Improving our data visualization design skills

Ashley Halsey Hemingway
Jun 16 · 5 min read

Learning as a team in a fully remote environment

Several sketches of data visualizations created by the team

Evolving user expectations

As the leader of a team that designs ad tech interfaces, it has been interesting to see changes in our user base over the years. Early ad tech users were very technical — sometimes they wouldn’t even use our UI, preferring instead to leverage the API or even build a custom UI of their own. Over time, advertisers and publishers have become more hands-on and directly involved with running their advertising business. As a result, there has been an increased need for our products to work well out-of-the-box and be intuitive.

Where early users often wanted their data in a raw format, or perhaps in a table, today’s users increasingly want to be able to understand their data at-a-glance. Our Xandr product UIs now utilize data visualizations more frequently to help users view their data in new and insightful ways.

A changing team environment

While some amount of our team collaboration has always been done remotely, 2020 was the first year where everyone in our group was fully remote (due to the pandemic). Product design, especially for enterprise products, is often detailed, heads-down work that benefits from quiet time at our desks. We are often solving challenging problems like creating an object architecture, documenting a detailed workflow, or prototyping a new design pattern. However, design is also a creative exercise — our design solutions greatly benefit from collaborating with others and ideating new, innovative solutions. We find that making space for both individual and collaborative work leads to the strongest design outcomes.

At our mid-year team retrospective, the whole Xandr Product Design Team agreed we missed that “grab a whiteboard” creative energy that’s easier to find when co-located. We also discussed which skills were becoming even more important in our day-to-day work. So, we decided to build up a skill that would improve the quality of our design work, while also providing the opportunity to collaborate with each other more closely.

As a result, we committed to spending six months improving our data visualization design skills as a team.

Creating a learning plan

Xandr’s Product Design team contains a “Design Ops” group – a subset of team members who are responsible for providing and maintaining the systems, tools, and processes that enable the wider team to do their work excellently, efficiently, and consistently. This group took on the challenge of translating our team’s goal into an actionable plan.

The first challenge was determining a reasonable time commitment that everyone on the team could actually fit into their schedules. Our work calendars had already gotten very tight due to many previously in-person chats morphing into scheduled meetings (as the wider company adapted to remote work). We are also a fairly small team supporting a wide range of products, so we needed to be extremely efficient with our time. Together, we decided to dedicate 15 of our weekly one hour team meetings to this topic.

The second challenge was finding an educational resource that was appropriate for our less-experienced team members, while still valuable and engaging for those of us more acquainted with data visualization design. After reviewing many different content sources (podcasts, websites, workshops, etc.) we chose a Coursera class as our guide — specifically Information Visualization: Advanced Techniques from New York University.

The course was arranged into four units:

  • Geographical Data – data associated with a location
  • Network Data – data associated with a relationship (parent/child, etc.)
  • Temporal Data – data associated with a time/date
  • Interactive Data Visualizations — exploring how the ability to interact with the visualization can extend its capabilities

We then broke each unit into a three week module format:

  • Week 1: Independent Study — watching / reading course materials
  • Week 2: Workshop — working on a design problem in small groups
  • Week 3: Discussion — sharing our learnings and discussing potential applications to current and future design work

Members of the Engineering and Product Management teams were also welcomed to participate.

The workshops we did as part of week two provided an opportunity to collaborate with team members we might not typically work with. We discovered new ways of sketching and sharing ideas remotely — from using tablets to taking photos of pencil drawings or sending images over Slack. Getting to build on each other’s ideas and incorporate them into our group “homework” assignments provided a fun and creative opportunity to work together and solve a shared design problem.

A breakdown of our course structure

Applying what we learned

It was interesting to analyze data visualization design from a theoretical perspective and one that was not specifically tied to ad tech. Ultimately, our goal was to learn something that we could then apply to our daily work. So, we decided to close out the year with a “Xandr Capstone” project: each designer could work independently or in one or more groups to create a new data visualization or improve an existing one within a Xandr product.

In the end, our product designers created a wide range of data visualization projects, including:

  • Improving a dashboard for a new metric — Ease Of Use
  • Applying learnings from an online shopping/browsing experience to our own inventory library feature
  • Using Sankey diagrams to help users better understand the impact of their “ad quality” settings on how many bids they see
  • Helping users better understand the effect that their edits have on the performance of a campaign
  • Transforming a PDF report that our TV teams use to populate an interactive, auto-generated dashboard
  • Improving the visual representation of a forecasting tool
  • Using interactive maps to give users a graphic interface for selecting geographical targeting

Reflections and next steps

When I look back on the investment we made in this team-learning goal, I feel that our time was well spent. I see the collective knowledge we gained being widely applied by the team in actual projects. As we’ve introduced new data visualizations to our products, the feedback we’ve received from users indicates that our enhancements are helpful, well designed, and exciting. Users are finding new insights in their data that data tables didn’t easily provide.

As the leader of the product design team, I now see increased collaboration and deeper creative relationships between designers. We have a new shared lexicon for giving design feedback about data visualization and when we Retro’ed the goal, team members were appreciative that Xandr’s leadership supported their professional growth.

Learning as a team was so well received that we decided to spend the first half of 2021 improving our user research skills together. As a team leader, I’ve seen firsthand that investing in team learning goals can improve the technical skills of the team while also providing a much needed opportunity for teamwork, individual growth, and creativity.

Xandr-Tech

Our latest thoughts, challenges, triumphs, try-again’s, most snarky and profound commit messages. Our proudest achievements, deepest darkest technical debt regrets (just kidding, maybe). All the humbling yet informative things you learn when you try to do things with computers.

Ashley Halsey Hemingway

Written by

Leading Product Design @ Xandr

Xandr-Tech

Our latest thoughts, challenges, triumphs, try-again’s, most snarky and profound commit messages. Our proudest achievements, deepest darkest technical debt regrets (just kidding, maybe). All the humbling yet informative things you learn when you try to do things with computers.