The current Google Drive Interface
For a course proejct, our group (consisting of Richmond Wong, Zach Porges, and Taylor Landis-Miller) decided to redesign Google Drive's interface. In order to do this, we first completed interviews with several users to find the most pertinent usability problems, which we then analyzed through heuristics. These problems are summarized below:
1. Files are hard to find because the user is not sure whether they are in “My Drive” or “Shared with me
2. Users with multiple accounts struggle to find files because they are not sure which account to check
3. Users who have folders within folders struggle to find what has been updated recently
4. Users rarely use the search bar; it is not well-integrated into the page
A more detailed report can be read below the images.
Figure 1: The default view of Drive will show a list of all documents (previously separated into “My Drive” and “Shared with me”), but users can filter through either the default labels or their custom labels. The default labels include labels that correspond to multiple Google accounts (here “Personal” and “Cornell”), “Starred,” “Docs I created,” “Docs Shared with me,” “Recent,” “Starred,” and an option to view the archive. In the prototype, other user-created labels also appear (“Web Design Project” and “Film Club”). The “Web Design Project” label icon indicates that it is shared with others. This solves the problem of not knowing which account files are assigned to, and not knowing what folders they are in. It also allows users to see all files at once, not hidden in folders.
In order to create a prototype that focused on the changes we made and the impact they have, we decided to do a low-fidelity pen-and-paper prototype. Our prototype needed to be efficient and easy to change, especially as we weighed different options from brainstorming. As an early-stage design, we want to allow space for easy edits following user testing. In order to create it, we referred to the current interfaces of Google Drive and Mail. Then we created a template that served as a background to different option selections, and finally chose a variety of selections that showed off our changes.
Envisioning our persona (which can be found below) interacting with the design, he may see that his “Web Design Project” label indicates that it is shared, easily allowing him to visually find web design files in the main screen (figure 1). Alternatively he could click on a label on the left column, and filter documents that way (figure 3). His documents are consolidated from his personal and Cornell Google accounts, indicated by a “Cornell” or “Personal” label, so he can see his film club files from both accounts at the same time (figure 3). The new placement of the search bar (figure 2) is more prominent, so that he is encouraged to use it when he cannot find a file.
Figure 2: The more prominent location of the search bar, more integrated into the spatial layout encourages users to use it when trying to find files. Tagging and searching makes recall and recognition of files easier to find than relying on folder hierarchies, improving ussability.
Figure 3: This shows what the interface looks like when a user clicks on a label (in this case, “Film Club”). Files can all be collected into one logical label, even if they belong to different Google accounts, or that were created by different owners, making the organization of files more closely resemble the user’s conceptual model of organizing information.
User Interviews and Data Collection
To construct a persona, each group member completed an interview with a Google Drive user, focusing on organizational features. Using contextual inquiries guidelines, we conducted our interviews with users by computers to be as close to the real use as possible. We asked them to show us their Google Drive accounts so we could get an idea of their motivations and strategies (Beyer & Holtzblatt, 1999). We asked broad questions that gave us insight into how users use Google Drive (such as “What do you use Google Drive for?”) as well as highlighting problems that users may have (such as “What sort of issues do you have?”). We also asked more specific questions as needed (see appendix for protocols). We each did one in-person interview, took notes, and recorded the audio of the interviews.
Broadly, we found that interviewees used Google Drive for uploading and backing up personal files, and collaborating and sharing with others for work or class projects. All had multiple Google accounts from which they used Google Drive. Problems they discussed related to not knowing where to find files, not remembering where a file’s location or if it was shared or not, and struggling to use Google Drive across multiple accounts.
Persona and Scenario
From our user interviews, we created a user persona.
James is a junior at Cornell University’s College of Arts and Sciences. He is computer literate; he surfs the web on Chrome using his MacBook Pro. He has two Google accounts that he uses frequently - a personal, and a school account.
James often collaborates with friends and classmates for the Film Club and his Intro Web Design Class using Google Drive. He occasionally uploads personal files too, such as his resume and a spreadsheet of potential summer jobs. A few Film Club files are organized in a folder, but most files are not categorized. For a group web design project, he created and shared a file of group tasks, but a different group member created and shared a file with the group’s design ideas. James finds Google Drive to be a powerful tool, but gets frustrated when he has to waste time finding files for his Web Design Class, as some files are in “My Drive” and some are in “Shared With Me.” Furthermore, he created a design rationale document for the project, but did so while logged into his personal Google account, causing him to have to change accounts in order to share it with his group members, which takes up time and makes him annoyed.
After brainstorming through a series of options and ideas, we discussed the tradeoffs, and then created a scenario based off of our persona and what we wanted the redesigned Google Drive to do.
After class, James decides to get some work done in Google Drive. First, he decides to start working on the User Testing protocol for his Intro Web Design Course, so he creates a document and assigns it the “Web Design Project” label so it is automatically shared with his team.
After getting some web design work done, he decides to work on a film script that he started last week. He doesn’t see it on the home screen, so he searches “film,” which returns all of the documents that either contain the word “Film” or have the “Film” or “Film club” label. The script is in his personal Google Account rather than his school Cmail account, but fortunately, he has a unified account, so it shows up. He clicks on it and opens his file.
Potential Design Evaluation Techniques
We would complete user testing on our prototype to see if it helps solve usability problems. If we were Google, we may change portions of the existing interface in code and run a/b testing on it immediately. But because of the low fidelity of our prototype, it is not ready to be tested in the field, so we would complete lab tests. Based on our persona and resources, our testers would be Cornell students who use Google Drive for both collaborative and personal purposes. Ideally, one member of our group would interact with the testers, giving tasks; one member would act as a computer to change interface data and screens as necessary; and one member would observe and take notes (in addition to video/audio recording of the test). We would use empirical analysis, and would likely ask users to complete tasks finding files in the system. We envision testers interacting with our prototype like other lab paper prototype user testing scenarios, although we might have to come up with more screens to handle other interactions. We would steer the testers to use certain search terms by asking them to do tasks that we are prepared for and have data for, but the answer we expect might not always be the one the user gives. We want to see if the learnability, efficiency, and satisfaction are improved in the new system. To see if we improved on the original design, we may take half our testers and have them complete the same tasks in the lab on the current Google Drive interface. We cannot compare time completion at the prototype’s fidelity, but we may compare the results based on user satisfaction, errors, and ability to complete tasks. At this stage, we are evaluating the higher level interaction and interface design, which is what our prototype filters in.
