The Bootstrapped Founder

Arvid: 00:00

Hey. I'm Arvid, and you're listening to the Bootstrap founder. This episode is sponsored by paddle.com. So if you're looking for a payment provider that does all the heavy lifting for you, so that you can actually focus on your business, go check out Paddle. I use them for all my SaaS properties.

Arvid: 00:19

I really recommend checking them out at paddle.com. They got a great API. They got a really, really good insight into your customer data kinda dashboard. It's spectacular. Now, brainstorming time.

Arvid: 00:33

Today, I wanna explore something that's been on my mind a lot lately. And that's the massive set of opportunities that I'm seeing in the podcasting space, particularly for software entrepreneurs. Let's talk business ideas in podcasting. You know that I've spent the last year building podscan.fm, talking about this non stop on the podcast. And being deep in the podcasting ecosystem, I've gotten really, really unique perspective on where things are heading, where things are coming from and where they're going.

Arvid: 01:01

And let me tell you, I think this market is absolutely massive and it's growing. We're looking at around 500,000,000 listeners by 2025. And here's what's really interesting. These aren't just casual listeners who tune in once and disappear. When people find a podcast that they like, they stick around for 100 of episodes.

Arvid: 01:19

Not maybe on average, right? Average likely is in the 30 or 40 episodes, but that's still a lot. Right? There are people who stick with podcasts for thousands of episodes at this point. It's incredibly sticky content and that makes the industry interesting because once you have a listener, they're hooked.

Arvid: 01:35

You can keep doing stuff with them and for them. And I think I am in this well quite unique position here to spot opportunities because I'm wearing 3 hats, different hats at once. I'm a podcaster myself as you can probably attest by just listening right now. So I know all the production headaches firsthand. I'm an avid listener.

Arvid: 01:54

I listen to podcasts all the time, so I get how hard it is to find the right stuff to listen to. And as the founder of a podcast data and analytics company, I get to see all these gaps in the market that entrepreneurs could potentially fill. It's like having a bird's eye view of the whole ecosystem from 3 different perspectives. So let me start with something that I think is really fascinating about podcasting to begin with. Whenever people ask me what the most effective way to reach an audience is, I always tell them now that I have some experience with this, that it's audio, particularly podcasting.

Arvid: 02:25

And here's why, just imagine how people consume podcasts compared to other media. When someone puts on their headphones, you are literally speaking directly into their mind. It is quite likely that I'm talking to you from inside your head at this very moment. Think about this for a second. TV, that's external.

Arvid: 02:46

You're looking at a screen somewhere else. Books, you hold this in your hands. You can feel that it's not inside you. You take it in, but it's still external. Websites, everything is on screens.

Arvid: 02:55

Everything is an external experience. But podcasting, you're right there in people's head. And it's the most intimate medium we have and it builds trust. It builds understanding. It builds connection like nothing else.

Arvid: 03:08

Now, here's where this gets interesting. The industry is seeing this massive growth. Just look not only at the numbers of listeners, but also what's happening in the entrepreneurial world, right? Spotify with their acquisitions. But at the same time, anyone can start a podcast with basically minimal equipment.

Arvid: 03:26

It's just weird paradox. You have this massive VC funded business side of things, and then this almost democratizing content creation part on the other side. And that's amazing. But it makes it really hard to convince people to pay for premium tools and services. Like why should I pay for this when I can just use my phone's microphone to record?

Arvid: 03:44

Or why should I pay for yet another service when Spotify is like $15 a month and I have access to every podcast out there? It's a problem, but it's not unsolvable. Let me break down where I see the biggest opportunities in podcasting right now. I think I'm gonna focus on 3 main areas, which is kind of a production, consumption and things that happen after people have released and consumed the podcast. And I'll tell you exactly why each of one of these matters based on what I've seen both in building PodScan and in running my own podcast.

Arvid: 04:14

The first I wanna start with is discovery community consumption. Like, almost the B2C angle here, right? Like everything that has to do with people consuming podcasts. You know what we really need? I think we need something like Shazam but for podcasts.

