Breaking Down Silos: Unifying Data Across 7 Business Units at Wells Fargo
Fortune 500 Financial Services Institution
THE CHALLENGE
When Bad Data Drives Million-Dollar Decisions
Wells Fargo, like many large financial institutions, had grown through acquisition and organic expansion, creating 7 distinct business units operating as independent kingdoms. Each unit had its own data systems, definitions, and reporting processes. The result? An enterprise blind spot that threatened strategic decision-making.
The data landscape was dangerously fragmented:
- 70% of cross-business unit reports contained incorrect data—errors that went undetected and unvetted by stakeholders
- No standardized metrics across business units—"revenue," "profitability," and "cost" meant different things to different teams
- Manual, weeks-long reporting processes that were outdated by the time they reached executives
- Zero cross-unit visibility—impossible to compare performance or identify best practices
- Siloed "kingdoms" where business unit leaders resisted standardization to protect their autonomy
The Critical Problem
The bad data wasn't just inconvenient—it was actively misleading leadership. Incorrect calculations of costs, timeframes, and profitability were driving strategic decisions. Developers had created "creative stopgaps" to fill data gaps, resulting in reports that looked complete but were fundamentally wrong.
Technical Chaos Masking as Solutions
Creative workarounds by siloed development teams had created a web of incorrect calculations. Cost allocations were wrong. Profitability metrics were unreliable. Timeframe data was inconsistent.
Political Minefields
Business unit leaders—the 'kings and queens' of their siloes—had little incentive to standardize. Each unit had built its own metrics that made them look good. Cross-unit comparisons threatened their autonomy.
Previous Failures
Multiple attempts at data unification had failed. On one memorable call, a consultant was fired on the spot for lack of progress—a stark reminder of how previous technical-first approaches had crashed against organizational resistance.
Executive Blindness
Leadership had no reliable way to answer fundamental questions: Which business units were truly profitable? Where should capital be allocated? Which operational models should be replicated across the enterprise?
"We thought we were hiring someone to build Tableau dashboards. What we actually needed was someone who understood that 70% of our cross-business data was wrong—and who could navigate the political minefield of getting seven independent kingdoms to accept a single source of truth."
The Stakes
With millions in strategic decisions being made based on faulty data, Wells Fargo needed more than dashboards—they needed data they could trust. Previous consultants had failed by treating this as a technical problem. The reality: this was an organizational change challenge disguised as a data engineering project.
THE SOLUTION
Listen First, Architect Second, Negotiate Always
The key insight: You can't force data standardization on resistant stakeholders—you have to make them want it. Rather than arriving with a predetermined technical solution, I began by listening to each business unit's unique needs, pain points, and concerns. Only after understanding their perspectives could I architect a unified framework that served everyone.
The approach followed organizational change theory: Unfreeze → Change → Refreeze. First, demonstrate why the current state was broken (the 70% error rate). Second, implement the technical solution while maintaining trust. Third, prove the value so teams would adopt the new standards permanently.
This required equal parts data engineering and diplomatic finesse—navigating egos, building credibility with skeptical internal developers, and securing buy-in from C-level stakeholders who controlled budget and political capital.
Trust-Based Discovery
What we did: Conducted deep listening sessions with all 7 business units to understand their unique objectives, metrics, and pain points. Identified what data existed, what was missing, and what 'creative stopgaps' were masking data quality issues.
Impact: Built credibility with resistant stakeholders by demonstrating understanding before proposing solutions. Turned skeptics into advocates.
Unified Data Model Architecture
What we did: Designed a common schema that didn't force-fit existing systems, but rather identified which metrics should exist across the enterprise and then reverse-engineered how to generate them from disparate source systems.
Impact: Created a single source of truth while respecting business unit autonomy—metrics were standardized, but units kept operational independence.
Phased Implementation with Quick Wins
What we did: Started with the most universally needed dashboards (financial performance, customer dynamics) to prove value quickly. Used early successes to build momentum for broader standardization across all 200+ metrics.
Impact: Eliminated resistance by demonstrating tangible value before asking for major changes. Each success made the next phase easier.
