Dashboards, Dashboards Everywhere, But I Don't Know What to Believe

Jeff Nieman is Senior Director of Data Strategy & Visualization at Best Buy and winner of the AI100 Award (MachineCon 2025). He has led data science and governance organizations at Cisco, McDonald’s, and Ford Motor Company, and also teaches data science as adjunct faculty at the City University of New York.

In an exclusive interview with CIOReview, he shared invaluable insights on how a small but focused Visualization Center of Excellence can transform data chaos into a culture of clarity, consistency, and business-aligned storytelling.

This is the challenge many companies face today. With the best of intentions, they invest in tools like Qlik, Tableau, or Power BI, empower teams to self-serve, and watch as well-meaning employees build thousands of dashboards across the enterprise. After all, “everyone can build a dashboard, right?”

The problem? These dashboards often contradict each other, are poorly constructed, rely on manual data feeds, and lack consistency. Instead of delivering clear, data-driven insights, leaders are left guessing which metric is correct and hoping the data refreshed on time that day.

So how do you fix this? Most companies don’t have the resources to centralize everything. But there is a better way. This article outlines a four-step approach to help your organization tell its story with data effectively, consistently, and accurately. The steps include creating a small central hub, building the right tools and processes, leveraging product teams for prioritization, and finally, enabling the broader organization.

1. Start with a Small Hub

Most people are familiar with the “hub and spoke” model. In analytics, this means a centralized team of analysts (the hub) supports embedded analysts or self-serve users in individual business units (the spokes). The challenge is often a lack of resources and the pressure for every team to self-serve.

I recommend a different path: build a small but nimble central hub called the Visualization Center of Excellence (COE). Its mission? To create top-tier visualizations for the company’s most important needs (more on how to prioritize in Step 3.) Focusing on the center allows us to define what excellent data storytelling looks like before scaling it outward.

2. Build the Tools and Processes for Excellence

Once you’ve established visualization COE, empower it to define what "excellence" means for your company. While there are plenty of great resources on best practices like those by Edward Tufte or Cole Nussbaumer Knaflic the challenge is tailoring them to your specific context. For example: How should your company’s fonts, colors, and logos be applied in dashboards? If your branding includes yellow or light grey, should you use those colors for line charts, even if they’re unreadable? (Answer: no!) What types of charts should be avoided altogether? (Pie charts, anyone?) Your experts make the call.

Your COE should define these standards in a style guide or playbook a living document that outlines what "great" looks like for your company. Mine is currently 62 pages and counting. This becomes your company’s single source of truth for data visualization and helps everyone “speak the same language.”

Next, build templates for your most common visualization tools. These templates should embed your fonts, color schemes, and layout conventions—KPIs at the top, filters on the side, and pre-set chart types. Templates not only speed up development but also ensure quality and consistency.

 ​Dashboards should tell one clear story, not a dozen conflicting ones, building excellence starts with a small hub, strong standards, and empowering teams to scale consistency  

Finally, establish a design review process. This is a space for peers to review each other’s dashboards, give feedback, and catch errors. When you've been staring at a dashboard for days, it’s easy to miss something obvious. Design reviews encourage thoughtful storytelling (not just “the business asked for it”) and foster a culture of feedback and continuous improvement.

3. Leverage Product Teams for Prioritization

Now that you’ve built a capable center with strong tools and processes, how do you decide what to build and when?

Our answer: adopt a product model. Inspired by thought leaders like Marty Cagan, this model assigns product managers to specific business domains. Their role is to deeply understand business needs and prioritize work accordingly.

If your company follows this approach, assign a COE member to each product team. The product manager can surface reporting needs and coordinate both back-end data work and front-end dashboard design. The designer begins with a wireframe based on the audience and use case, aligns with the product manager, and builds the final dashboard in parallel with data development. This tight alignment ensures dashboards are not only technically sound, but also relevant and timely. If your company doesn’t use a product model, the COE will need to prioritize requests themselves.

4. Enable Everyone

In the rush to build dashboards, many companies skip the first three steps. They default to a “Wild West” model of unrestricted self-serve. Analytics teams often justify this to focus on “more strategic priorities.” But the result is a mess: conflicting metrics, poor user experience, and unused dashboards cluttering your environment. I’m not saying to eliminate self-serve but to support it by sharing the tools and processes built in Step 2.

To increase awareness, We host regular guild meetings for anyone in the company with access to a visualization tool. Each session includes design highlights, template demos, solutions to common wireframing challenges (using our style guide), and an invitation to participate in design reviews.

Instead of tolerating dashboard chaos, we open the door for the entire company to benefit from the COE’s work. Adoption has been easy, who wouldn’t want a clean, ready-to-use template or a guide that answers common design questions? It saves time and boosts quality. This is how we realize the vision: build in the hub, with an eye toward empowering the spokes.

Conclusion

There is a path to taming the chaos of runaway data visualization. It starts by forming a small team of expert designers in a central hub. Give them the mandate and the tools to define excellence. Use a product model to prioritize what gets built. Then open those tools, processes, and standards to the entire enterprise. These four steps will help your company tell its story through data clearly, consistently, and effectively.