How to clean and organize your BI environment?

Imagine this: You’re looking at a report that shows your company has 80 products under its purview. Then you glance at another dashboard, and it says 72. Which one would you trust? Probably neither. You’d end up doing what most stakeholders do—send off an ad-hoc request for a spreadsheet download and start counting the numbers yourself. It’s frustrating, but it’s a reality that those of us in business intelligence face all the time.

And this isn't just a one-off scenario. It’s a pattern. Business teams keep coming back with questions like, “This number seems off,” “This metric doesn't align with the new definition,” or “This metric is reported differently on another dashboard.” These aren’t just annoying questions; they’re signs of a bigger problem. Inconsistencies like these erode trust in the data and the decisions that come from it.

If I reflect on my experience as a BI Analyst so far, so much of my time has been spent on remediation and operational work. Fixing a broken data pipeline, hunting down an issue in a report, and going around in circles just to agree on the basic definitions of metrics. I’d love to be innovating—using data to find new insights and drive the business forward. But instead, I’m stuck in the trenches, just trying to keep things running smoothly. And I know I’m not alone; this is the harsh reality for many of us in business intelligence. It’s frustrating because these problems seem so basic, yet we still can't get them right.

What is the problem?

It’s not a lack of resources. We have all the tech stack, we have plenty of data, we have plenty of dashboards, and we have plenty of analysts. A new business question pops up, and we immediately create a new dashboard (or at least add a new chart to an existing one). But we’re treating the symptoms, not the disease. Over time, this results in a rapid increase of dashboards that are underutilized or redundant, leading to inefficiency and confusion.

So, what’s the real problem here? It’s not lack of resources but rather a lack of structure and intention in how dashboards are created, maintained, and used.

Identifying the Root Causes

To address the problem effectively, we first need to understand why dashboards go unused. Common issues include:

1. Dashboard Overload: With every new business question, new dashboards are created, leading to a cluttered environment where users are overwhelmed by the sheer volume of available dashboards.
2. Lack of Consistency: When metrics and definitions vary across dashboards, users lose trust in the data, leading to decreased usage.
3. Poor Usability: Dashboards may not be user-friendly or fit for purpose, reducing engagement.
4. Lack of Documentation: Without clear documentation, understanding the data sources, logic, and purpose behind each dashboard becomes challenging, especially when team members leave or projects change hands.

How We Addressed These Challenges in Our Team

Faced with these issues, our team decided to take a step back and reassess our approach to dashboards. We paused all new dashboard requests and conducted a comprehensive review of all dashboards delivered to date. Here’s the process we followed to clean up and reorganize our dashboard environment:

The Framework for Dashboard Cleanup and Optimization

Step 1: Assess Usage and Identify Key Dashboards

Goal: Determine which dashboards are actively used and which are not.

• Action: We analyzed dashboard usage metrics, focusing on the number of views each dashboard received. This allowed us to quickly identify the top dashboards that were used most frequently by our stakeholders.
• Outcome: This step provided immediate insights into which dashboards were valuable and widely used, giving us a clear starting point for prioritization.

Step 2: Ensure Consistency Across Dashboards

Goal: Verify that metrics and data across dashboards are consistent and reliable.

• Action: We reviewed the frequently used dashboards to ensure that metrics reported were consistent across all dashboards. This included synchronizing definitions and ensuring that each dashboard provided enough context to reduce stakeholder questions about metric definitions.
• Outcome: This helped us regain the trust of stakeholders, as they could now rely on the accuracy and consistency of the data presented.

Step 3: Evaluate Fit for Purpose

Goal: Improve the usability and relevance of dashboards that were not frequently used.

• Action: We shifted our focus to underutilized dashboards, seeking feedback from stakeholders on how these could be improved or made more actionable. Given the turnover in our team, we faced challenges in understanding the original logic and data pipelines behind these dashboards.
• Outcome: This process highlighted the need for better documentation and the creation of structured, well-documented workflows.

Step 4: Implement a Structured Documentation System

Goal: Create a robust documentation system to make all dashboards easier to understand, maintain, and update.

• Actions:
  • Naming Conventions: We standardized the naming of Alteryx workflows to match the dashboard names, making it easier to trace workflows back to the corresponding dashboards.
  • Template Usage: We utilized Alteryx templates to document the purpose, schedule, and any changes made to workflows.
  • Metric Definitions: We documented the functional definitions of each metric, including its data source, table, and any additional details. Links to these definitions were added to the dashboards for easy reference by stakeholders.
  • Repository Creation: We set up a repository with a folder for each project, containing a README file that outlines the project’s purpose, stakeholders, related dashboards, and any relevant SQL code or workflow information. This repository is actively maintained and updated with any changes.
• Outcome: This structured approach to documentation has significantly improved our ability to manage changes, onboard new team members, and ensure transparency and continuity in our work.

By implementing these steps, our team has greatly enhanced the usability and reliability of our dashboards. The structured documentation ensures that any team member can quickly understand the background, purpose, and logic behind each dashboard. This has made change management easier and reduced the risks associated with changing team members. By following this framework, any team can enhance the value and effectiveness of their dashboards.

However, this is just the beginning. To further elevate the effectiveness of the dashboards we deliver and to build more trust in our data, we can do a lot more.

What more can be done?

1. Continuous Engagement with Business: We can hold working group sessions with stakeholders twice a year to review the dashboards provided to them and assess if they continue to meet their needs. As the business evolves, our dashboards should evolve too.
2. Assign Ownership: It's crucial to assign a business owner and a technical owner to each dashboard to ensure accountability and maintain data quality. Additionally, the data assets powering each dashboard should have designated data owners responsible for ensuring data accuracy.
3. Open Communication: Establishing clear communication channels is vital. A centralized hub for updates and incident reports can keep everyone informed and aligned, preventing surprises when changes are made based on a request from one stakeholder that others are unaware of. We can also schedule office hours to provide stakeholders with greater visibility into dashboard features and guidance on how to use them effectively. This is also an opportunity for stakeholders to ask questions and clarify any concerns they might have.
4. Data Audits: Regular audits are essential to maintaining data integrity. By periodically reviewing our data sources, transformations, and pipelines, we can identify and address discrepancies or errors before they impact dashboards.

By redefining our approach to dashboard management, we're not just addressing surface-level issues; we're fundamentally changing how we leverage data to drive meaningful insights. It's about moving past the reactive work of fixing errors and redefining metrics to create a more structured, transparent, and intentional data environment. This shift allows us to build dashboards that stakeholders can trust and use effectively, ultimately fostering a culture where data is a catalyst for innovation rather than a source of frustration. While there's always room for improvement, these efforts are paving the way for a smarter, more agile approach to business intelligence.