Data: in the hands of users or not?
Reading time : 2 min
When it comes to sensitive governance issues, the conditions for deploying self-BI tools are a significant topic of discussion between IT departments and other corporate departments.
Why is this?
The digitisation of organisations produces volumes of data that are growing continuously and rapidly at all levels, whatever the business line is concerned with.
No one can escape this phenomenon.
Management Boards sustain this trend.
They have long understood the strategic importance of intelligent data use.
However, in most organisations, data processing is entrusted to the IT department for apparent reasons inherited from a culture in which all IT-type processing is handled by a department with:
technical expertise, human resources (data architects, data engineers, business analysts, ....)
methods (needs analysis, data modelling, validation of the definition of KPIs, data storage management, etc.)
the capacity
to ensure optimum system security,
to deliver standardised and validated reports and dashboards.
In terms of governance, you can't do any better.
But this approach is reaching its limits in a world where the volumes of data to be processed are exploding.
Business departments' needs are no longer being met within acceptable timescales in relation to the speed at which they need to make decisions or simply answer questions they are asking themselves, the answers to which lie nestled in their own data or links with external data sources.
Self-service BI tools have emerged in response to this situation.
This software exploits data extracted from various operational systems that have previously been stored in a data lake, whether this data is structured or linked.
This software takes on much of the data ingestion work typically carried out by data organisation experts, making analysis much more accessible.
Answers to questions come much more quickly.
In terms of autonomy and agility, this is the best.
However, this approach is not without risks.
From a governance point of view, every organisation legitimately expects the same questions to lead to the same answers.
This guarantee disappears in the case of self-BI because not all users have the basic skills needed for data analysis.
This is where the issue of cohabitation between the two models arises, and the definition of a 'hybrid' governance model becomes clear.
The model will redefine the roles and responsibilities of IT and the new 'Citizen Data Scientists' within the company's business lines.
It will find the proper positioning for the cursor to ensure reasonable autonomy for the business lines and compliance with precise data quality rules.
The areas for discussion include aspects such as:
data control and access
data quality and uniformity
the effectiveness of the model (offering flexibility and capacity for innovation, with reaction times acceptable to the business)
budgetary impact
training and skills development
compliance and security
strategic alignment with the company's priority objectives.
The debate is open and urgent.
Whatever the case, even if the sensitive issue of governance cannot be dismissed out of hand, developing self-BI skills within the organisation is a real priority because even with non-strategic data that is often processed in Excel, learning tools such as Power BI offers impressive productivity gains.
But that's another story.