Executive Visibility Platform
What we did: Deployed 25 Tableau dashboards providing daily updates on financial performance, customer dynamics, operational efficiency, and strategic growth metrics. Gave executives side-by-side business unit comparisons for the first time.
Impact: Enabled data-driven decision-making in hours instead of weeks. Uncovered performance gaps and best practices that were invisible before.
Technical Implementation Details
The Data Unification Challenge:
Unlike typical schema consolidation projects, this wasn't about merging 7 existing data models into one. Instead, it required:
- Metric Inventory: Analyzing what data existed in each division and what was missing
- Ideal State Design: Determining what should exist for enterprise-wide decision-making
- Reverse Engineering: Figuring out how to generate standardized metrics from inconsistent source systems
- Gap Filling: Building new data pipelines where metrics didn't exist at all
The Architecture:
- Source Systems: Oracle databases across 7 business units with inconsistent schemas
- Data Integration: Python-based ETL processes to extract, transform, and load into unified warehouse
- Standardization Layer: Common KPI framework covering financial performance, customer dynamics, operational efficiency, and strategic growth
- Visualization Layer: 25 Tableau dashboards with daily automated updates (minimum)
Technology Stack
What Made This Work
The turning point came when early dashboards revealed performance gaps between business units that executives hadn't been able to see before. Suddenly, the value of standardization became obvious—not because I mandated it, but because the data told a compelling story that leadership couldn't ignore. Business unit leaders who initially resisted standardization began requesting access to the unified metrics to understand why their units were underperforming compared to peers. The "kingdoms" started voluntarily adopting common KPIs because visibility created accountability.
THE RESULTS
From Blind Spots to Clear Vision
Data Quality Transformation
Eliminated 70% error rate in cross-business reports. Identified and corrected bad data affecting profitability calculations, cost allocations, and timeframe projections across the enterprise.
Reporting Speed
Transformed manual reporting processes taking 2-3 weeks into automated daily dashboard updates. Executives now access current data in hours instead of weeks.
Business Units Unified
Created single source of truth across 7 previously siloed business units. Enabled apples-to-apples performance comparisons for the first time in company history.
Metrics Standardized
Standardized 200+ metrics across financial performance, customer dynamics, operational efficiency, and strategic growth. Deployed in 25 Tableau dashboards with daily updates.
Cross-Unit Visibility
Gave executives their first-ever ability to compare business unit performance side-by-side. Uncovered best practices and performance gaps previously invisible in siloed data.
Hours Saved Monthly
Eliminated hundreds of hours of manual reporting work through automation. Teams shifted from data compilation to data analysis and strategic decision-making.
Additional Wins
Prevented Costly Mistakes: The 70% error rate wasn't just an accuracy problem—it was actively misleading strategic decisions. Incorrect profitability data could have led to misallocation of millions in capital. Faulty cost calculations could have resulted in unprofitable product launches. By establishing data integrity, we prevented decisions based on fundamentally flawed information.
Created Accountability Through Transparency: Before unification, business unit leaders could hide behind inconsistent metrics. After implementation, transparent performance comparisons created natural accountability. Underperforming units couldn't blame "different definitions"—everyone was measured the same way.
Enabled Evidence-Based Strategy: Leadership could finally answer critical questions: Which business models should be replicated? Where should capital be invested? Which operational practices should be standardized across the enterprise? These decisions shifted from intuition and politics to data-driven analysis.
Cultural Shift from Data Compilation to Analysis: By eliminating weeks of manual report generation, teams pivoted from being data gatherers to data analysts. The organization moved up the value chain—less time formatting spreadsheets, more time generating insights.
Established Foundation for Future Innovation: The unified data model and automated pipelines created infrastructure for advanced analytics, machine learning, and predictive modeling that would have been impossible with siloed, low-quality data.
"We hired a consultant to build Tableau dashboards. What we got was someone who understood that our real problem was 70% of our cross-business data was fundamentally wrong—and who had the political savvy to get seven independent kingdoms to accept a single source of truth. Previous consultants crashed against the organizational resistance. This one navigated it like a diplomat while delivering enterprise-grade technical solutions."