Bringing our writers to the forefront of our search experience
We thought adding Atlantic authors to our existing results page would be simple. We were wrong.
“We’d like you to design a way to help users find Atlantic writers.”
This was the challenge proposed to me in August 2022 by our Executive Director of Product Carson Trobich. A few months earlier we had launched a custom-built search feature on The Atlantic website and app, and our Data Science team had found two interesting trends: of the top 100 search terms on our site, 30% were Atlantic writers; of the top 20 search terms, half of those were Atlantic writers. We had introduced search to help readers browse our site, but many of them were also using it to quickly find content from specific authors.
At the time, there wasn’t an optimal way to find all articles by a particular writer on the site. Profile pages exist for our writers and contributors, but those pages were only accessible by clicking an author byline on an article page or headline. Our data suggested that readers, unable to locate this information with ease, resorted to search. With this insight, we knew that it was time to bring our writers to the forefront of our search experience.
Defining the problem
We started the process by asking the following two questions:
1) Where should we surface writers on our site?
Should we highlight them in the main navigation or more prominently on our homepage? Instead of creating a new path that readers would have to learn to navigate, we decided to place writers where readers naturally gravitated–the existing search experience.
2) How should we surface these writers?
In a competitive landscape analysis of how other publishers presented writers in search tools, a common approach was to combine multiple media and content types—such as videos, photos, articles, and authors—into a unified experience. Some of the most common design patterns leveraged filters to distinguish between content types and to refine results.
After weighing the pros and cons of these approaches and thinking through possible user flows, it was time for some design exploration.
The design process
We came up with countless iterations, hoping to find one that would meet the needs of readers and the expectations of our editorial team. Defining the UX and content strategy was more challenging than designing the UI. Would we mix articles and writers together in one list or would that confuse and overwhelm readers? How much writer information would we display in the search results? Would we refer to them as writers, authors, or contributors when we use all three terms throughout the site?
There were many types of metadata to consider—writer names, headshots, and bios to name a few. We played around with tags and filters as ways to refine a search query. The Atlantic has been publishing since 1857, so providing a way for people to narrow authors by time period was top of mind. Additionally, in interviews with readers we heard that contextualizing authors through the topics they write about could be beneficial, especially for those in academia.
Given the breadth and history of our publication, data consistency has been a challenge for our organization. From an engineering perspective, the lack of uniform data on our authors was one of the greatest obstacles in this project. For some writers, we have high-quality headshots and full bios; for others, only a first and last name to work with. These inconsistencies made it difficult to land on a design that looked cohesive. To add more complexity, the design needed to be flexible enough to accommodate podcasts, newsletters, and other kinds of content we might want to add to the search experience down the road. It became clear that our site infrastructure would not be able to support advanced filters and topic tags for launch, because that would take a massive data clean-up effort.
Validating our ideas
After all these considerations, we arrived at a design that uses horizontal tabs to separate articles and authors. It is a common UI pattern across the web, but would people understand how to navigate this tabbed layout? Were the labels intuitive enough? Would readers be able to anticipate what they’d see in the author tab?
When there are questions about the clarity of a design, it’s best to conduct a usability test. Watching people interact in real time with our ideas is an invaluable way of identifying areas for improvement. Testing is a low-stakes way to measure the intuitiveness of interactions, navigation, hierarchy, and copy before time and resources are invested in building the design.
To test the design, we used a prototype—a mockup that mimics the live experience without full functionality. We created a prototype in Figma just interactive enough for people to be able to complete essential tasks, and we launched an unmoderated usability test on UserTesting.com.
To our team’s delight, the test validated our design approach. All ten testers clearly understood the purpose of the tabbed layout and completed tasks with ease. They appreciated the tabs, especially in the absence of numeric pagination and refinement tools like filters.
There were opportunities for improvement too. Although testers correctly predicted that the Author tab displayed Atlantic-published writers, confusion set in once they navigated there.
After completing our task (to search for an author named Hillary), testers were met with three people results, one of which was Hillary Rodham Clinton. The presence of a well-known public figure gave testers the impression that they were not seeing Atlantic writers, but instead people who are mentioned on the website. It turns out that many of our authors have also been subjects of our coverage, meaning they have both written for and been written about in The Atlantic. With public figures like Robert Frost, Helen Keller, and Frederick Douglass on our writer roll call, we knew that we would have to modify the design to add clarity.
We came up with a simple solution: a line of text above the search results that reads “Showing Atlantic contributors’’. Sometimes all that’s needed to improve usability is helper text or a small tweak in copy.
Looking for the best version of search
We launched the new version of the search experience this past March. Now, our readers can see authors clearly differentiated from other types of results. We are eager to incorporate more tools to help readers in their journey—like topics or time period filters, but that will require a deeper assessment of how we tag content and approach information architecture on the site.
Presenting authors in search has brought some new challenges for us to tackle as well. Now that our search properly highlights our writers, we have turned our attention to ways we might make our author pages more engaging and helpful.
When we build digital products, not all our ideas are going to make it to launch. In the spirit of shipping early and often, we must determine which features will deliver the greatest benefit with the time and resources we have available. We conduct research to understand readers’ needs, expectations and behaviors and gain insight into where to head next. For search, that will mean iterating and finding new ways to optimize our readers’ experience.
Acknowledgments: Aldana Vales, Carson Trobich, Jim Quindlen, and Lauren Olasov.
