- Discovery Studies
- Maps & Flows
- Sketches & Wireframes
- Usability Testing
Improving a Trader's morning routine through a trust-worthy dashboard
About the Project
Traders spend their day analyzing, optimizing, and reporting on advertising campaigns. To support this need, I designed a landing page that provides a quick overview into how their campaigns are delivering and performing. Traders can identify which campaigns need their attention and can take swift action to correct it. To accomplish this I:
- Conducted user research to learn what specific metrics were helpful to Traders when diagnosing issues.
- Pair designed with a designer to push the solution beyond our initial ideas.
- Ran usability tests to validate that notifications and alerts were triggered at appropriate times.
I co-owned the project with another designer. Together we conducted research, brainstormed and sketched ideas, put together wireframes and specs for development, and gathered feedback from stakeholders and users.
When a Trader logs into AppNexus’s Console, they are greeted with a list of all Advertisers that their company works with, a lot of stats for those Advertisers, and a way to navigate to campaigns for each one. While this may seem straight forward and simple, for a Trader who is anxious to check the pulse of campaigns they are responsible for, this view is not helpful. It does not show them the relevant information and requires them to navigate in and out of pages to finally determine statuses.
To improve this experience we first needed to understand what Traders needed so I, along with a designer, a product manager and a UX researcher, spent a couple of hours observing a Trader during their morning routine. Our focus was on how they immediately analyze, diagnose and take action on campaigns. We arrived at 9am and split up into two groups of two with distinctive roles; a notetaker and a facilitator who asks clarifying questions.
By the end of the 2 hour session, we observed the Trader doing the following things:
- They were interrupted by emails and requests to do other tasks throughout the morning.
- Reports were pulled to address discrepencies between Console and their Advertiser’s ad server's stats.
- adjusted daily budgets to make up for over- or under-spend.
- Targeting and bid strategies were tweaked to correct or improve performance.
We repeated this exercise with two more users at another company where we observed some differences and some similarities.
- They were more focused on completing each tasks before continuing on to the next.
- They had less interruptions throughout the morning.
- The first part of the day was focused on delivery and spend with the second half on performance objectives.
- Discrepencies were also addressed by pulling reports.
- daily spend was calculated in Excel and adjusted using the API to bulk update fields.
- Campaigns that weren’t performing well were deactivated and bid strategies were increased on campaigns that were.
After gathering enough information the Product Manager committed to two main goals for the project and we revised our existing Trader persona with the new information.
Goal 1As a Trader, I am quickly able to check which campaigns need my attention so that I can save time in my morning routine.
Goal 2As an agency or network, I can handle more orders per Trader so that I can increase efficiency.
Now that we learned what our user needs and defined product goals, I was ready to begin brainstorming how the UI should work. Mike, the other designer on the project, worked in San Francisco and I worked in New York. We logged into our video conferencing tool to pair design for the next couple of weeks (learn how we pair designed remotely). We explored different layouts and interaction with the goal of supporting many objects in a single view.
A grid of cards
Rows of cards
An inbox-style layout
We thought through a few different charts, possible metrics that could be useful, interactions that would be delightful to see when first logging in in the morning, and navigation to other areas of the product. After honing the design to something we were happy with we still had some questions we wanted to ask users.
We utilized the UX team’s research program Allies, to see if the design was usable, gague their initial reaction, and to see if the information provided helped them identify campaigns that needed attention. With facilitation help from a UX researcher, we showed users an image of the design and asked questions like, "What does this number signify?" and "Do you see any campaigns that are not delivering well?" and asked them to complete tasks as a part of a fictional scenario.
Colors were more distracting than helpfulThis contributed to an overwhelming sense at first glance. Users needed a little extra time to understand all of the data we were presenting but could identify which campaigns were delivering poorly because of the red indication.
Terminology used in labels were understoodUsers were able to site that 'Yesterday's Spend' was the most important stat to focus on. They were able to express the calculations behind some of the data points as expected and identified which cards were indicating poor delivery.
The goal line was understood but didn't seem necessaryUsers were able explain the purpose for the goal line on the graph and thought it was interesting. In practice though, they didn't anticipate using it.
We used the results from the tests to inform the next iteration where we removed the goal line, reduced the amount of colors to just red for delivery or performance issues, blue for flight issues and grey for all other cases. As we iterated and added features based on product requirements, we continued to validate and increased the fidelity and functionality of our mocks and prototypes. We validated animations, transitions, color and font hierarchies.
As the project got more fine-tuned, it became unnecessary to have two designers working on the same project. I ultimately rolled off and Mike became the sole leader on the project while I worked on other things and provided feedback to Mike.