Podcasts are broken, Twitter is only just beginning to solve its business problems, and I couldn’t be more bullish on both. For starters, 348 million people listened to podcasts in 2016 — roughly the same number of people who use Twitter. Many of those listeners aren’t currently on Twitter which could be an interesting growth opportunity, especially considering podcasts are one of the best ways to find people to follow. But 348 million is just the tip of the iceberg.
Podcasts are the digital evolution of radio and radio is probably more popular than you realize. According to the Pew Research Center, more Americans listen to radio than use Facebook, meaning podcasts could have well over a billion listeners when radio is finally disrupted.
With radio, listeners are constrained to a handful of stations and whatever happens to be live at the moment. But with podcasts, listeners can access just about any audio available on the internet whenever they want. The combination of the near ubiquity of radio and the limitations of the technology it relies upon should make it a prime candidate for disruption. So why hasn’t software eaten radio?
It’s because podcast discovery sucks.
For all of its limitations, radio is simple. You turn on your car, press a button, and you’re listening to something. In contrast the podcast user experience is high friction. You have to know what shows exist, search for them, subscribe to them, and check each one for new episodes. Even if you build up knowledge of many shows over time, you still have no effective way to parse all of the episodes currently being published across the entire podcast ecosystem — you don’t know what exists, what you’re missing out on, or if a given episode is especially good, bad, or uniquely appealing to you before giving it a listen.
The signals that are used to help discover content on other sites like Netflix, Facebook, Instagram, Twitter, Medium, and Nuzzel aren’t harnessed in podcast apps. Things like: of the friends, authors, journalists, celebrities, etc. I follow, who recommends this episode? Are any of them hosts or guests? What episodes are trending right now? And what episodes do other people with similar tastes to mine like?
In order to fix podcast discovery, a podcast app must understand the interests of its users, enable peer to peer sharing of episodes, and deliver high quality recommendations of episodes by tapping into those signals. It’s interesting to note that Twitter was born out of Odeo, a web-based podcast platform. Odeo, like all of Ev Williams’ companies, was focused on democratizing information. But as it turns out, the broken thing about podcasts — and the real opportunity to democratize the information in them — is not by giving every single person a microphone and record button; it’s by giving listeners the power to curate, like, and share the episodes they find meaningful with their friends and others who follow them. And no podcast app in existence today is focused on this.
The concepts of feed-based discovery and social notifications aren’t new — Facebook invented the news feed in 2006 and dynamic feeds have been in play at Facebook, Twitter, Instagram, etc. for a while. It’s just that no podcast app focuses on building out social and interest graphs and couples them with a dynamic home feed of recommended episodes. This is also why so far, podcasts have been unable to “go viral” — podcast apps themselves don’t enable the mechanisms to support social propagation. It should be just as easy to discover a great podcast episode as it is to find a great tweet, Facebook post, or story on Medium.
And Twitter is better positioned than anyone to do it — in large part because what you search for (Google), who you happen to know in real life (Facebook), and what you buy (Amazon) are weak indicators for what you want to listen to compared with who and what you are interested in (Twitter). Twitter is already the best place for podcast discovery, it’s just that no one has extracted the information and coalesced it where the listening selection happens — in a podcast app.
And as it just so happens, the majority of podcast hosts and guests are already active and accessible on Twitter. But there are still more reasons this is a uniquely compelling play for Twitter and great for podcasts.
Twitter Has a Cold Start Problem
Twitter has had near-stagnant growth for the last 3 years, currently resting at 330 million active monthly users. However, Twitter has the potential to have well over a billion active users. One reason is that when you include 3rd party digital embeds and tweets shown on TV, 1.6 billion people see tweets each month. This suggests the content on Twitter is relevant, but that most people are ok with not logging onto Twitter itself to consume it.
The biggest reason more people don’t use Twitter is that when a new user signs up, the app knows practically nothing about who they are or what they care about. Twitter is then faced with the product challenge of constructing an onboarding experience that not only explains what Twitter is and how to use it, but that also susses out the interests of a new user, accordingly recommends accounts to follow, and delivers them into an initial timeline experience that is compelling enough to get them to come back again.
Just how challenging is Twitter’s cold start problem? One *billion* people — 3 out of every 4 to ever sign up — have tried Twitter and not come back. Twitter’s onboarding flow asks new users to select high level categories that interest them — Technology, Sports, etc. — which eventually map to static lists of recommended accounts to follow. This means that some subset of the same several hundred accounts gets shown to every single new user. Twitter fails to get personalized enough with its cookie cutter recommendations and so many users leave and don’t return.
