Beijing-built App Castbox Speeds Up Your Research With Podcast Audio Search
This post comes courtesy of our content partners at TechNode.
Editor’s note: This was contributed by Elliott Zaagman a trainer, coach, and change management consultant who specializes in aiding Chinese companies as they globalize. To contact him, check him out on LinkedIn, or add ezaagman on WeChat.
I am a podcast addict. I’ve never been much of a reader, and tend to spend lots of time on trains, streets, and planes, so the revolution in downloadable audio content has opened up a whole new world of information for me.
As I’ve developed somewhat of a pseudo-career as a writer and blogger, I’ve relied on podcasts as sources of both inspiration and information for the topics I’m researching. However, I’ve come to find this method incredibly annoying and time-consuming on most podcast apps. The specificity of the search results is often woefully inadequate.
Take, for example, the popular app Stitcher. Here’s what happened when I tried to search the topic of “blockchain,” the decentralized, anonymous ledger system used by most crypto-currencies:
One hundred and seventy-three results, all for blockchain-themed podcasts. In order to find the information that I’m looking for, I will have to go through each podcast, episode-by-episode. Some episodes are filled with great info, others aren’t, and the only way to sort the good from the bad is to listen to them, or maybe try to judge by episode title.
This process has gotten a lot easier, however, since I’ve discovered Castbox, the project of former Googler, Peking University alumnus, and Hebei native Renee Wang (王小雨). Castbox aims to improve the search quality of audio files. Take a look at what happens when I conduct the same search, just on Castbox instead of Stitcher.
They offer similar results when categorized under “channel:”
When you search under the “episode” tab, it gets a little more specific, organizing by specific episodes. This is great when looking to access to-the-point summaries from top experts on a topic:
But where this app really stands out is in their “in-audio search,” which combs podcasts for references to certain terms, and then allows the user to, in just one click, go directly to the point in the episode when that keyword is discussed:
How it works
Castbox does this with a Natural Language Processing (NLP) algorithm which transcribes the audio of a podcast into text. They also use a machine-learning process to customize the search results based on the user’s listening habits.
This provides a user experience that I personally find preferable to the one provided by apps like Stitcher. “Traditionally, audio search results crawl channel tags or keywords of an episode and title descriptions,” explains Tina Kuan, Castbox’s Chief Marketing Office. “This age-old approach becomes ripe for manipulation by a keyword savvy podcaster – much like the issues plaguing SEO.”
Chinese roots, but focusing on the overseas market
While Castbox has a presence in the Bay Area, its HQ and the majority of its development team is based in Beijing. “Beijing has a large number of internet talents, but with relatively low cost, so it makes sense to operate from there,” says Renee Wang.
In basing itself in China but targeting an overseas user base, Castbox is hoping to follow somewhat in the footsteps of Musical.ly, the China-based video platform rarely used in the middle kingdom, but which is a hit with Western teenagers. In November, Chinese super-unicorn Bytedance (parent company of news aggregator Jinri Toutiao) acquired Musical.ly for what is reported to be between 800 million and 1 billion USD.
Despite its China roots, Wang emphasizes the desire and need to have a diverse team that is representative of the app’s user base. “Reaching global resources is very important to us, and that includes a push to hire talent from other countries,” explains Wang.
Room to improve
I’m certainly not the only one who sees the value in what Castbox offers. In October, they announced a US$12.8 million series A investment, from some of the same investors who were Musical.ly’s early backers.
However, I must say that their in-audio search function still is far from perfect. For example, I recently searched for content from the wildly popular and well-respected China tech news blog Technode, the total search results were definitely better than Stitcher, giving me both more channel options, as well as episodes of other podcasts where Technode reporters were guests.
Stitcher:
Castbox:
However, the in-audio search was a bit more disappointing:
Keep trying, Castbox. You’re already my new favorite app, but you’ve still got a bit of a way to go.
Images courtesy of TechNode