MakeSense

MakeSense aims to help emerging content creators on Twitch using analytics to expand their presence on the platform. We leverage machine learning and NLP models to help streamers understand their audience and their most successful content. This in turn enables them to strategize the growth of their channel.

Our Mission

MakeSense’s mission is to enable emerging content creators on Twitch to grow their channel and understand their audiences through analytics.

The Problem

Live streaming has greatly risen in popularity, especially in the last few years. The live streaming market is expected to be worth 330.51 billion dollars by 2030. By far the largest live streaming platform is Twitch, which on average had over 15 million daily active users and 3.8 million unique broadcasters in February 2020. According to statista.com, during one fiscal quarter in 2021, 6.51 billion hours were watched on Twitch.

Most of the money and the views go to the top channels on the platform, leaving new creators a difficult ladder to climb to be successful. In addition, top creators often have management teams and agencies behind them that help them to maintain their Twitch channels’ popularity and engagement with their audiences. Meanwhile, there are plenty of creators on Twitch with the potential to grow and thrive on the platform, but lack the infrastructure and support to get there. Twitch itself values its creators but will tend to prioritize the bigger channels that already have larger audiences since they are more interacted with and thus affect Twitch’s traffic and revenue the most.

Our Solution

Utilizing natural language processing and recommendation models, MakeSense has created a dashboard for emerging creators who average between 500 and 2000 subscribers. This dashboard provides those creators with real-time information and insights on their viewership during their live streams. It can help creators gauge how their audience is interacting with their content. It also allows the content creator to look at their video analytics after the stream and zero in on segments where there were a spike in metrics, such as new subscribers or views, as well as give content recommendations based on their performance.

How it works

The user can choose to see their channel analytics during a livestream or after the stream. Some of the audience insights and statistics are derived from the Twitch API (i.e., viewers, subs) while others are derived from models built into MakeSense’s backend, like audience sentiment scores (created using a double tuned BERT model) and the topic recommendation algorithm. MakeSense aims to provide granular insights to a livestream such as when viewership and sentiment were highest during a stream, or which segments of the stream had the lowest follower growth. This data enables the streamer to, at a glance, understand their performance landscape, and tailor their content strategy to maximize growth, retention, viewership, or following.