Iron - Saúde ao seu alcance

Tadeu Kwiatkowski Ribeiro

Data Governance Steward

Tadeu Kwiatkowski Ribeiro is a data and analytics leader with over 10 years of experience designing, implementing, and scaling solutions across healthcare, mining, travel and tourism, and other sectors. Tadeu combines technical expertise with business strategy and transformation, grounded in a Computer Engineering degree, an Executive MBA in Business Analytics and Big Data, continuing studies in strategic management and leadership at Stanford University, and hands-on experience building businesses and advising organizations in complex enterprise environments.

There is a moment in a meeting that many professionals recognize. A report is on the screen, and someone says, "That number does not look right." The conversation stops being about the business and becomes about the data. Five minutes later, someone opens a spreadsheet they trust more than the official report. Something has broken that is harder to fix than any data pipeline.

This pattern repeats across organizations. The instinct is to look for a technical error, but the cause is increasingly not there. It is often an accumulation of interpretations along a chain that starts before the data is even processed, and sometimes before the system records the first transaction.

Where the Loop Actually Begins

The actual problem almost never starts with a calculation error. It is a business definition that everyone assumed was precise and was not. What counts as an "active customer" or a "completed order" feels obvious in conversation but turns ambiguous once encoded into a product, a system, a query, or a dashboard.

That ambiguity then travels. A business rule becomes a requirement, then a product behavior, then a database structure, then data pipelines, transformations, calculations, filters, and charts. Each step is an interpretation. By the time a number reaches an executive, it has passed through a dozen translations made by teams that do not always share the same context or language.

  ​What does not get automated is the understanding behind the numbers: business acumen, communication, and the judgment to turn insight into business value.  

Disagreement is rarely about right and wrong. Sometimes two reports diverge because each is the end of a different chain: one counts a sale at order confirmation, another at fulfillment, one defines churn at sixty days, another at ninety. Sometimes the dashboard is technically correct, but the business rule behind it was never clear enough to encode the same way twice. Sometimes the product behaves differently because the requirement was incomplete, or because users adopted the product in a way nobody mapped. And sometimes everything seems aligned until the source system never stored the field needed to answer the business question.

These are not always data and analytics mistakes. Many come from what happened before the data reached the data team: unclear business needs, weak SLAs, incomplete requirements, ambiguous rules, disconnected teams, and systems that capture what was built rather than what the business later needs to understand.

The Quiet Cost

The expensive part is not the rework. The deeper cost is that meaningful business questions stop being answered cleanly. Teams build workarounds: manual reconciliation, side spreadsheet, a number forced to match the expected total. Time gets wasted hunting where the cause does not live. Analytics loses its authority.

Many organizations believe they have a data lake or a data warehouse, but what they actually have is more like a data swamp. The signs are familiar: weak schemas, foreign keys broken during migrations, the same concept spelled five ways, fixes stacked on fixes. The simplest questions stop having stable answers. The infrastructure works, but the trust is gone.

Why Technical Cleanup Will Not Save You

The instinct is to fix this with cleanup, refinement, standardization, or another layer of validation. All are necessary, but they often treat the symptom. Cleanup repairs records. Refinement polishes the surface. Standardization aligns formats. None solves the upstream confusion where divergence begins.

This is where governance earns its keep, and the term needs precision. It is not a committee, a policy binder, or a layer of approvals. Its job is narrower and more important: making information reliable enough to act on. That requires definitions, ownership, lineage, and the discipline to connect business rules, product behavior, system design, and analytics logic. Most of that work depends less on tools and more on communication.

Rebuilding the Trust

Trust comes back through consistency, not announcements: the same definitions, in the same place, every time, until people stop checking. What closes the loop is the standing work of the data and analytics function.

That function must sustain three conversations. With the business team, it needs to understand goals, rules, needs, and SLAs. With the product team, it needs to understand how the product was designed, what was delivered, and how users behave in real life. With developers and IT, it needs to understand how systems behave in the backend, including what is stored, what is not stored, and what was never designed to answer future business questions.

What surfaces in dashboards usually traces upstream. It may be a gap in business rules, a missing SLA, a requirement that was never clear, a product behavior that does not match the process, an unmapped user journey, or a source system that cannot record the minimum context needed for reliable analysis. A good data leader earns the role by finding those gaps, translating them, and helping the right teams close them.

Data engineering, data modeling, algorithms, statistical models, and calculations remain the foundation. But this is also where automation and AI have advanced fastest, reducing repetitive work and lowering human error for data engineers, data scientists, data analysts, and other data professionals. What does not get automated is the understanding behind the numbers: business acumen, communication, and the judgment to turn insight into business value.

Promoting data leaders on technical skill alone limits maturity. Prioritizing technology over business objectives stalls initiatives. The function is increasingly a business, product, and communication discipline that uses technology. Success sits in the balance between technical command and business judgment.