Where did I start?
Brainstorming (Lean UX)
The project was kicked off by facilitating a Lean UX exercise. This was an entirely new product for Paylocity, so the ability to work quickly while incorporating constant feedback from multiple channels demanded a more unique perspective, such as Lean UX, rather than the older waterfall model, which wasn't fluid enough. Establishing business outcomes, proto-personas, and hypotheses is a great way to guide the business forward when there isn't a lot of top-down feedback to start with.
User Research
It was important to create proto-personas during our Lean UX session so we could start answering questions around 'why' clients want retention data and 'how' they would best consume it. Presenting data like this early on to stakeholders is important to get the go-ahead for reaching out to real clients.
Once we had buy-in from the business owners, we quickly set up moderated sessions with some of our highest revenue-generating clients to assess the accuracy of our assumptions and learn from where they were not. The moderated sessions consisted of 1-on-1s and group sessions via Zoom.
Wireframes
After accumulating an enormous amount of retention-specific data such as location, department, manager, tenure, drive time, and salary, early wireframing began using Balsamiq. I created a dozen low-fidelity wireframes to test the effectiveness of content and data visualizations depending on where they were placed on the page. We heard time and time again that high-level statistics held the highest value for our clients, so adopting a dashboard panel view across the top of the page became the consistent approach moving forward.
High fidelity designs
As noted from the wireframing work, high-level statistics were what clients valued the most. So when it came to the final designs, I reduced the number of dashboard panels from six or more in the wireframes down to a maximum of four. It was important to limit the number of dashboard panels since the factors involved in creating those panels could be quite complex. For example, in the design above, our clients are seeing 8 At-Risk Employees, where the Median Total Compensation of those employees is $45,000.00, their Median Tenure is 5.3 years, and the Median Drive Time is 25 minutes.
What about when a client wants to drill down further into a specific employee's retention data? Since the new data visualization views were so well received during our research efforts, I wanted to integrate those visuals into what our clients are most accustomed to seeing, which are lists of their employees. Introducing an expandable/collapsible accordion control to the list view was a great way to keep the default screen clean yet robust.
As I continued requesting feedback from internal and external users, I quickly discovered that going straight from high-level statistics down to the employee level wasn't the only thing users wanted. An admin or superuser, who oversees more than just a few direct reports, requested to see retention data at the department level. The design above is an example of how I proposed showing group data that could be sorted in ascending or descending order and specified down to the cost center level.
