At Google Play, I saw as a high impact, low-ish hanging fruit worth developing. So I went ahead and conducted research, created prototypes, ran tests, and pitched the idea to bring a team together to build it. This is the story of ‘Eastwood 2.0’
At Google Play, I saw as a high impact, low-ish hanging fruit worth developing. So I went ahead and conducted research, created prototypes, ran tests, and pitched the idea to bring a team together to build it. This is the story of ‘Eastwood 2.0’
When watching movies on the Play Movies app, users could pause the film, then tap on actors to see their data right on screen - their metadata, other films they’ve been in, latest news, etc. In order for these actors to be tappable, an internal tool called Eastwood was use to tag actors manually.
While observing the usage of the Eastwood tool, I noticed that although it served its basic function of tagging actors in the Play Movies app, it lacked intuitive design and efficiency. Users had to navigate through a cluttered interface, making manual tagging a time-consuming and cumbersome process.
Recognizing the potential for improvement, I engaged with the vendor team (primary users of the tool) to gain insights into their workflow and challenges. By observing their usage patterns and attempting to use the tool myself, I uncovered several pain points and opportunities for enhancement. This hands-on approach allowed me to empathize with the users' experiences and informed my strategy for proposing enhancements to the tool.
In the existing UI of the tool, users faced challenges with the suggested actors feature and associated key bindings, which often led to a slower and less efficient tagging process. To address these issues, several key improvements were implemented, but I will just focus on two aspects for this portfolio overview: UI arrangement and cluster suggestions.
For my first step, I rearranged the existing tool's placements. These changes were mostly based on intuition after getting familiar with the tool. Two key points to note in the screenshot above are the suggested actors being moved to the right and having their associated numbers mirror the layout of the numpad on the keyboard. Additionally, the current tagging cluster was placed under the main movie screen.
The cluster suggestions are a collection of frames of an actor within a movie. This was moved from the upper left corner to a more prominent location below the movie screen. This adjustment aimed to enhance visibility and make it easier for users to navigate and tag actors.
Furthermore, the organization of previously tagged frames to the left and upcoming clusters to the right was introduced. This allowed users to quickly scan for inaccuracies in previous tags or prepare to tag upcoming actors efficiently.
To minimize disruption and enhance user familiarity, the frequency of changes to the key assignments for suggested actors was reduced to zero on numbers 1-6 (customizable), and kept suggestions on 7, 8, and 9. This change aimed to help users develop a sense of proficiency in tagging the main actors, contributing to a smoother tagging workflow.
These improvements were designed to address the cumbersome aspects of the existing UI and streamline the tagging process, ultimately improving the user experience and efficiency of the tool.
Recognizing the importance of user comfort and usability, I opted for a darker interface design in response to feedback from users who experienced eye strain from the brightness of the previous all-white version. Considering that users often spend more than seven hours a day interacting with the Eastwood tool, prioritizing their comfort and reducing eye strain was important for both usability and efficiency.
Additionally, the darker aesthetic not only addressed usability concerns but also aligned with popular movie editing software of the time, which felt appropriate. That... and I thought it looked cooler anyway.
After refining the prototype based on valuable feedback from both personal testing and the vendor team, I was confident in the viability of the project and ready to present it to stakeholders for official approval.
Before pitching the project to stakeholders for official approval and resource allocation, I conducted thorough analysis to ensure the numbers were substantial and justified the investment of engineering resources. I felt good that my evaluation process ensured that the project was practical and worthwhile to pursue.
Once my final prototype was completed, I put together a deck and pitched it to stakeholders, and was given the green light.
Eastwood 2.0 was developed and ready for live use in less than three months. Since its launch, it has demonstrated significant improvements, including a 54% increase in tagging efficiency within the first week, approximately 30% fewer errors in actor-tagging, and new users achieving proficiency in less than two weeks!
In addition to these quantitative metrics, qualitative feedback highlighted the improved user experience, with users finding the tool more pleasant to use and easier on the eyes.
Although the project's appearance may now seem dated, at the time, I aimed for a design reminiscent of professional movie editing software like Final Cut Pro or Adobe Premiere. Reflecting on my time at Google, I’m most proud of spearheading the Eastwood 2.0 project.
Despite this project being grizzled and ancient by the time of this reading, it provided invaluable experience across all design stages early in my career.