Twitter should rearchitect its onboarding to dig deeper into subjective interests, but another important solution is to know something about a user’s interests before they sign up. How would that be possible? Twitter would have to already own another standalone app that is able to collect that information. Similar to Amazon acquiring GoodReads to help with the discovery and recommendation of books to purchase on Kindle. So what standalone app experience could Twitter own that would serve as a gateway to peoples’ interests?
There are other interesting candidates like books, articles, etc. but podcasts— like Twitter — span to cover what going on in the world: from breaking news to celebrity gossip, sports to comedy, science to philosophy, and every topic in between. And podcasts, like tweets, are getting churned out everyday and directly feature the most interesting people in the world.
The Future of Twitter: The Interest Graph
Amazon is no longer just a digital book store, Google no longer just a search engine, and Facebook no longer a private social network. Ultimately, Twitter will grow to become something beyond the native Twitter app and the time to start that expansion is now.
Live video is interesting, but I think the fundamental aspect that makes Twitter so uniquely appealing isn’t merely that it’s live. It’s that Twitter is the only all-encompassing social interest graph on the internet. With Facebook, you’re limited to seeing random viral videos and updates from people you know in real life, but on Twitter you get the chance to see what the people you find most interesting in the world — people you would otherwise not have access to — are talking and thinking about. Facebook can’t even successfully pay celebrities to get them to use their site in this manner while Twitter is basically a live 24/7 Reddit AMA between everyone.
And many people are already using Twitter to discover other things in the world. Not just who to follow or what podcasts to listen to, but what newsletters, blogs, books, articles, TV shows, movies, restaurants, bars, live events, songs, video games, board games, travel destinations, apps, etc. to check out.
If Twitter builds out an ecosystem of discovery apps to solve each one of these use cases, it will provide utility to its existing users and probably unlock new revenue streams. More importantly, it will give the 2.8 billion people who are online and not already on Twitter many new reasons to enter its ecosystem and begin to build out their own interest graph. The whole is greater than the sum of its parts and frankly the discovery of many things on the internet could be way better than they are now.
Why Should Twitter Build a Podcast App?
To quickly recap:
- Growth and Monetization: The podcast opportunity can easily scale past a billion users and hundreds of millions in revenue. Software is eating the world and the disruption of radio is inevitable — also the podcast space is largely overlooked and without meaningful competition. Podcast ad spend has nearly doubled in the last 3 years, outpacing listenership growth, and given that Twitter already has an existing ad network it should be able to take advantage of this — just as Facebook plugged Insta into its existing one. Twitter could also experiment with other new monetization streams with podcasts like subscriptions, donations, selling of merch, surveys, and the selling of live tickets.
- Unique Advantages: The bottleneck to disrupting podcasts is solving discovery and Twitter is uniquely well positioned to do that because graph-based content discovery is a core aspect of the Twitter experience. The out-of-the-gate utility Twitter’s podcast app would offer to its existing 330 million users would likely make it the 2nd most used podcast app in the world very quickly. After demonstrating what social discovery can unlock and creating the first app where podcasts can go viral, I wouldn’t be surprised if podcast hosts will want to work with Twitter to expand their audience and have other problems solved. This could result in hosts asking their listeners directly to use the Twitter podcast app — a very compelling growth channel — and opens up the possibility of Twitter entering the audio hosting an RSS feed management space. A move that would enable Twitter the freedom to operate independently of Apple and other entrenched players in the field, moving quickly and unilaterally to solve problems for creators and to bypass existing systemic friction in the podcast ecosystem (e.g. creating an interface and new data standards for creators to automatically link Twitter handles of hosts and guests in the RSS feed itself).
- Long Term Business Strategy: Podcasts are a low-risk, low-resource investment experiment with high-upside potential and if successful, could unlock a much larger business strategy for Twitter that enables it to double down on one of its most unique and compelling properties — the interest graph. Discovery 1.0 on the internet was search. Discovery 2.0 will be a serendipitous experience that combines everything it knows about what you’ve seen, what you like, and what the people you care about like to offer you suggestions you didn’t know you wanted — not only would that be the biggest upgrade to online discovery since Google won search, but it could be the next defining chapter for how we use the internet.
Podcast apps are living in the dark ages and don’t even incorporate the digital norms of discovering information that were mainstream 10 years ago. It should be no surprise that listeners love their podcasts but don’t extend that love to the app they use to listen. Despite how bad podcast apps are, total listenership has been steadily increasing for years, ad spend is accelerating, and several hundred million people are listening. Imagine how many more would listen if Twitter finally made the first podcast app that solved discovery.