Lauren's Portfolio

Appearance

Assignment 6: User Testing ​

Interview 1: Margaret Wang ​

Margaret seemed to understand how to use the interface pretty well. She instantly realized how to log in and create an account. When I asked her to follow the account of the US President, she first entered ‘potus’ into the search bar that filters posts by author. When that didn’t turn up anything, she scrolled through the list of posts under the default ‘All’ tab and clicked on the users’ names to follow them. An issue here I noticed is that the search bar is case sensitive, so even though she essentially knew the username to type in, the search didn’t turn anything up due to the fact that it was lowercase. I then asked her to add a context to the MrPresident account and upvote it, which she was able to quickly do, and then navigated away easily.

After that, she was asked to write a post with her thoughts on NPR’s article. She did not read the linked article, merely leaving a post stating that it was “so real.” She also did not click into the contexts of the NPR article to read their contents before posting. This behavior is rather typical and is in fact what my app was designed to combat. In the artificial data, the most upvoted (thus visible) context for the article was not very descriptive. Perhaps I could have made the dummy context more helpful, but I think that a hyperlink would be extremely useful in promoting the actual reading of posted articles. When I asked her about her experience using this feature in end of interview debriefing, she stated that she would have liked for an image or video preview to accompany the link.

I then asked her to unfollow all the users she had followed. She unfollowed NPR and POTUS from the “Following” tab, then navigated back to the “All” tab to unfollow PrincessDonut.

Finally, I asked her to log out. She navigated to the settings page and found the log out button with no interview, but then asked if she should delete her user instead. I figured that was probably fine, so she deleted her user.

After she logged out, I noticed that the context she wrote for MrPresident’s post now appeared as a blank box since the user deletion did not propagate to the upvotes. This is a bug.

Overall, she did not experience many issues and took less time than I expected to complete all the tasks.

Interview 2: Grace Yang ​

I first asked Grace to register for an account, which she was quickly able to do. When I asked her to follow the account of the US President, she scrolled through the posts lists and followed the ‘MrPresident’ account, which was a parody. She did not seem to notice the POTUS account. I then asked her to follow the PrincessDonut account. She spent several minutes trying different variations of the username in the search bar, including ‘princess donut,’ ‘Princess Donut,’ ‘princess doughnut,’ and then just ‘doughnut.’ No posts turned up, as not only was the name she typing in not a 100% match, but also because PrincessDonut had not made any posts, only written contexts. At one point she switched to the “Following” tab to search, which was empty as she did not have any users. Eventually she switched back to the “All” tab and scrolled through to find PrincessDonut, whose name she then clicked on to follow. Once she followed PrincesDonut and went back home, the feed automatically defaulted to “Following,” which she was confused by for about 15 seconds before she realized that the feed had switched. The third account she followed was the ABC account.

I then asked her to leave a context on MrPresident’s post and upvote it. She did leave a context easily, but then upvoted the autogenerated context, which I didn’t notice at the time.

After that, she was asked to write a post with her thoughts on NPR’s article. She did not click into the contexts, but she did read the article. She found the fact that she had to manually copy-paste the hyperlink into the search bar to be frustrating. I then asked her to unfollow all the users that she followed. She quickly unfollowed MrPresident and ABC, but temporarily forgot that she had followed PrincessDonut and thought that she had already unfollowed everybody. She eventually remembered, but it was evident from this that the lack of a “follow list” is an issue.

I then asked her to log out, which she did with no issues. She then asked if she should delete her user. Due to the bug I found during the interview with Margaret, I asked her to remove her upvotes and delete her context and post before deleting her user. When she tried to log back in, she accidentally used the register form instead of the log in.

Beyond the issues encountered while testing, during the debrief, she stated that she would have liked a user profile page for her own user and also a way to retrieve her password if she forgot it.

While reviewing the screen recording, I noticed that she kept using the back button on the browser to go back, which may be a slight navigational inconvenience.

Opportunities for Improvement ​

  • User Searching: users currently do not have a good way of searching for other users to follow, especially those that have not made any posts. They either have to scroll through the feed or search on post authors. This is a major conceptual flaw. A user search bar could be combined with the post search bar or just be separate, but there needs to be a way to search on users. Grace took several minutes to figure out how to follow PrincessDonut.
  • Log In vs Register: the log in and registration forms are currently adjacent within the Log In page and are visually nearly identical. This is a minor physical / linguistic flaw. I could potentially place the log in and registration forms on separate pages so that there's no confusion.
  • Email Verification: while this didn't pop up during the actual test, Grace noted that accounts were not linked to emails, meaning that there is no password retrieval and minimal security. I agree this is an issue, especially for an app that depends so heavily on trustworthiness of its users. This is a minor to major conceptual flaw in that it doesn't really add friction during regular usage, until it suddenly might become a serious problem.
  • Hyperlinking: It was frustrating for Grace to access links by copy pasting the text from the post into the browser search bar. Two ways this could be addressed are by turning links into hyperlinks automatically or allowing users to paste their urls in using a hyperlink insertion feature. This is a moderate conceptual flaw.