Inspirational Designs

Three Branching Ideas

We wanted to break out three different designs and make each have very distinctive interactions. We ideated a number of parameters on which the idea could differ and then selected five of these that we perceived as being core to the each associated experience. We also identified six other parameters which did not significantly branch the experience and design. The core ideas: how to match partners, how to select prompts, whether to show biographies or not, chat size, and how post-conversational reviews are handled.

Design Idea

Idea #1

Idea #2

Idea #3


Random w/Filter

Pick Topic

Pick Person

Focus on filtering certain traits

Focus on topics

Choosing individuals


Suggest topics then vote on which are interesting

Room creator sets title of room which becomes topic

No prompts by default, click to get prompts

The people are brought together mostly randomly and so having some voting on topics is helpful for good conversation

The creator has responsibility for choosing what the room will be talking about and getting people to join

We want natural progression of conversation, and only use prompts if/when they get stuck.


Brief read, decide whether to chat


Relatively heavy abstract/brief.  People can choose what information comes up in searches and have tags/bios/etc. (Potentially Tinder)

Since they are brought together randomly, getting background on the other person is helpful to start conversation off well

The emphasis of this interaction is on the topics that are picked.  The individual users are self-selecting based on topics and chat room specifics.

This idea lets you fully browse people before they choose to invest in a chat

Chat size




Random chats with random people are most likely

When picking a topic, there can be active discussions like a forum setting.

This idea is specifically targeted towards people who want private conversations.

Post-Talk Reviews




We don’t want people to be filtered out by ratings. Ratings would cause people who are naturally quiet or not good at facilitating conversations get stuck with other low rating members.

The focus of this idea is not on individual members, so more anonymity is appropriate.

We want to make sure users get reviews for accountability and to give feedback to other users.

Idea #1:


This design is based on random selection of chat partners.  Although a user can choose some basic filters (To be determined, but some possibilities are language, position, location, etc.), the final decision on who each user talks to is relatively randomized.  Each user writes a short autobiography that can fit into a small text box, and this autobiography is displayed to the partner while both sides choose favorable topics from a random group of topics as a start.  Once both users are done choosing, the conversation begins.

In the conversation, both users see a chat prompt they wanted to talk about.  They can use the prompt as a start if they wish, but are free to move to any topic they’d like.  If the conversation reaches a lull, then the users may request a new chat prompt.  Like any standard video chatting service, the interface also has a thumbnail picture of himself, a chat box for text, and the ability to toggle sound and video.  

What we like

  • Required user input is simple and straightforward. 
  • Finding chat partners is fast.  If there are a pretty good number of users online, then a random chat partner can be found almost immediately
  • Minimal biography displayed before the conversation means that users don’t spend too much time picking and choosing.
  • Worldly Winston would be able to make good use of the massively variable conversational topics that could be brought about by talking with people he’d never know otherwise.
  • Listener Leroy can explore the world vicariously.

What we don’t like

  • Scumbag Steve can fairly easily troll other users.
  • Driver Debby cannot choose her topic or ground she likes debating very easily. (Though this can be counteracted by her own assertiveness)
  • Double Listener Leroy’s are quite capable of being matched together.

Idea #2:


This design is based on creating video chat rooms for multiple users with the same interests meet together to discuss relevant topics. A user first chooses a subject that he is interested in. The user will enter a subject forum which contains video chat rooms with specific topics of the subject. Any member may start their own room and name their rooms more specifically to the topic they wish to talk about. Each room can contain up to five users.

For example, a member may choose history and then go into the room named “Social ramifications of the Marshall Plan” where up to four other people may be discussing this history. If the member does not find a room that he is interested in, he may also create his own room.

What we like

  • Group chat will allow more diverse views to be shared in a single conversation.
  • Focus is on topics, which will make it easier for users with similar interests to find each other.
  • Driver Debby would thrive in these chat rooms because their focus is on arguing and persuading others, so Driver Debby would gain a large audience with this idea.

What we don’t like

  • Not personal enough for one of our persona, Listener Leroy. He desires one on one conversations to get to know a person, but the group chat is more topic focused, with less personal one on one interactions.
  • The anonymity of the chat rooms encourages our anti-user, Scumbag Steve. It would be more difficult to filter out rooms that have inappropriate titles, and keep out unwanted users.

Idea #3:


This idea represents the extreme end of user control in selection. The user is able to browse the biographies for all people who are online, view their ratings and reviews, and then decide whether they would send them an invite. The person who receives the invite can then view the inviter’s bio and then either accept or reject. This idea is also extremely people focused and not prompt focused and so does not include an initial prompt for the conversation.