Arvid: 04:31

I wanna be clear, it's weird, like just the idea of this because it's spoken word. And I know it sounds a bit out there, but hear me out on this one. We'd need some kind of serious audio fingerprinting technology and that's probably not the same technology that Shazam has been using for the last decade or so. It's probably gonna be AI embedding space that you could identify not just what people are saying but how they're saying it. Because I wanna be able to take my phone and tell it, hey, somebody just told me about this podcast where like this one scientific guy talks about sleeping patterns and the other guy just mentioned that particular thing.

Arvid: 05:04

What show was this? It's kind of a search, but it's semantic. And it's has more to do with mood and vibe, right? Because think about it. Summaries and transcripts don't really capture the actual experience of an audio show.

Arvid: 05:18

And if you reduce it to text, which is the way it's currently searchable? And that's what PodScan does as well. Right? Transcribe all these podcasts. But it still loses this kind of the the voice, the tone.

Arvid: 05:28

So, what we need is something that can analyze the audio and say, okay, these hosts have great chemistry. They use a lot of cultural references from the nineties. They explain technical concepts in an accessible way for people from this particular background, that kind of thing. Would probably need to synthesize all this information into something comparable, like the audio fingerprint, so you could find shows or even episodes or even segments of episodes that match the vibe that you're looking for. That is generally the discovery problem of podcasting because the medium of audio is so hard to access that people come up with all these abstractions, summaries, show notes, all of this, but that loses a lot of data and that data itself is interesting.

Arvid: 06:08

And here's what makes this technically challenging. You would need to be able to peek inside of all of these different platforms that house the audio data somehow. Because right now, if a podcast is only on Spotify or only on their website or hidden behind some other kind of walled garden, it's really hard to get access to that audio content for searching. Google can only see what's publicly available on websites and we need to bridge these gaps somehow. Now, technically, all podcasts are supposed to be built on top of the RSS platform, right?

Arvid: 06:40

The files are on some hosts and there's an RSS feed that tells every consuming audio player, every website that integrates podcasts, where new episodes are coming from. But if you only host your podcast on a big portal, you really can't get to the audio data. And that is a problem. And then beyond that, there's a community aspect, which I still would see as a B2C opportunity. That's something I'm really excited about.

Arvid: 07:07

Podcast audiences are incredibly loyal. I talked about the stickiness and I see this in my own analytics at all times for every episode. But here's what's interesting, different podcast communities need completely different tools. Like if you're running a fantasy football podcast your community needs totally different features than let's say a Harry Potter fan podcast community, right? You talk about different things clearly but it's also different in how people interact with each other.

Arvid: 07:35

Like, do they give each other tips? That's probably more fantasy football. And do they just talk about the things that they love about some story that they found somewhere? That's probably more the the Harry Potter thing. So, with more shows doing live episodes on top of this I think there's a whole new layer of community interaction tooling that we need to think about.

Arvid: 07:56

So don't just see a podcast as this audio artifact that gets deployed into a store and then people listen to it. The podcast itself is this whole living breathing being that is very much based on the community of its listeners. And you'll find this for almost every podcast that is very popular out there. There are a lot of communities like unofficial ones and often very official communities, Discords, chat rooms, forums, like this, Slacks sometimes that exists because the podcast needs the listenership and the other way around obviously too. And that is where tooling can happen and should happen.

Arvid: 08:38

And often this is also where budget exists because those podcasts that already have this large audience often have ways to get some money. Right? Patreon is a big thing in the podcasting community, but sponsorships are also very big. And often just subscribers, they have paying subscribers for bonus feeds and that kind of stuff. So there's budget in this b to c world.

Arvid: 09:02

Let me move on to the production side of things, mostly because I care more about this, like being a podcast producing person myself. And there are also clearer indication of budget in this side of things. This is something I deal with every week on my own show. I can really talk about this. I run an audio only podcast, as you can probably tell.

Arvid: 09:21

And let me tell you why. Because video, something that I used to do is just way too much work. And here's the thing though. I actually record video for all of my interviews with the people that I talk to. I just don't use it.

Arvid: 09:35

So I have it but I don't use it. Why? Because editing video is incredibly time consuming and expensive. Because if you want it done right, you have to pay somebody and it takes a long time to do video and cuts and transitions and all that. With audio, still a lot of work, but it's not as, let's call it artistic in a sense.

