Visualization Discovery for Dashboards that Stick
The Business Intelligence team at Norman Doors Inc gained access to the business expense application data. Recognizing the value this could provide the business, they set out to develop a data visualization. A developer developed a dashboard, iterating on feedback from his internal Business Intelligence team. However, once business users began testing, there were a lot of features missing and even some data sources that would need to be included for the dashboard to be useful to the business. The BI team would need to return to development and add the new requested features in order to deliver something with business value for Norman Doors.
This may sound like a familiar and frustrating scenario to analytics teams. With new tools like Tableau and Power BI making it easier to rapidly develop an analytics tool or dashboard, it is tempting to rush into development. But this can lead to wasteful design practices, unnecessary development iterations, and a final product that lacks business value. Had the BI team at Norman Doors sought to understand the needs of users before design and development, a final, finished, and useful product could have been produced when hands went to the keyboard. Without early engagement with users, it is all too common that dashboards are developed but don’t actually “stick.”
At Slalom, we use a methodology called the Product Engineering Methodology (PEM), focused on agility and the end-user, to assist in developing visualizations that add real business value. The first phase, called Discovery, is the framework for understanding business users and their needs.
GETTING TO KNOW THE USER
Discovery is arguably the most important step of any data visualization project, but the one that might be most tempting to skip over. Our Visual Analytics team recommends dedicating roughly a quarter to a third of the development initiative on the Discovery phase — and the benefits will be clear from the time investment.
Here are some guiding principles you can leverage in a discovery phase:
1. People: the most effective visualizations are developed with the end-user in mind. Identify anyone in the business who might be impacted by the roll-out of the new visualization. Conduct interviews and workshops to understand pain points and the current technology toolset. It takes practice and experience to ask the right questions during this process. By asking the right questions, you will gain an understanding of the underlying need business need.
2. Analysis: synthesize all the information collected from discussions with users to in order to integrate it into the design. Creating user personas and journeys are a great way to do this because they are concrete deliverables easily discussed and amended in a meeting for example. User personas are a UX concept where you create 3–5 fictional representations of your end users with needs, objectives, and roles and responsibilities. Here is an example of a user personal of a regional director:
Using the personas as a guide, develop user journeys to visually display a user’s workflow through the dashboard. User journeys are a map with steps users may take while interacting with your dashboard. What problems might users come to the dashboard with and what steps will the user take for the dashboard to deliver actionable insights? Will the user navigates to different dashboards or need to get additional information in an application outside the dashboard? We’ve found that user journeys help clarify the design of the dashboard and how the information will be used to make decisions. Including user journey design in your process simply yields a better dashboard.
3. Build a Wireframe: The end goal should be to build a high-fidelity wireframe, like the one below, though it can be good to start out with a looser wireframe which facilitates conversations around color, branding, data flow, and general layout. Get feedback and iterate on the wireframe with stakeholders. Wireframing a dashboard is one of the most crucial steps in the Discovery phase to ensure the dashboard will meet the needs of the persona and accommodate the user journey. Building a dashboard without a wireframe is like trying to build a house without a blueprint.
A successful Discovery phase accounts all three areas of design: Personas, Wireframes, and Use Cases. While dashboards can be developed leveraging these concepts individually, we find a much better result occurs when all 3 things are used in the development workflow for a dashboard.
Discovery takes more time upfront than the siloed development approach, but it enables the developer to understand the end user's needs. We have found that the investment yields longer-lasting, more impactful dashboards.
