Twitch Case Study — “Twitchlets”
As the premiere live-stream platform, Twitch has grown significantly over the past several years. A platform built for “community-first” thinking, Twitch has gone from just a gaming site to a site for all communities to have a space to hang out with their favorite streaming personalities.
How might we get users to watch more on Twitch.tv?
With Twitch, the metric primarily used is “watch time.” The main competition for Twitch isn’t just the likes of Youtube, Facebook, and smaller live-streaming platforms. It’s competing against the likes of HBO, Netflix, Hulu, etc. They want users to not only come to the site, but they want them to stay on the site as long as possible. If there is nothing engaging about the service, users will surf the web to find the next hook catching them. Twitch needs an edge over its competitors, and for most features Twitch is. Clipping live-streams for highlights, cheering and involving the community with crowd-sourcing content, events and more are just a few of the things which Twitch was either first to the board or has done it better than the rest. However with Youtube and Facebook still climbing steadily behind Twitch (especially Youtube at this point) Twitch needs to create a system that can attract new users and engage veterans of the site to maintain their top spot.
Target Users
Twitch is home for the “younger” generation. It wants to be inclusive to all walks of life, but the site was heavily cultivated by those who grew up during the exponential growth of the Internet. Twitch is also home to many modern Internet jokes, modern media, and alternatives to traditional media like television. This is not to say Twitch is “for kids.” This is seen in their Terms of Service, where users who want to stream must be at least 13 years of age. A lot of toxicity has ebbed-and-flowed through the service and Twitch has been adamant on creating an area of little hostility and instead have a place where users can be expressive as they want to be without fear of negative judgement.
Limitations on research
Being a student working on this case studies didn’t come without restraints. I was learning on how to develop a study and create a prototype while also building this case for Twitch. This meant less time on refining the quality of work and more on making sure timelines were met. Also, being the only individual on the team to work on the study meant I had to spread myself thin to hit all marks. I didn’t get to interview and prototype with as many users as I wanted to, including more avid viewers on the platform. The 2020 pandemic also clamped down on my timeline even more as I had to accomplish more in less time. This has been one of the more difficult projects I’ve ever worked on.
Lo-Fi Concept
Going into the study, and into building the prototype, I had a general idea on what I wanted to accomplish. Twitch needed a service native to their site which would a higher-level of interaction between the stream and the user participating. I emulated Mixer’s “MixPlay” feature for the project because Mixer was one of the higher-standing competitors against Twitch until its collapse in July 2020. Though Mixer is no longer around, a lot of its features and services have since been implemented by Twitch, Youtube, Facebook, and more including their technology to develop a near-zero delay between the live-stream and users. This was imperative for MixPlay to work as MixPlay was developed as a tool for users to interact with a stream with a simple press of a button. I have taken that idea to Twitch with “Twitchlets.” Twitchlets are a Twitch-applet system where users can pay Channel Points or Bits (monetary value) to affect the gameplay of streamers. Such mechanics would include disabling crouching in a shooter, invert steering controls in racing, drops of loot in an RPG, etc.
The most basic form would have a variety of options shown for users to browse during a streamer’s broadcast. In the example AnneMunition is playing Escape From Tarkov, a survival first-person shooter. The options shown are “Invert Aim,” “Disable Crouch,” and “Play Sound.” There more options hidden that can be shown by expanding the Twitchlets area, but the highlighted options are place at the top and are never hidden. After making a selection, a confirmation window would appear for the user to double-check if this is the selection they wanted. This window would included the action selected and the amount of Bits or Channel Points a user would spend to perform the action. After confirm the selection, the action would be performed on the stream for a fixed amount of time. “Invert Aim,” for instance, could be set for two minutes (at the streamer’s discretion). This would also be a feature not every user or streamer would be required to participate in and could disable it through their respective settings.
Users always want to be involved in some capacity with the community and the stream they are watching. Currently Twitch has crowd-sourcing events, “cheering” with bits, gifting subscriptions, viewers clipping the best moments from a stream, and much more. However, this has a slight disconnect as they are not directly affecting the streamer or the stream they are watching. With “Twitchlets,” this system will bridge the gap to almost nothing between users and a stream. Users will be able to help or hinder games on a real-time basis native to Twitch without the need to use third-party systems and setups.
Ideation and Creation
Now that I have the basic idea of what I want to accomplish with Twitchlets, it was time to work on the steps needed to high-level optimization. The .5 version of the prototype was shown to a group of and I was given a wealth of information on how to remove unnecessary steps and clutter so users would could get to seeing their action selected with less hassle. This included removing the confirmation window for a timer system on the bottom of the screen that is large enough to see and undo if the user does not want the action selected. Also, I had the chance to talk to some volunteers and discuss what types of improvements can be given to the prototype and aggregated these answers to have more differentiation for “Twitchlets” versus the rest of the site. Both my group of peers and volunteers did note the process to perform actions were straightforward and easy to follow as well.
Following the feedback process, I have since iterated and developed a prototype which follows what was needed to change to become an effective addition to Twitch. The buttons are now shaped in contrast to the rest of the site’s layout, making it easy to find on the screen. Gone is the confirmation screen in favor of a progress bar that fills after a fixed amount of time has passed. When a user presses the “undo” function or if the action can’t be processed, the bar will fail to fill and a “red X” will appear. No breaks from the action, straightforward direction, and easy accessibility will make this project an effective tool which Twitch can use to boost their numbers and move ahead of their competition in the coming years.
Ending Thoughts and Moving Forward
Obviously this won’t be a “patch in and solve everything” kind of project. There will be complications that need to be further assessed on including support between game makers & Twitch themselves, accessibility for the streamer and “Twitchlets” service, and of course small Quality of Life adjustments like multi-queuing actions. However given the timeline of the project and the scope, I believe this will be a great jumping point for Twitch if they wish to increase engagement and watch time with users and their favorite streams. In fact, as of writing this article, Google is currently beta testing a system similar to this with Youtube Live. Also, there are resources Twitch could capitalize on that already uses their platform to push this project out sooner and better. Crowd Control, for instance, is an independent group of individuals who have developing a system similar to “Twitchlets”already for classic games like Tony Hawk’s Pro Skater and Super Mario World. If Twitch wanted to, they could possibly collaborate with Crowd Control to build “Twitchlets” and have an succinct advantage over competitors who are building their own “Twitchlets.” The opportunity is there, Twitch just needs to capitalize quickly on what is given to them.