Arvid: 09:55

Right? Video, there are certain expectations that people have and that takes a lot of time. We're talking about a full day of editing for just 1 hour of footage when it comes to video. That often includes the work on audio as well or may not include it and that's additional work. It's just a lot of work.

Arvid: 10:08

And when it comes to automating video production, here's what I'm thinking could be done. Because if your final edited audio file might be from any kind of tool, could be Pro Tools or Descript or Premiere Pro Audacity, whatever you add it in doesn't matter. You have some kind of commonly understood file format that shows the edit compared to the original audio files. So the tool that I would like to see in the market that I haven't yet found would be able to take that audio file plus the raw video recordings and the edit file, like the file that describes how the edit happened, and then automatically sync them up into a video format, such as matching timestamps. It needs to make smart decisions when to switch between shots.

Arvid: 10:50

So there's probably some kind of diarization, detecting who's speaking, how to handle transitions, how to handle cuts when, like, parts of the audio were removed, maybe even identifying engaging moments where you wanna zoom in on someone's reaction. All of this stuff sounds like AI work that can be done. It's also complicated, but I feel that would be a massive thing that I would personally pay for immediately. If all I had to do was upload my source files and my Pro Tools file, then, yeah, I'll take the video for this. And for the audio cleaning side, I think we should also be thinking about what tools we can offer to make podcast audio better.

Arvid: 11:27

And I know there are things out there like Adobe's podcast stuff and the script studio mode and that good, but their general purpose. I would like to see AI models specifically trained on conversational audio podcast style. What if we had something that really understood podcast conversations? We understood not just how it declinliness of the audio is supposed to be, but also how it's supposed to fit together as a conversation. Like, it could distinguish between a thoughtful pause and dead air.

Arvid: 11:55

These are 2 different things. Or know the difference between background ambiance that adds atmosphere and genuine noise that needs to be removed. Yeah, there's a lot of stuff in production. And I have so many ideas. That's just the audio side of things.

Arvid: 12:08

Let's even look at, tracking who is listening. Like for something that seems to be solved, that really isn't solved yet is tracking who is listening to a show. Like they do this with RS prefixing. Like it's pretty much like a link forwarding for the RS feed that a podcast has for better analytics. So that anybody who wants to download an episode kind of clicks on a link that is registered on one website that then forwards that click to the next website that forwards that to the next and next and next.

Arvid: 12:39

And ultimately it ends up at the original audio file, but it's gone through like 3 or 4 different forwarding agents, each of which registered that as an analytics, as a click or something. That's how podcasting works right now. You should see, where popular podcasts, where those audio files are hosted. There is often 8 to 12 prefixes to the URL that all go through these different server hops just to register clicks or your attribute who is listening and that kind of stuff. It is crazy, but it's not been solved really well because you don't really get this information still.

Arvid: 13:15

Like, it's still super hard to get analytics about podcast that is reliable. And that is something that should be looked into. I don't really know what exactly to do here because that's kind of out of my depth, particularly attribution of who a person is and who they're not. And I think it's still a privacy issue, but you still can't tell who's listening, when they stop listening, what parts they skipped, none of this because RSS is just like a download, that's it. There's a huge opportunity here for someone to crack the problem while still respecting user privacy.

Arvid: 13:46

I think that's still very important because that's something that podcasting does really well. If you don't wanna say who you are as a consumer, you don't have to. Well, unless you're logged into Apple or Spotify, but there are alternatives out there. And you know what also would be amazing? Some kind of AI powered tool that could help me and don't forget producing podcasts, right, during interviews.

Arvid: 14:06

Now I talk to people a lot and I think I'm okay at asking questions, but I don't know about you. When I'm interviewing someone, I'm trying to listen to what they're saying and think about my next question at the same time. Remember what we talked about earlier and keep track of time all at once. It's not easy. It's quite stressful.

Arvid: 14:23

And I consider myself introverted in many ways, probably extroverted when I talk about things that I like, like what I'm doing right now. But it's not the easiest or energy conserving thing to me that just talking to somebody is really, really tough. So I would love to have an AI assistant that's listening to the conversation with me in real time and then constantly suggests interesting questions or follow ups for the next 30 seconds. Something I could do over the next minute or so. I could even set different goals for it.