What we like

  • The users know exactly what they are getting into when they actually end up in a conversation
  • Scumbag Steves get very quickly marginalized by having bad reviews
  • Listener Leroy likes the people focused aspect of developing a stronger personal connection with people

What we don’t like

  • The lack of focus on prompts is non-ideal for Worldly Winston’s focus on learning and exploring new ideas
  • Through user discussions, we discovered that often when people invest heavily in a choice they are more likely to regret their choice and potentially simply leave the conversation.
  • It is potentially creepier and people may be reluctant to share their bio information knowing people can access it
  • Scumbag Steve can easily troll specific users and find them over and over again.

Synthesizing a Single Design

After thoroughly developing our three distinct ideas and discussion with potential users, we moved on to attempting to synthesize the positive aspects for each idea into a single cohesive design. In comparison to our previous three ideas, we pulled largely from #1, some from #2, and very little from #3. We created storyboards for this idea, but in the interested of not duplicating information, we lumped them in with the paper prototypes in this report.

        The user flow of our final design is to allow the user to choose which regions they would like to look for others to chat as well as what topic categories they are interested in. The region selection will be done with a map where users can indicate which regions they are interested in with a live indicator of how many people are looking for people in your location. This will allow the user to be able to quickly weigh the option of finding someone quickly versus finding someone from the exact area they are interested in.

After the selection has been made there will be a behind the scenes algorithm that starts to match you with appropriate other user. After the match has been made the user will be able to view a brief bio provided by the other person, see where the other user is located, as well as select 3 out of 5 provided prompts that they would enjoy discussing. This step will allow the users to familiarize themselves with one another and develop some context to aid them throughout their conversation.

After a bit of time the two users will be joined in video chat with one another. During the chat the primary focus will be on enabling positive and enjoyable conversation. The prompt will be displayed at the top for people’s reference. They will also have a chat box where they can chat via text if they’d like, as well as being able to share links to relevant articles and the like.

After the conversation ends, if the user violated the terms of service with their behavior and get reported their account will be warned and potentially banned. Their phone number / Facebook account will also be blocked from creating a new account. If they did not directly violate the terms of service but were exceptionally rude they can be thumbed down, which won’t immediately ban them but if someone gets a large amount of thumbs down they may be banned or only matched with other rude individuals. Lastly, a user can also indicate that they would like to friend the other person and if both users indicate this, they will be able to stay in touch and have future conversations. The design is intended to encourage friendly conversations while filtering our users of Scumbag Steve.

We did make a couple of compromises when choosing which aspects of our design to move forward with, and this does cause some potential shortcomings. One of these is that since the users cannot view one another’s biography before they are matched, they may read one another’s biography and decide they are not interested in chatting after all. Idealy, they would then simply amicably go their separate ways but their is the possibility that they may just leave right away. A second potential shortcoming is the lack of refined control on the exact prompt and location of discussion partner, potentially causing less valuable conversations. This was a compromise to allow for faster matching of users so they wouldn’t have to wait for the exact perfect match, which could take a very long time.

        We created paper prototypes based on the single idea that we are moving forward with. These prototypes are shown below.

Results of Testing

For user testing, we spoke to 9 users total, 7 users from the US and 2 users from Japan to gain insights. We figured that because our design will target users internationally, we figured that we should speak with some potential users that live outside of the United States first.

        After testing this idea with the two users in Japan, we found that they would be most interested in using this site to meet English speaking foreigners to improve their English without cost. They said they usually pay foreigners to speak with them to improve their English, but this platform would allow them to speak to native foreigners without payment or scheduling. One of our users also noted that he is not fluent enough in English to be able to carry on a deep conversation about the various topics we showed him. They both emphasized that many people would be very interested in a website like this that allowed them to speak to other English speaking foreigners to practice their English.

        These user testing sessions made us realize that users from non English speaking countries will be using this platform heavily to improve their English. We initially wanted to make sure we are not designing for them, but there seems to be a great enough demand for a place to improve their English in Japan, and most likely in other non English speaking countries, to warrant blocking them outright. While we will not emphasize language improvement, If language improvement is their motive, they should be allowed to use the platform. Also, we need a way to specify the language that will be spoken. We did not consider how to resolve language barriers before doing these interviews.

We gained several insights related to the region / topic selection screen from our users. One piece of feedback was the map should start out unselected and rather than selecting countries we should consider other systems, like drawing circles on the map that represent where users are interested in finding discussion partners.

We also looked specifically at the prompts for selecting a topic and region to discuss with. Users who were interviewed felt that if prompts such as “Select a region to talk to” were better if phrased as questions. These declarative prompts felt too forced and official. Since this application is about communicating and questions, users felt makes sense to extend this  curious feel to our prompt displays. As a result we will phrase these prompts in a manner similar to, “Where would you like to talk to?”