Design Refinement

We have designed a video chatting platform that allows users to talk to strangers online from different countries. The problem with the existing video chat platforms is that there are many men who only go on these sites to meet girls and engage in inappropriate behavior. Our video chat platform is built from the bottom up to discourage those kinds of behavior and promote good conversation between people. This includes having a high barrier to entry and easy report, which will lessen the anonymity of these accounts.

        This chatting platform works by first making the users log in through their Facebook account. The user will then select countries and themes to filter where the conversation partner will be from and what the chat will be about. Next, the user will select from a list of prompts to indicate what they would be most interested in discussing with the stranger. Finally, the user will be paired with a stranger and begin the chat. When the user exits the chat, the user will be taken to a page that allows them to give feedback on their chat, friend their partner, or find a new partner to start chatting with. The user can also report their conversation partner if the other user violated the terms of use of this website. We also allow the user to down-vote their partner when the partner does not violate the terms of use, but was unpleasant to talk to in any way. The idea is that the down-vote will be notified to the maintenance staff, and if the user is down-voted by multiple other users, that user will be matched with other similarly down-voted users. We may also send users with low ratings a warning or a suggestion email explaining that something may be wrong with the way they are interacting with other users.

        Overall, from paper to demo, there were not too many changes because the system turned out to be relatively simple.  However, the users’ major changes and feedback were focused on the map and geographical selection of the interaction.  We also determined the overall color scheme and realized that certain required interaction screens needed to be made before we could consider the interaction fully prototyped.

The initial reaction revealed that the users enjoyed the framework for selecting countries.  However, one issue that was brought up was the strong possibility that a user does not know the exact location of a country, thereby making the map selection not useful.  From this feedback, we implemented a search bar where users can search for the countries they may be interested in.  Though we acknowledge that this feature requires the user to be able to spell the country’s name correctly.  

        Another major feature of the map feedback involves the visibility of the users looking for people in your location.  Due to the nature of geographic selection, an common issue would be the inability to find matched users due to an insufficient number of people online.  To counter this, we remade the map to include a heatmap indicating active countries.  A natural mapping between color saturation and number of users should be sufficient to give the user an overview of other users.  

While prototyping we found that a blue overall color scheme indicated user density very well while at the same time being kinder on the eyes of the users.  This color neither alerted the user (Like a red or yellow), nor dropped into the background (Like a gray).  Although red would have probably indicated more strongly, the connotations and interactions with red would be too far in the danger zone.  Similarly, gray would have worked had gray not been a culturally accepted color of non-interactivity.

Our last major change was the inclusion of new pages required for the interaction to be considered full.  Though we had the capability to add a friend after a call, there were no friend pages.  Though we wanted to show a user’s bio, there was no profile screen.  We had to scale our project a little larger in order to account for these new pages.

Initially in our first conversations with users we learned that context was quite important to establishing deep conversations with people. We assumed that providing a biography would add enough context. However, we learned that often only adds superficial details to the conversation. To enable truly deeper conversations we need to create a friend’s list that allows people to reconnect with partners after their conversation ends. This will create conversations that go deeper by getting additional context through repeated interaction.

        By testing our paper prototypes with potential users we gained a number of insights. One of these insights was how users would select which countries to find a partner in. Initially, we were going to simply provide a map and have users click on the countries they were interested. However, we quickly discovered that users generally were not always aware which countries were which. We fixed this issue by listing the names of selected countries and allowing users to add countries by entering the countries name in a text field which gives intelligent autocomplete suggestions.

        A second insight is that while we had always assumed there would be some form of a profile screen, it was not until we developed our paper prototypes that we realized we hadn’t actually given much thought to the design of this page. After this realization, we were able to quickly develop a basic profile screen in our more complete prototype. This profile allows the user to input the information about themselves that is given when they are about to start a conversation.

        The biggest question we have going into the final stage of this project is how our product would appeal to an international audience. While we did get some feedback from two users in Japan everyone else we have interviewed has been from the US. We need to make this appeal to people across all countries in order to be successful, as such it is important we get feedback from a more diverse, international audience.

We are also aware that we do not have extensive documentation available for our users on how to use our chatting platform. If this idea were to be pursued further, we would add a tutorial session or a help doc that guides the user through our platform.

We also got feedback from our users in a run-through of the program that it was unclear to them how one would be notified if someone was trying to talk to them and what that experience would be like using the application. They said that this piece of the experience was missing in our demo, so we will need to consider this piece when moving forward.

It was also suggested that we have a way of giving a thumbs up rating that allows users to give positive ratings to good conversation experiences. This could help us identify good conversation partners that can be matched with other good conversation partners to allow them to have positive conversation experiences.