A deep dive into a single user research insight that had a big impact.
This is an in-depth look into one of the key user research insights that drove major design decisions during the redesign of MIT Technology Review.
Fore more context, read the full case study on the MIT Technology Review redesign.
How the reader reads
During the initial research phase of the redesign, my colleague Vanessa DeCollibus and I conducted an extensive research project where we surveyed around 2000 people and interviewed 19.
We found that when deciding what to read, people are interested in the topic, not the story type or format. In other words, it doesn’t matter if a story is classified as news, opinion, or photo essay as long as it covers something the reader is interested in. We heard and saw this confirmed over and over as people reported the behavior and we observed it when we asked them to find something to read.
This matters, because before the redesign, our entire site architecture was structured by story type. The primary navigation contained the following links (all story types!):
- News and Analysis
So—From the first scan of the page, we were making it harder for people to find a story they’d be interested in.
Uncovering the editorial perspective
We enlisted the excellent team at Upstatement to partner with us for the next phase of the redesign. We presented our research to them and had several deep discussions about our findings and what it could mean for the newly designed experience.
Fully informed and equipped, the next thing we did together was hold a series of stakeholder interviews with other departments to gather their input and learn about their points of view on the user experience. During one of the conversations, the editorial department revealed an illuminating insight:
Editorially, our content was categorized into six topics: Business, Biomedicine, Computing, Energy, Mobile, and Robotics. This much was well-known, at least internally, and was built into the site navigation (though buried a few layers deep and difficult to find).
But here’s what I didn’t know: Each topic had six themes which evolved a few times per year as markets shift and trends emerge. For example, the themes within the Robotics topic are practical robots, interfaces, automated cars, machine learning, robot breakthroughs, and tomorrow’s machines.
Together, the themes tell a story about what’s important to MIT Technology Review. They are signals for what matters in science and technology, where things are headed, and our editorial point of view. Yet they were completely invisible to readers.
We started to connect this new understanding of the nuance and intention within our editorial topics to the role that a story’s topic plays in how readers choose what to read… and as these things percolated, another problem came into focus, but this time it was with the publishing platform.
Before the redesign, we had a unique Drupal template with its own unique layout for each story type. The last time I counted there were over 25 templates.
We designed and built the Drupal system several years ago without anticipating the implications of having so many templates to maintain.
Intertwining design templates and story types created a structure that didn’t allow flexibility in how the editors could put a story together. When they wanted to make tweaks to a story that the templates didn’t allow, we’d write custom code to force the template to fit a specific scenario or create a new variation of the template that accommodated the request.
This kind of ad-hoc design and development wasn’t a very future-friendly practice and it created a bloated system that was difficult to maintain.
So with the opportunity to rebuild it better, we deconstructed the templates and broke them down into all their discreet parts.
Our new CMS would have a library of storytelling components that any article could use, no matter the story type or topic.
(By the way, we went with a headless Drupal setup with a custom publishing admin that controls the story layouts. Because the developers I work with are superb.)
We untangled the content and presentation, and suddenly there were many more possibilities for creating an engaging story.
With the content and design no longer beholden to story types behind the scenes, it made sense to tie this idea into how the user would discover content as well.
We removed the story type as an organizing principle in the information architecture and surfaced the topics and themes instead. Now, our taxonomy reflects the nuance and intention of the topics and themes, and allows users to more easily discover content they are interested in.
This one little research insight had a major impact, where the business goals, technical strategy, and user needs aligned perfectly. A UX dream.