Overview
In this episode of The Dashboard Effect, Brick Thompson and Caleb Oaks address one of the most fundamental and most frequently underestimated challenges in business intelligence: data that cannot be trusted. The conversation moves from a clear taxonomy of the four types of untrustworthy data to a set of practical strategies that organizations can implement to improve data quality, assign accountability, and build the cultural infrastructure that keeps data reliable over time.
For any organization where reporting inconsistencies have eroded confidence in the data, or where teams are spending more time questioning numbers than acting on them, this episode provides both a diagnostic framework and an actionable path forward. See how Blue Margin’s Managed Data Service helps organizations establish the data governance and quality practices that make reporting trustworthy and keep it that way.
What This Episode Covers
The Four Types of Untrustworthy Data (0:26 – 1:04)
The hosts open by categorizing the ways data fails: stale data that is outdated or no longer actively collected, unclear data where conflicting naming conventions make consistent aggregation impossible, inaccurate data that is simply wrong, and missing data where no information exists at all. The taxonomy is useful because different types of data quality problems require different remediation approaches, and organizations that treat all data quality issues the same way tend to apply solutions that address the symptom they are most familiar with while leaving other categories unaddressed.
Define Your Metrics (4:27 – 5:46)
Company-wide alignment on how key metrics are defined is a prerequisite for consistent reporting. When different teams calculate net profit, revenue, or customer count using different formulas or different data sources, the resulting inconsistency is not a data quality problem in the traditional sense. It is a definitional one, and it produces the same erosion of trust as genuinely bad data. Establishing shared definitions and documenting them in a place where they are visible and accessible is foundational work that many organizations defer and pay for repeatedly in reconciliation cycles and disputed reports.
Assign Ownership (5:50 – 6:53)
Designating specific owners for metrics throughout the organization creates accountability for the quality of those metrics rather than leaving data quality as everyone’s general responsibility and therefore no one’s specific one. Metric owners are motivated to proactively address issues in their domain because those issues reflect directly on the accuracy of outputs they are accountable for. That accountability structure is more reliable than quality checks that depend on someone noticing a problem without having a stake in the outcome.
Expose the Data (6:57 – 8:19)
Giving frontline workers visibility into broader data views changes how they relate to the data they generate. When employees can see how their inputs contribute to aggregate reporting, they are more likely to notice when something looks wrong and more motivated to correct it. The person who enters a customer name inconsistently is unlikely to discover the downstream aggregation problem that inconsistency creates. The person who can see the aggregation has every reason to fix the entry at the source.
Utilize Data Quality Services (8:25 – 9:43)
Tools and processes that standardize and map disparate entries to a clean master list address the unclear data problem systematically rather than case by case. Product names, customer names, and other fields that accumulate variations over time require a mechanism for resolving those variations into a consistent format that reporting can rely on. Implementing that mechanism programmatically is more scalable and more reliable than manual review and correction.
Foster a Data Culture (10:33 – 11:30)
Data quality is not a project with a completion date. It is an ongoing operational commitment that requires the same kind of sustained attention as any other business-critical process. Organizations that approach data cleanup as a one-time initiative tend to find themselves back in the same position twelve months later as the data environment continues to evolve. Shifting the organizational mindset from data project to data practice is what converts a cleanup effort into a sustained capability.
Build a Community of Practice (11:31 – 11:57)
Creating a dedicated space where employees can ask questions about the data, understand how it is structured, and surface concerns about quality builds the trust and organizational knowledge that sustains a data culture over time. A community of practice turns data literacy from an individual capability into a shared one, and the questions that surface through it often identify quality issues and definitional gaps that would otherwise go unnoticed until they affect a report that reaches leadership.
Who It’s For
This episode is worth your time if you are a data or BI team lead dealing with ongoing challenges to reporting credibility and wanting a structured approach to addressing data quality at the source rather than in the reporting layer, a business leader whose team spends more time debating whether the numbers are right than deciding what to do about them, a data governance or operations professional trying to establish accountability structures around data quality without creating bureaucratic overhead, or any organization that has delivered data infrastructure and found that inconsistent data quality is undermining the adoption of the tools built on top of it.
Why It’s Worth a Listen
The four-type taxonomy is the most immediately useful part of the episode because it converts a vague organizational complaint about bad data into a specific diagnostic. Different types of data quality problems require different solutions, and the ability to identify which type you are dealing with is the prerequisite for choosing the right response. Organizations that treat all data quality issues the same way tend to solve one category while the others continue to erode trust.
The ownership model is worth taking seriously as a governance design principle. The absence of clear metric ownership is one of the most consistent structural gaps in mid-market data environments, and the problems it creates, inconsistent calculations, unresolved discrepancies, disputes over which version of a number is correct, are all symptoms of the same underlying accountability gap. Assigning owners does not require a new system or a significant investment. It requires a decision and a conversation, and this episode makes the case for having both.
And the community of practice recommendation is the most scalable part of the strategy for organizations trying to build data quality into the culture rather than policing it through oversight. A forum where people can ask questions without fear of judgment produces more useful information about where the data is actually failing than any audit process, and it builds the organizational relationship with data that turns quality improvement from a compliance activity into a shared interest.