Arvid: 14:52

Like if I wanted to make the conversation more educational, then I could press that button. I don't know, on my stream deck or something and it would come up with more, you know, chatty questions. If I wanted to be more in-depth, I could press a button and it would come up with more like precision questions. It would be really cool. I can tell you how many times I finished an interview and thought, oh, I should have asked about that.

Arvid: 15:14

This would be like having a really smart producer in your ear, helping you create better content and being highly flexible during the process and being effectively free, right? Don't forget about that. Like if you're an AI agent running right next to you for an hour a week, that doesn't really cost you as much as an actual producer. And this could make podcasts so much better, right? If you listen to people who are starting out interviewing and you just need to listen to my first couple episodes where I do this, it's really hard.

Arvid: 15:42

You make a lot of mistakes and you leave a lot on the table. Having an AI that can constantly help you increase the quality of your content, it's a good idea. Okay. Another example from the production side. That's something more on the sharing and marketing side of things.

Arvid: 15:57

See, there's so many angles to production. Right? Do you know these audiograms that everybody shares on Twitter and social media? The the ones that have the waveform and the captions, and then they kinda just take a segment of a podcast and you share it as a video of the waveform. But, you know, this just really text being shown and audio being played.

Arvid: 16:15

I think we need better tools for this. I'm thinking something that could take the audio to full audio of a show, transcribe it, then automatically pick the best highlights and generate brandable, customizable audiograms that you could just share right away. And I know that tools like Descript have this built in, but it's not their main focus. There's room for someone to come in and just nail the specific problem for podcast producers of a certain category or of a certain size, maybe even just solo shows or just interview shows, stuff like this. There's still specialization potential in this niche.

Arvid: 16:48

Okay, so we talked about consumption, we talked about production. Let's talk about the thing that happens after a podcast is being released, the data and monetization too, like turning the podcast into a business because that is where PodScan, my business operates. We help people monitor mentions of their keywords, their names, their brands and manage sponsorship opportunities or guest bookings, but there's so much more potential in this space. Think about all the knowledge trapped in podcast archives. In untranscribed podcast archives, I'm working on this trying to transcribe all of them, but there's still a lot out there.

Arvid: 17:21

What if you could turn that into an interactive experience? And I've seen people experiment with this. So I'm just gonna share like the rough outline. Here's what I mean. Imagine taking everything that someone like Tim Ferris or Rob Walling from our community has ever said in their podcasts.

Arvid: 17:37

And for Rob, that's now over what 500 episodes or a 1000. Like, there's a lot of stuff that Rob has shared over the last couple of decades. And then you feed this into an AI agent and the AI learns everything that Rob Walling or Tim Ferris have ever said. And then you are able to have a conversation with that version, the podcast version of them, whenever you want. You could ask them about their expertise, you could ask them to critique what you're doing, you could have them share their experiences because they have shared all these stories, it's all in there.

Arvid: 18:08

And you get answers based on everything they have ever publicly shared. And this isn't just for the big names. Think about every podcast out there where an expert has something to say. All the cooking podcasts, all the chefs, the craftspeople from the woodworking podcasts, experts in their fields who've shared their knowledge through podcasts. It's crazy.

Arvid: 18:26

There's so much going on. I've seen people use this in medicine, in real estate. It's bizarre. You can take every single niche, take all the experts and condense these millions of words into AI models that you can talk to. That you can have look at your documents, look at your screenshots, look at your photos of whatever you're building and then have them critique it.

Arvid: 18:45

It's crazy, but that is possible. So I see a lot of opportunity in this field. And here's another thing. Podcasts are incredibly current and compared to other media, right? Compared to websites that can often be like years old.

Arvid: 18:58

Podcasts usually are super current. At least the last episode is the latest one. Unlike YouTube videos that might cover topics, spending years of history and just be very old, Podcasts are often about what's happening right now, this week, what's coming up next. And there's a huge opportunity here in trend tracking and forecasting based on podcast content that nobody's really tapped into just yet. I see Google Trends, that kind of stuff, but podcasting still seems to not really factor into this.

