The Data Steward Flywheel
Jim Collins, in his work with Amazon, described the flywheel concept:
The flywheel effect occurs when small wins accumulate over time, creating momentum that keeps your business growing. Achieving the flywheel effect requires removing friction and applying force.
Another description refers to the initial effort of getting the flywheel to move, which is very hard, and it is all about the accumulation of little wins that stack up over time.
Nothing is more true about this effect when it comes to moving some of the data management responsibilities from IT to Business.
Figure 1. Helping Data Stewards overcome the challenges of taking Ownership of Data
When you set up a Data Management Program, you have to work hard on reducing the resistance. You need to listen for the resistance and be prepared to respond appropriately. Establishing effective training programs takes lots of time, so you should also establish Communities of Interest to help the stewards get tips and tricks on how to create the deliverables.
The Data Steward Flywheel works like this:
Quick Learning “How-To” Produce Deliverables, which
Improves DM Capabilities, which
Lowers Resistance, which
Increases Business Ownership, which
Needs more Learning How-To
I noticed Jim Collins focus on getting the WHO right (on and off the bus) before the WHAT. Applying the Zachman ontology to our case would mean focusing on our Data Stewardship Community before the Data Strategy.
A suggestion for a Data Flywheel is:
More data, which
Improves our learning, which
Produces better experiences, which
Increases our Users, which provides more data
Simply stated, the more data we get from our customers, the better the insights we can produce for our Customers.
Figure 2. The WHO (Data Steward) Flywheel and the WHAT (Data) Flywheel
If we connect the two flywheels, better Business (Value) Engineering leads to better Data Engineering.
Now, how much Business Engineering can we expect from our Data Stewards?
I presented a list of fundamental Data Steward Capabilities we need to train. One or two people suggested that I had lost the plot and we should consider ourselves lucky if we can get Data Stewards to:
Classify, Curate & Catalogue Data
Control the Business Vocabularies
A further comment suggested that expecting our Data Stewards to model data when Data Modelling is dead is crazy.
Figure 3. Data Steward Capability Map
My experience is quite different. Enabling the Data Stewards to own the Subject Area Model and the Business Conceptual Model was very liberating for the business data stewards. From then on, any Technical Data Steward entering the Business Department needed to understand and conform to the business data steward-owned models.
This business ownership of data management artefacts is precisely the outcome we are looking for.
Empowering our business to drive business value from data is what we require. Still, we have to apply the flywheel approach when working with organisations that are not data-centric.
What does this have to do with the Unified Metamodel?
Real learning can not happen in a formal classroom or virtual learning session. It happens when you have to apply your knowledge to create deliverables.
Your competency is further enhanced during construction feedback and review using the appropriate scorecard. Intense reviews are extremely helpful in testing your understanding. Take advantage of these learning opportunities.
To ensure that we get the most out of the learning whilst doing experience, following the procedures defined by your organisation provides both parties with valuable feedback:
Are the practices helpful in achieving the desired results?
Are you improving your Data Management Capabilities?
I would like to get your feedback.
Please can you answer the following questions by providing a score between 1 and 5. 1 is “Definitely Not” to a 5, which is Definitely:
Do you believe that we should expect more than Classify & Vocabularies from our Data Stewards?
Do you believe that Data Stewards can or should learn about Data Modelling?
Do you agree that Data Steward Training has to be more hands-on (completing How-To Excel templates)?
Join the conversation on LinkedIn