Practices in Change Management, Prosci Research 

[ThM23] Thierbach M., Münster G: The Future of Version Control for Tabular Models, 17.3.2023 https://youtu.be/XbDDyTvclVs , abgerufen am 25.3.2023