Renee Wang is the founder/CEO of Castbox, a leading innovator in the podcast and audio space.
What opportunities are you seeing in the podcast space?
By 2020, 30% of all broadband sessions will done without a screen. These activities are getting more and more popular and audio is one of the main channels for users to access information. This is a global trend. In particular, paid audio content is getting stronger. In China, before 2015, nobody was paying for audio content. Since 2016, it has grown quickly. For example, in 2017, $500 mil USD was spent on paid audio content. By 2019, that figure will be US$1 billion. There are similar signals coming from other countries too, including Korea and Japan. Paid content will create incentives for better content creation.
Looking back, you can see all the ways people have accessed information. People used to watch TV, now it’s Netflix. People used to read magazines, now it’s Instagram. People still listen to radio but this will be replaced by a more personalized and smarter platform like podcasts. Machine learning and natural language processing is also getting better at predictions so this will drive further adoption.
Castbox solves a big discovery problem by providing in-audio searches. How are you using machine learning to process the episodes?
The way it currently works is we transcribe all audio into text. We then add timestamps to every phrase and index all the text into a search engine. Then, using natural language processing, we figure out what the key words mean. For example, if you searched the term “blockchain”, it will tell you where in the episode they mentioned blockchain. It may not be the title or description, but we can show that it was mentioned at 52 min and 21 seconds into the episode.
However, in the future, this will change. I believe the best UI is no UI. Currently you still have to take a lot of actions on the screen. Once Google Assistant and Siri get smarter, we will be able to leverage such voice assistances to make a truly voice and audio only interface. You will simply talk to your device, whether that is a smart phone, smart speaker, or smart car. All the search will be done on the backend and we’ll deliver to you the best results. If you like the episode you’re listening to, you can let it keep playing. If not, you can skip it. Using your response, the machine will more intelligently serve more personalized results to you. I believe this will happen soon, especially when you look at how powerful Google’s new assistant is.
How long does the above process take and is it fully automated?
For a one hour episode, it takes about one hour and 10 minutes to do the whole analysis. It is fully automated.
Tell us about the vision you have for Castbox original content.
We are positioning ourselves as the Netflix for audio. However, the current strategy is not to do exclusive content. Rather, we are using an open distribution strategy and launch the content on all platforms. Because our original content is well both produced and interesting, it’s been getting a lot of listeners. One of our top shows, Don’t Mess by Christine Sydelko, was number 1 in the iTunes comedy category on the first day it launched. By reminding the users that the content is produced by Castbox, we are using the content as a way to attract them to our platform. we are confident that once users will see that great functions we have, they will stay with our podcast player. Also, because we are a platform, we also have the data on what content users like best. So we try to produce it tactically. We try to understand what kind of niches need to be filled. For example, we’ve noted that the majority of podcast listeners are males. So we are trying to target more female listeners to that they have more content too. This is important because females are more likely to share and comment. We’ve seen this across our original content, with 60–70% of our audience being female and the engagement rate is very high.
Being a global company with a global audience, are you targeting your listeners geographically?
Content is a very localized business. If you’re not localized enough, it’s very hard to make great content. Thankfully, we have very talented team members in the US so we can create great content for the US audience. At present, we’re making mainly English content, and we’re targeting the US because of its culture. Pop culture is very important in the US and by making content based around such culture, it can have a global audience. Also, by creating original content for the US audience, we have our own IP which we can license to turn it into a series. Making content for other markets is more difficult. For example, if we make content for Indonesia, it’s unlikely to have global reach. That being said, we do have a team in Korea and we are working with Korean publishers and producers to create content for them. We’re not doing too much in China because there are already three unicorn companies that are doing great work in the audio space. We don’t want to waste money or resources competing with them.
Revenue generation for podcasters has so far been limited to ads and donations through Patreon. Tell us how why you’ve launched a paywall service into Castbox.
Every day as a company we ask ourselves, ‘how can we help our creators make money?’ Because that is our only chance to grow bigger. The current situation has a few pain points. Creators in general are not being rewarded for their real contributions. Instead, the middle man takes most of the revenue. Look at Spotify where around 55–60% of revenue goes to the record companies and only 11% goes to the musicians. As a donation system, Patreon is great, but not a lot of users are on it and only 5% of them are willing to donate. Also, it’s centered in the US and internationally the percentage of donations is even smaller. When we think about the consumers, the users that do the sharing and engaging with comments also get nothing. It is unfair that they are not being rewarded for their contributions. I want to create really fair terms for all stakeholders. We just launched a paywall service in conjunction with Wonderyas the lead partner about three weeks ago. Although it’s still too early to see an uptick on subscriptions yet, we’re optimistic that this will help support the industry further.
You’ve recently integrated blockchain into Castbox for payments. How will this work?
We think blockchain can resolve some of the problems described earlier. We have integrated a digital wallet inside Castbox and, for certain podcasts, have opened the reward function. We finished all the development but have not activated because we want to get permission from the creators first. The way it will work is that all behaviors will be recorded in a smart contract. As the content is consumed, all stakeholders participating in that, including the content creator and the listener that shares or comments, receive a percentage of the revenue. We are still working on the tokenomics and incentive structure. We want to get this right and will be focused on implementing a framework for how listeners are rewarded for their tasks and creators for their value of content. Our token is called BOX, and it will be freely tradeable once we are listed on an exchange. It is an ERC20 token for now. We also plan to distribute BOX tokens as follows: 25% presale, 15% team, 30% ecosystem incentives, 20% foundation, and 10% partnerships. I don’t believe this should be a zero-sum game and I know that by opening this up, the whole content industry can boom.