In every sphere of entertainment, from books and movies to music and YouTube channels, supply exceeds demand. Podcasts are no exception to this trend. With millions of hours of podcasts available on the internet, it's common for users to feel frustrated while trying to discover podcasts that pique their interest.
This "discoverability problem" isn't new, as any avid podcast listener would confirm. So when Paul English, podcast enthusiast and long-time partner of Arbisoft, finally decided to address this issue, he took inspiration from TikTok.
TikTok became famous partly because of its smart algorithm, which was excellent at finding relevant content for users. So what if there was an app with a discovery element with a similar algorithm that made it easier for listeners to find podcasts that fit their taste?
This idea was the starting point for the creation of Moonbeam.
The development journey began in November 2019 when Paul started discussing the podcast app idea with his friend, Youngme Moon. As a Harvard Business School professor and the host of her own podcast. Youngme proved to be an indispensable asset in the initial phase. In fact, the name "Moonbeam" was inspired by Youngme's last name!
By March 2020, Paul had started drawing screens for the basic app design and was working on brand identity and UX. Once satisfied with the general layout, he brought the idea to us. We helped create the first version of the app and released it on iOS in October 2020.
That was just the start. By February 2021, we had received considerable feedback from early users that the discoverability element of the app needed more work. This feedback and our research helped us pivot the approach and add the missing piece of the puzzle - a feature called "Beam".
"Beam" presented the best curated moments from some of the best podcasts. This automatically eliminated the hassle of digging through traditional podcast directories containing millions of hours of content. Users could swipe up if they didn't like the curated moments from a particular episode and the app would play something else that might fit their liking. Each action by the listener, whether a swipe to dismiss an episode or time spent listening to a clip, provided valuable information to the algorithm that perfected its recommendations based on the user’s activity while recommending other content that listeners with shared interests were consuming.
Once the initial user feedback was addressed, we relaunched Moonbeam in June 2021. Almost immediately, we experienced a surge of interest, and the app was even featured on Product Hunt, TechCrunch, Apple, Yahoo,Techradar, and several others. Over the next couple of months that followed, we iteratively enhanced the product based on user feedback, which resulted in Moonbeam having all the features of a traditional podcast directory alongside the discovery element.
This largely validated our product strategy and gave us the confidence to begin merger and acquisition efforts for Moonbeam. We knew we had built a technology that an existing platform could acquire and use to become much more effective without spending considerable time developing such tech. While at the same time, Moonbeam was doing well for itself with a consistent number of new and regular users. We knew the potential it could have with a large user base, something that was possible only if an established audio platform took this technology and integrated it into itself.
The iterative improvements to the app kept on coming. We started working on establishing a creator-friendly platform, allowing podcast hosts to select clips of their show to be featured in Beam. We also made Moonbeam stand out from the crowd by adding a new "Tip" feature that lets listeners tip their favorite shows in just a few taps.
We also implemented our speech-to-text transcription technology (hosted on our in-house servers), primarily used to transcribe the curated moments (clips) to provide better recommendations to the listeners. We could’ve utilized existing services for this purpose. However, given the volume of content that we had and our team’s ML expertise, it was much more cost-effective to build our solution.
That’s something we do quite often at Arbisoft– building our own products and solutions to meet our needs when third-party services don’t seem feasible or provide the value we’re looking for.
Our negotiations with prospective buyers intensified in March 2022. After meeting several potential acquirers, our discussions closed with Audacy - the second-largest radio company in the United States. We found that Moonbeam was very nicely aligned with Audacy’s vision for audio discovery, and what’s even better is that our technology and codebase proved to be closely compatible with their architecture and workflow.
As a result, Audacy decided to acquire Moonbeam, such that they’d not only be able to use the underlying tech but more importantly, make use of the team that’s behind the development of the cutting-edge and innovative technology that powers Moonbeam.
Although it wasn’t smooth sailing all the way, the team's dedication, ingenuity, and creativity helped Moonbeam reach its potential. From obscure priorities to shifting relationships between the internal team and client, the Moonbeam team overcame many obstacles to ensure the app created value for users and stood out from the crowd. Moonbeam was one of the most advanced technical implementations done by Arbisoft in a short period. The Moonbeam mobile app is available worldwide on both the App Store and Google Play.