Method
I developed this six-field framework with Dr. Mariana Hebborn and Prof. Dr. Heiko Beier. The approach has proved its value repeatedly in my work. We brought the method and our experience with it together in “How to Become a Data-Driven Organization”.
A report may take hours of manual work each month. You may need to choose a platform or hire the first person for your data team. We look at what the task requires and what we can put in place for the tasks that follow.
These tasks are also where we work on culture. When business teams agree on a data definition, they practise making decisions together. When your team operates a solution itself, it develops skills and takes ownership. The six fields help me shape these changes alongside the technology.
Choose a task to see the related work in the other fields. We use these connections to plan the next useful step.
ILLUSTRATIVE EXAMPLE
Sales leadership gets a report for its variance analysis. The existing manual preparation should be retired.
People: An internal team member helps maintain the process and later takes it over.
Data Quality: Validation rules stop an incomplete delivery from being released.
Data Architecture: The required data connection is documented. Other reports can reuse it when the scope fits.
Data Governance: The business team decides which orders count and how cancellations are treated.
Automation: Calculation and checks run on a schedule; failures follow an agreed handling process.
Analytics & AI: The report answers the agreed business question. An ML model is not required.
POSSIBLE NEXT STEPUse the next regular reporting cycle to check which manual work is still needed.
The examples and distances in the diagram are illustrative, not measured client results or maturity levels. A single action does not have to improve every field.
Who will take responsibility?
We involve the person who will look after the solution. If that person is not yet in place, we define the role and the skills it requires.
Which error would distort the decision?
The business team identifies which data needs to be reliable for the task. We implement suitable checks and agree who handles errors.
How does the application get the right data?
I work with the team to review existing connections and identify what is missing. Freshness, permissions and operation inform the decision.
Who decides what a metric means?
We clarify terms, permitted use and business ownership for the specific use case. We also agree how changes are handled.
Which manual steps can we remove?
We make recurring preparation and checks executable. The process includes notifications and a way to restart after failures.
What should happen as a result?
We work with the future users to define which decision an analysis or model should support. We compare its usefulness with the current approach.
After the first implementation, we follow a regular working cycle. We check whether the result helps, where manual work remains and whether someone in the team can take over.
An agreed definition can be reused in the next data product. Another application can use the same data connection. Metadata makes dependencies visible, and your team learns to use it in day-to-day operations. Connecting these steps across the six fields builds the organizational and technical foundations for a data fabric.
More about our bookTell me what you want to do with your data and where you need support. We can work out whether I’m the right person to help.