Arvid: 19:26

So that might be something to build. I might be working on this. Who knows? So let's go into the the final thing here, monetization, for this part because I think that's still really underserved too. Most podcasts are really stuck in the sponsorship and ads model.

Arvid: 19:41

And what if there was something, well, like Patreon or OnlyFans, but just for podcast content. Better ways to get direct listener payments and direct listener interaction. The loyalty of podcast audiences is incredible and I think we deserve better tools to help creators capitalize on that. And I think we can build specific tools that integrate into the audio experience more than these tools that are trying to integrate into any kind of fandom that we have. And there's something that nobody else has cracked yet too that would be voice training and coaching for podcast hosts.

Arvid: 20:15

I think this is all over the place because this technically is a production thing, but you know how singers have voice coaches? I think podcasters need that too. I certainly need somebody to help me speak better. This probably might be evidenced, on this very podcast. And I think most people don't realize how crucial good vocal delivery is.

Arvid: 20:33

I mean, listen to my early episodes sometimes and you will just cringe. Not just the audio quality, but also just how how I'm speaking, you know, it's just hard. It's hard to learn. So what if we had an AI system or a real person, a service business that could analyze your speaking patterns, help you improve. Like it could tell you when you're meandering too much or help you be more precise in how you phrase things, give you tips on pronunciation.

Arvid: 20:56

Because let's be honest, nobody wants to listen to someone drone on forever about a topic that could have been explained in half the time. The nature of the podcast is that it's audio and audio can take a long time. So you can't really condense it, you can't really skim it. So you wanna be able to deliver just right. There's also opportunity I think in other languages that that's something we really need to talk about the translation and dubbing.

Arvid: 21:21

It's a huge opportunity and you can already see YouTube working on this, right? YouTube has been automatically dubbing certain shows, which is really cool. But I think it's way more complex than people realize because if you dub a German podcast into English, it's not just about converting the words. You need to understand the German market context, the industry specific terminology, certain cultural things, references that might not translate directly. And you'd need an AI model that's not just language aware, but also context aware, industry aware, industry specific models for different types of content.

Arvid: 21:51

I think there's opportunity here. And same for show notes. This is really technically interesting. You could use an AI to not just transcribe the conversation and pick out the most important things, but to actually understand the structure, the flow of the conversation. It could identify when you introduce a new topic, timestamp that.

Arvid: 22:08

When you're making a key point, timestamp that. And when you're giving an example, pull them together, time stamp that. Right? You could automatically format all that into show notes that are optimized for both human readers and for search engines. And it could also filter out all the URLs and product mentions and find links to those to put into your show notes, giving all these things a boost as well from an SEO perspective.

Arvid: 22:32

Maybe you could even generate different versions of show notes for different platforms because what works on Apple Podcasts might be different from what works on YouTube and Spotify. Let me wrap this up with what I think is the biggest opportunity of all and that is finding that sweet spot between what podcasters need and what listeners want. Becoming the kind of the transition between these two audiences. That's where the next generation of successful podcasting tools will emerge, between production and consumption. And with how fast AI and audio processing tech is developing, I honestly can't think of a better time to build in this space, like right now is a good time.

Arvid: 23:08

So before I go, I guess I would love to hear your thoughts on this. Have you spotted any opportunities that I may have missed? Because I mean, I see a lot, but I don't see everything. Are you working on something in this space? I would really love to hear.

Arvid: 23:20

So drop a comment or let's discuss it on Twitter. I'm always open to this. And if you found this helpful, don't forget to subscribe to this fine podcast because I've got more episodes like this coming where I break down different markets and opportunities that I'm seeing, not just in podcasting, but in the wild world of SaaS in general. That's it for today. Thank you so much for listening to the Booster Founder.

Arvid: 23:40

You can find me on Twitter at Avidka, a I v a d k a h l, and you'll find my books in my Twitter core set too. If you wanna support me and this show, please go to radestpodcast.com/founder and leave a rating and a review, and tell everyone you know about podscan.fm. It makes a massive difference if you show up there because then the podcast will show up in other people's feeds, and PodScan will find new customers. I would love that. Any of this we're up to show.

Arvid: 24:05

Thank you so much for listening. Have a wonderful day, and bye bye.