Project: Star+ Home Hero Design • 2021 • Role: UX Manager, Human-centered designer
Disney+ viewers that regularly (and quickly) find something to watch are more likely to stay subscribed at the time of renewal. One of the key design elements on the homepage that helps people discover new content is the large hero area. Prior to launch we spent numerous design cycles on getting this feature right. My team conducted multiple rounds of research collecting customer feedback and our insights drove many of the product decisions for the MVP. However, after launch, we learned that while people rely on the hero area to discover new content, it wasn’t driving as many video starts as we would have liked. One possible reason for this that people navigate down the page and miss the rotation of tiles on the carousel.
To address this behavior, we started iterating on the hero carousel and how we could expose the best content on Disney+. As designs were being finalized and engineering estimates were coming in, our leadership announced a new general entertainment streaming service called Star+. Many of our teams shifted focus from iterating on Disney+ to launch-mode for Star+.
Because this new service would include a broader range of content, including live sports and special events, the hero module would need to accommodate these different programming types. Fortunately, the team was already exploring concepts that could meet this challenge.
I was pulled in to help the team iterate through the late-stage designs through a process called RITE (Rapid Iterative Testing and Evaluation). A partnership between research, product design and prototyping, we ran three rounds of this iterative design approach to refine the new concepts and validate the interaction model.
A principal researcher on my team led the interviews with viewers while I facilitated alignment and debrief workshops with our teammates — designers and product managers. As a team, we all made design recommendations based on direct observations of users. This was a new process/methodology we were trying out for the first time at Disney Streaming, so getting buy-in was part of the up-front effort I led.
Rapid Iterative Testing & Evaluation
To set the context of our work, we often frame where we are in terms of the double diamond. This well-known framework enables the team to think in terms of research that helps ‘Make the Right Thing (1st diamond) vs research that helps ‘Make the Thing Right” (2nd diamond).
RITE is a method that emphasizes making design iterations as soon as a problem is observed; therefore, design questions need to be targeted. Incremental changes to strengthen visual affordances are the priority.
RITE has very specific rituals and processes that my research partner and I guided our stakeholders through, alternating between observation and design iteration. The objective is to reach consensus about a design problem using evidence, ideally, in a single day. The sequence of steps includes a planning session in the morning to decide the goals, followed immediately by 3 research sessions attended by the entire team. We then de-brief to analyze our observations and determine if the users encountered obstacles in the design.
If an issue is identified in the first round, the design can be updated before the second round. We then test with 3 more people and debrief again to see how any changes impacted the usability. At the end of the day, the team decides together what actions should be pursued.
The hero design we tested exposes 5 featured content tiles in the first row. As the user moves focus to the tile, the hero area animates to display the poster art of the selected content. This was an evolution from the the rotating hero carousel on Disney+ which only shows one title at a time. Our hypothesis was that by showing the user all five featured content tiles at once (rather than rotating through one-at-a-time), they would be more likely to discover something new.
Goal Setting and Research
The first step in RITE Testing is setting goals. In our morning session, we determined the questions we expected to answer by the end of the day. I then led an activity to help our observers recognize if a user was sucessful or unsuccesful in completing a task. For example, if we wanted to validate that users understood the relationship between the row of tiles and the large hero image, we expected to observe the participant navigating the row or commenting about the relationship. We then went straight to our first session.
Over the course of 6 sessions we noticed that while people understood how to use the new hero, they didn’t have enough information to make a decision for each of the tiles when they weren’t familiar with the title. The density and placement of the meta-data impacted both usability and engagement. In the first round, we observed that people didn’t have enough information in the hero to make a confident viewing decision and that they prefered to see more details.
In Round 2, we added a brief description as well as ratings, runtime, genre and the year made. A header identifying the content as “Star+ Highlights” was added to editorialize the first row as a featured row. However, we observed that users’ overlooked the label as their eyes were drawn down the page. We saw a disconnect between the title being displayed in the hero and the tiles in the featured first row. People didn’t understand why they were selected or grouped there.
After two rounds we decided to simplify the hero – similar to the first round, but adding an editorial badge to all 5 tiles, creating a unifying visual affordance.
The last iteration removed the long description so the team wanted to understand the impact of removing that text. Would users have enough information from the image and badge alone? We learned that while some viewers may miss the description, it didn’t create an obstacle to interacting with the hero or impede interest in viewing the content. In fact, the badging communicated a merchandising message that was engaging and users understood that they could get more information on a details page.
After 3 rounds of RITE the team’s decision was to proceed with this design.
This work launched in Q3 of 2021 so the final impact in terms of KPI’s like Minutes Watched or Time To Play is still unknown; however, as a design team, we feel confident in the usability and usefulness of this design element.
While some on the design team were reluctant to participate in RITE testing early on, perhaps due to the time commitment and group decision-making, this method proved to be an invaluable tool for iterating quickly to a design solution. By the end of the process, each member of the cross-functional team became an advocate for the approach. We have since used it on multiple features to quickly iterate our designs.