Last month, I discussed the challenges and opportunities of self-service BI as an approach to enable non-tech-savvy business users to directly access data and explore it on their own. In this blog, the focus turns to some of the reasons why self-service has failed and understanding the approaches that have worked.
For reference, I use the terms ‘power users’ and ‘citizen developers’ interchangeably in this piece, to represent those business users outside the BI team. These are the people who will be enabled through self-service BI to address their own and others’ informational needs.
Self-service BI failures
There are several possible reasons why self-service BI fails. Some that have been identified are:
- Unrealistic expectations: Companies who let novice users loose on organisational data face the potential of bad quality reports and inconsistent reporting. This results in a huge distrust of data in general.
- Reporting chaos: With no governance structures in place, there will be redundant reports from different users. Because they work in silos and use different filters and terminology, they will deliver conflicting results despite using the same underlying data.
- Lack of adoption: BI tools, environments, and processes may be easy for specialists. However, we must keep in mind that casual users do not have the same background and skills. The complexity and ‘newness’ of these environments can be quite intimidating for novices.
- Lack of support: Citizen developers are not trained on BI processes and tools. Without proper support and handholding, any self-service initiative is bound to fail. Organisations must factor in the time and resources essential to deliver this support.
- Poor data quality: If the power and downstream users do not trust the data, they will stop using it. Even worse, if the business starts distrusting the data there is a high likelihood that siloed pockets of departmental BI initiatives will get started.
Making self-service BI work
Through a combination of my own experiences and several industry sources, below are a few ways a business can go about establishing self-service BI successfully:
- Identify the user population: It would be complete chaos if self-service BI was accessible to the entire company. In a data-mature organisation, only 25% of all users can be labelled as potential ‘power users’. In my experience, it is better to bring these ‘citizen developers’ on board in small groups, each with a specific focus and guidance. It is also useful to involve them in a community of interest.
- Set a self-service BI strategy: Self-service can mean a lot of different things. The business must therefore be clear about the scale of implementation, the types of users, their technical proficiency, the expectations of deliverables, and the approach to be used. It is also important to not try and boil the ocean! Starting small and building focused business areas one at a time works well.
- Keep stakeholders informed: The company must keep not only the power users but also their managers and the intended users of their work up to date. There must also be channels set up for feedback throughout the process.
- Set up comprehensive governance and quality assurance: In my mind, this is one of the most important aspects to get right. A company must put policies and processes in place to ensure what is delivered to the business is complete, accurate, timely, and relevant. A business cannot allow inaccurate or inconsistent information to be reported to the business. Once data is distrusted, it becomes an almost insurmountable obstacle to regain that trust. To this end, peer review processes are very useful.
- Use an appropriate tool: Although most of the reporting tools out there claim to support self-service BI, some are more suitable than others. I have found that tools that support curated semantic models are better for ensuring consistent and accurate reporting. Additionally, they reduce some of the technical complexities where users have to identify and implement various types of joins and unions across datasets.
- Establish a single source of the truth: Even a well-architected data warehouse and reporting environment may be too complex and detailed for citizen developers. A well-curated and quality-assured semantic data model, with pre-joined and de-normalised high-level entities presented in business-language terms works very well. This requires a lot of planning, designing and implementation before letting power users loose on the model.
- Establish a dictionary and metadata: Not only must the data be surfaced in business terms, but it should also be properly catalogued and documented. The documentation must be readily available and easily accessible by the business users. Likewise, there should also be a catalogue of existing reports and dashboards. These must also be easily accessible to enable users to search the catalogue for similar reports before embarking on a new development.
- Educate the power users: Users must be educated on the use of the tool as well as the data. Aspects of visualisation theory are also very important. It is also very useful to explain the data model to power users through workshops and hands-on implementation sessions.
- Refice and adapt: Like any good strategy implementation, regular monitoring, review, feedback, and adjustment are always useful.
While I did not discuss it in length, there must also be a close alignment between self-service BI and the broader data governance function. Self-service BI users often detect data quality and consistency issues. This means there should be good and open communication to bring these issues to the fore and make the organisation aware of how they are being handled.
