I recently had an enlightening conversation with a new friend, Zoë Zeiders, who has had a successful career in project & program management while heavily leveraging data to make decisions and move work forward.
The goal was to hear her perspective, then enhance the community that I built so that data-adjacent professionals could learn how to work better with analytics professionals, and vice versa. This naturally led to a more meta conversation about analysts, data scientists, engineers and their stakeholders.
I thought it would be helpful to unpack some of her insights here as well as some of my own experience and research on these issues, and propose some potential solutions.
The Core Problem
Communication gaps or lack of data literacy can prevent stakeholders from clearly expressing their needs to analysts (or the broader analytics team).
On the other hand, analysts may struggle to convey complex data findings in a business-actionable way.
This disconnect leads to technically sound reports but are too noisy, miss the actual objective, and ultimately end up in the dashboard graveyard because they fail to drive stakeholder decisions and business value.
Start by Improving Communication and Collaboration
Gathering Requirements
If you want to be a successful analyst, you must deliver on requests for analyses and find the insights that drive decision making or solve problems. In order to do that accurately, you must understand the problem the business is facing.
During the requirements gathering phase of the analytics workflow, don’t be afraid to stop the conversation and ask the opposite party for clarity. A 2-minute discussion to clarify and provide context can save hours of wasted time later. Repeat their responses, as you understand them, to confirm clarity as needed.
“Why” & “What” Questions Will Help You Understand the Problem:
Ask questions that start with “why” e.g. “Why is it important for the business to solve this problem, and why now?”
After answering “why” questions, move to “what” questions e.g. “What 3 metrics will you need to make this decision effectively” and “What decisions do you anticipate making based on the insights derived from this analysis?”
Only after addressing the “why” and “what” can you effectively go after the “how”.
Effective Data Storytelling
The main objective of data storytelling is to convert raw data into clear, engaging narratives that drive both understanding and decisive action. This process starts with defining your specific goals and thoroughly understanding your stakeholders’ challenges.
Once you have gathered your requirements and performed your analysis to unearth the required insights, craft a compelling narrative. This should include a clear plot, essential context, and a distinct call to action, using appropriate and consistent visualizations to enhance your message.
It's vital to present data with objectivity and transparency, integrate subject matter expertise for context, simplify complexity using plain language, and incorporate interactivity where possible.
Being crystal clear on your core message cannot be overstated. Going back to my conversation with Zoe, she said to me that she would sometimes just title a page in a report with exactly what insight she wants the stakeholder to leave with.
For example, “Sales are down 2% quarter-over-quarter.” Even if the chart makes this clear as day, this ensures the key message is front and center.
Building Reports that Don’t End Up in the Graveyard
Mastering Business Intelligence (BI) Requirements Engineering
You asked the important "why" and "what" questions during your initial requirements gathering, dived deep into the actual analysis, and started to craft the story your data tells.
Now, you can begin thinking about how you can best visualize that story and translate it into an effective BI solution that drives action and decision-making. Before you jump into building dashboards, it's vital to further refine what "good" looks like for your stakeholders.
This means prioritizing key data sources and impactful KPIs to prevent information overload. Remember, too much noise on a dashboard, especially for non-technical users or those with less data literacy, often means it will simply be ignored. We want clarity and impact.
To help solidify this and ensure the BI solution is aligned with the stakeholder’s “Vision of Good”, apply the requirements you initially gathered to the Pain, Need/Dream, Fix model:
Pain: Current frustrations and operational challenges.
Need: Immediate, essential requirements for functionality.
Dream: Ideal future aspirations and capabilities. This uncovers richer requirements, moving beyond reactive fixes.
Fix: This is your solution and how it solves the “Pain” and should deliver immediately on the “Need”, eventually on the “Dream”.
Stakeholder-Centric Report Design
Adopting a stakeholder-centric design philosophy is key for creating intuitive and actionable reports. Every design choice should be guided by a deep understanding of end-user needs, their specific decision-making processes, and their level of data literacy.
Key strategies include tailoring reports to your audience by adapting to their required decisions, relevant metrics, desired detail, and preferred formats. Always prioritize the most critical takeaways by presenting them upfront for immediate clarity and impact.
Remember that the entire design process itself should be iterative, constantly refined through continuous stakeholder feedback. Consider holding “office hours” during your UAT (User Acceptance Testing) period so that end-users can interact with it and offer suggested feedback.
Optimizing Report Utility and Value
Provide Flexible Data Views: Empower users with filters, selectable columns, and saved configurations (with appropriate governance).
Strategic Refresh Cadence: Align report refresh frequency with business needs and decision-making windows. In a previous role, I had to work with reports where refreshes often would get delayed by a few hours. Because they were already scheduled to refresh close to when I needed them, this stalled decision making. I had to make the case to change that refresh cadence.
Articulate Value Proposition: Clearly define the benefits each report provides. Is this report actually needed? Run a Cost-Benefit-Analysis if needed. Is there already another report that provides similar data? If so, what does this new report uniquely provide that the other one does not?
Reporting Implementation: Avoiding Common Pitfalls
Slow report generation performance: Nothing kills a report more than it taking forever to load. It’s one thing if the initial load takes some time, but if every filter & slicer takes additional time, it’s going to be abandoned.
Insufficient user training leading to poor adoption: in a previous role I held, I would create training documentation when I would release new reporting for leaders. This would drastically increase adoption, and often lead to users reaching out with educated questions and enhancement suggestions.
Choosing seemingly cheaper solutions with high hidden costs: The cost of making bad decisions with poor data far outweighs the cost of purchasing a necessary BI tool.
Treating BI solely as an IT project, neglecting its business purpose: Any tool that you implement just for the sake of implementing a tool will be ultimately a waste of resources and end up unused.
Expecting users to easily abandon familiar tools (e.g., spreadsheets): I joke plenty about the common “can this be exported to Excel?” but the reality is, some folks want to play with the data outside of your dashboard. Meet them where they are.
Accelerating Data Democratization & Fostering a Data-Driven Culture
In a nutshell, data democratization is making data accessible and usable for all employees, regardless of technical skill. This enables informed decision-making throughout the organization.
To achieve data democratization, organizations must ensure data is easily accessible via user-friendly platforms and presented in an understandable, relevant manner. This involves breaking down silos and investing in tools that simplify data interpretation for all. Tools like Power BI, Tableau and Looker have made this easier in recent years, but providing training for these tools is crucial in ensuring everyone knows how to use them.
Secondly, fostering data literacy is crucial. This means equipping employees across all levels with the skills to effectively understand, interpret, and communicate data insights, enabling them to confidently use data in their roles. You’re teaching everyone to speak the same language. Create documentation that explains the rationale behind calculations, the reason why work was performed a certain way. This helps prevent “tribal knowledge” and improves data literacy.
Finally, robust data governance and security measures, such as clear policies and role-based access, must be intertwined with all of these efforts. This ensures data is used securely, ethically, and appropriately. Doing so builds the necessary trust to give more people access to data.
Conclusion
Disconnects between data teams and the business are incredibly common. It doesn’t represent a failure of your organization, it simply represents that your teams aren’t speaking the same language.
By improving communication and collaboration, designing analyses and reports around the stakeholders and enabling data democratization, your organization can increase data-driven decision making.