Enterprise Dashboards for Supply Chain

Role: UX Designer
Project Length: 8 months
Type: Enterprise dashboards for Supply Chain leadership, leveraging Microsoft Power BI.

Project Background

I designed enterprise dashboards for Supply Chain leadership, who use these reports to analyze operational data on 53 Distribution Centers across the United States. These Distribution Centers supply auto parts to 6,000+ NAPA stores, other warehouses, and some commercial customers.
‍

What the inside of a typical GPC Distribution Center (or warehouse) looks like.

Notable Challenges: Lack of established design standards for reports, no streamlined workflow for gathering/validating report requirements and designs, navigating constraints within Power BI as a third party tool.

I also established design standards and re-usable components that developers can use to build reports for other functions, such as Commercial Sales, HR, Pricing, Merchandising, Finance, and Store Operations.

These reports are displayed in the enterprise business intelligence tool, Power BI (by Microsoft) and are updated daily with data from a cloud data warehouse (Google Big Query), which is part of the modern data and analytics strategy of Genuine Parts Company

Research Process

Unique Highlights: Primary users of these reports are internal employees, ranging from General Managers to Supply Chain Directors and Executives. This made it easier to schedule time for user interviews, but also limited the number of actual users.

Since each report has varying users (usually a Sr. Director or VP of Supply Chain) I worked closely with Product Managers and UX Researchers to identify requirements through Discovery Research. This process began from when the required data was extracted, prepared, and analyzed. I also involved Development early on to ideate, identify technical feasibility, and understand limitations.

Sometimes, there were existing reports in the legacy BI application (Qlik) that we could reference.

Research focused on understanding what is important to see within each metric, and how to surface this data while allowing for deeper analysis. Discovery Research was conducted through interviews, as users were responsible for different areas of warehouse operations such as Safety or Performance. Later in the project, we planned to implement surveys for NPS scores and additional feedback.

Challenges & Workshopping

Initially, there were no established report design standards or style guides with each Power BI Developer designing reports as they saw fit. Besides lacking consistency, reports were not accessible and report creation was the "Wild West" with lots of back and forth on what "looked" right or not. This led to a lot of repeat work, as each developer had their own way of creating reports.

With the Lead Power BI Developer for Supply Chain, we established documentation for other developers to reference along with style guides and examples based on learnings from Data Visualization books and courses.
For example, I drew heavy inspiration from the Okabe-Ito color palette to use for our enterprise reports, which is important for people with color vision deficiencies as color is often used to compare and differentiate data points.

A main point was Accessibility, and making sure that all reports accommodated for a wide variety of users. I also conducted a team workshop to identify common dashboard use cases, and worked with Developers to create standardized JSON components within Power BI. Basically re-usable report building blocks.

Examples

Below are several examples of the reports that I designed for various Supply Chain metrics. You can see what the reports looked like before (ad-hoc) and after the enterprise design standards were applied. The goal was to present the data in a clear, and easy to understand way while allowing for further exploration and analysis. Actual data has been redacted.
‍

Screenshots and notes from existing reports helped create quick concepts that could be validated.

Sample page from an ad-hoc report, without enterprise design standards applied.

The same report, with enterprise design standards applied.

Sample page from an ad-hoc report, without enterprise design standards applied.

The same report, with enterprise design standards applied.

Re-usable template pages that covered standard Power BI report use cases.

I also conducted heuristic analyses of dashboard projects for other initiatives, and provided recommendations based on the enterprise standards that we established. These analyses combined direct changes in Power BI, as well as call outs based on Nielsen Norman Usability Heuristics.

Reflection

With a streamlined workflow and the team becoming accustomed to building reports with enterprise design standards, we noticed faster project completions with more visually consistent data visualizations that met Web Content Accessibility Guidelines.

Feedback from the Business was positive, stating that these reports helped reduce the manual work spent analyzing data in Excel spreadsheets, and sourcing data from various databases and email threads. Since this data comes directly from the source, there was increased confidence in data accuracy and validity - important in decision making.

Future work includes creating more standardized JSON files that developers can re-use while building Power BI reports, enhancing documentation and Power BI training resources, and eventually replacing all reports in the legacy business intelligence application (Qlik) as all the company functions adopt a modernized data platform.

I also had the opportunity to share my learnings about Data Visualization with the UX Design & Research Team, which helped solidify my own understanding and prompt further discussions and questions. Resources included the following books and courses.

Good Charts: The HBR Guide to Making Smarter, More Persuasive Data Visualizations
The Big Book of Dashboards: Visualizing Your Data Using Real-World Business Scenarios
Fundamentals of Data Visualization: A Primer on Making Informative and Compelling Figures
Information Visualization Course | IxDF (Interaction Design Foundation)
Analyzing/Presenting Data/Information: An online course taught by Edward Tufte