When Data Democratization Backfires: The Paradox of Freedom
Finding balance between freedom and usability for non-technical users
France was plagued by economic hardship, starvation, and discontent in the late 18th century.
The french people hoped each year would be better than the last; none were.
The king’s—Louis XVI—inability to address these issues fueled the growing unrest.
The dam broke.
The commoners stormed the bastille. The royals—overthrown!
What had been reserved for the royals—luxurious food, palacial buildings—was now distributed among the people.
Finally! At long last!
The people could breath easy now. The world would be fairer and more just.
…for about 5 minutes.
Once people saw what was up for grabs, everyone wanted a slice. More and more revolutionary groups ate each other.
For years, no one knew who would end up on top.
Cities burned; blood ran in the streets; people starved.
This historic upheaval reminds us how shifts in power dynamics can lead to unexpected results, ones far from the predicted results.
The French Revolution in summary:
Goal: equality across people, plentiful bounty shared, happiness
Result: horrible bloodshed, famine, oppressive regime installations1
The modern business environment (usually) involves fewer guillotines than eras past.
The guiding principles of human interactions though, are the same.
As the French Revolution ushered in a new era aiming for democracy and equality, the modern business landscape is experiencing a revolution of its own: the democratization of data.
This breaking down of traditional silos aims to empower all levels of an organization to access, analyze, and leverage data.
No longer confined to data scientists and IT specialists, data democratization aims to create a culture where informed decision-making and innovation thrive across the board.
And it might
…for about five minutes.
In the push for data democratization, we often hear that giving everyone access to data will lead to better decisions across an organization.
But there’s a crucial aspect that's often missed: the human balance between freedom in choosing metrics and how usable the data actually is for non-technical users.
Drawing from Benn Stancil's Substack post on metric architecture, we can see an inverted U-curve where the level of constraints on metric selection and real data democracy interact.
…self-serve is a misunderstood (or, at least, misrepresented) problem.
Because the most common question people have is “How often did this thing happen?,” effective self-serve is less about complex analysis and more about metric extraction.
People “want to choose from a list of understood KPIs, apply it to a filtered set of records, and aggregate it by a particular dimension.
It's analytical Mad Libs—show me average order size for orders that used gift cards by month.”
This blog post explores why fewer constraints in metric selection can lead to a situation where only technical people can effectively use the data, thus undermining data democratization.
Data Democratization in summary:
Goal: employee equality, plentiful (business) bounty shared, happiness
Result: horrible data security, few accurate insights formed, oppressive tech stack installations
The Inverted U-Curve of Data Usability
Think of metrics calculation freedom as freedom from a political regime. As Nelson has laid out below, there are hefty costs for BOTH a nonexistent or heavily tyrannical government.
Just switch ‘government’ with ‘IT’ or ‘Data Ops’2
At one end of the spectrum, we have a highly constrained environment where strict controls govern which metrics can be accessed and how. This creates a bottleneck where only IT or specialized data teams can perform analyses, leaving business users waiting for insights.
Constraints hinder data democratization by restricting access and slowing down decision-making processes.
On the other end, a completely unconstrained environment, where users can freely select and define their metrics, sounds ideal. It leads to chaos.
Without standardized definitions and a clear understanding of data sources, users might pull incorrect metrics, misinterpret data, or make poor decisions based on flawed analyses.
In this scenario, only those with deep technical knowledge can be trusted to navigate the complexity and extract meaningful insights. Back to square one.
The Goldilocks Zone of Metric Selection
The sweet spot lies in the middle—what we might call the Goldilocks Zone of metric selection. Here, there are enough constraints to ensure consistency, accuracy, and reliability of the metrics, but not so many that it stifles access and usability for non-technical users.
This balance allows true data democratization, where users across the organization can confidently access and use data without relying on technical intermediaries.
Practical Implications
Standardized Definitions: Implementing standardized metric definitions ensures everyone speaks the same language when it comes to data. This reduces misinterpretation and keeps metrics consistent across reports and analyses.
Con: This is technically a constraint. Someone arbitrarily defines a term and everyone has to learn that term to use the data.
User Training and Support: Providing training and support for business users helps bridge the gap between technical and non-technical staff. When users understand the data they are working with, they are more likely to use it correctly.
Con: This puts a lot of load onto the trainers in addition to their existing responsibility
Self-Service Analytics Tools: Investing in intuitive self-service analytics tools empowers users to perform analyses within defined constraints. These tools can offer guided workflows that help users pull the correct metrics without needing deep technical skills.
Con: Another tool, another data schema, another round of vendor lock-in
Governance and Oversight: Establishing a governance framework that oversees metric selection and usage helps maintain the balance between freedom and control. Regular audits and feedback loops ensure the system evolves based on user needs and organizational goals.
Con: Usually requires significnat time investment
Conclusion
Complete freedom in metric selection sounds appealing.
In Thomas More’s utopia, slaves had golden handcuffs because the concept of non-human-exploiting automation didn’t exist.
In the same way, our current zeitgeist around metric utopia is just as limited a view.
Total freedom from constraints often leads to a scenario where only the technically inclined can navigate the messy data landscape—the same place enterprise data was 30 years ago.
We need to break out of our view and find a path that meets the human element where it is.
To achieve functional data democratization, we have to balance constraint and freedom.
We need to protect users from creating bad data but also allowing freedom to make informed decisions.
If you’re careful, you can foster a data-driven culture that is inclusive, accurate, and effective—But I’m not convinced that the common methods to achieve that end state are effective.
By navigating the inverted U-curve instead of flying blindly to one end, the right constraints in data usage will lead us to consistently richer, more accurate, more actionable insights.
And quick de-installations for the most part.
Sorry IT and Data Ops guys. Its not your fault. It’s part of the beast.