While working on my latest software project, Podline, I've been thinking a lot about how I plan to onboard customers and deliver the service offered by my SaaS business. And that's what I want to share with you today. Podline is effectively an integration that people can add to their existing podcast hosting platforms. If they have a podcast and want to communicate with their audience, adding a Podline voice messaging embed or link to the website, that's the way to do it.
And it's the only way to do it. And in a way, that's quite the ask, because Podline is not like basic podcast hosting or an analytics tool with automatic integration. Instead, users have to add Podline to their platform themselves intentionally and with a little bit of effort. It's easy, just an iframe embed or a link. But I still need to educate my customers on how to use it.
So to improve the customer experience, the whole onboarding experience, really, of setting up Podline for their podcasts, I've been considering a wide glove service approach. And this idea comes from Paul Graham's essay on doing things that don't scale, which is an essay that I highly recommend, even though it's kind of more aimed at the VC funded founder. You just do things knowing that you're not going to be profitable because you have some kind of financial backing.
But it still works for bootstrappers, because in the end, in the beginning, we don't have anybody as a customer anyway. So we might just as well put in the effort to lend a couple that may not be immediately profitable, but they guide the way. And that's kind of the point of wide glove as a service approach. You really get it. But I'm going already into something I wanted to talk about later.
So let's just think about the idea of wide glove service being something that involves helping people by just taking over the task that they would normally have to do themselves, using your product in the early stages of your business. Because I work in the podcast industry, I know what it takes to build and set up a podcast.
And it's not just about the technical aspect, like the mic or the room audio treatment or even the content creation aspects, like writing a script or actually recording the whole thing. You also need to establish a business ecosystem around the podcast itself. This includes having a place where people can go to learn more about your podcast or look into older episodes and subscribe if they enjoyed it in the first place. You need a website. You need a place for people to go. You need a domain.
And you need to bundle it up all into something you can mention on your show, like the fact that you could just go to tbf.fm, the WhatsApp founder.fm, and you would go to the podcast landing page for this podcast. That's important to have, because tbf.fm is easy to say and easy for people to type. And it needs to exist for anybody who wants to know about the show to learn more.
And many podcast hosting services, like Transistor or Anchor or all the other competitors, offer this option just to set up a website. And I've had experience running two podcasts before, so I know how these things work, how you set it up. And I just realized just how crucial this is, because it's vital for me to understand what my customers expect from a podcast landing page and any potential integrations to be able to render a white glove service. I need to know what it looks like.
And then it kind of dawned on me, if you've never experienced white glove service before, how can you provide it? This seems to be a common problem in the customer service industry I recently learned. Many customer service hires just don't know what this kind of service looks like, because they've never experienced it before. And many founders don't know about that either. And I thought, just in preparation for today's episode, this is a thought worth exploring.
Some founders, they just see a problem. They create a software solution, like an MVP with all the fancy graphs and dashboards, but they just don't understand how this fits into their users workflow or the larger context of their work. They see the technical implementation, but they don't see the contextual needs of their customers, their prospective customers. And this kind of white glove service works best, or even just at all, when you understand the full job to be done.
For a podcast host, that job is to create a space where people can chat, find out more, and engage with the podcast. That's the job to be done. The goal isn't just to set up a voicemail system on a website or integrate Podline into your HTML, in my case. That's a part of it. But it's to build stronger bonds with the community and make it easy for people to reach out. And that is a reframing that, if you are so deep in the trenches of building a software product, you might lose.
You might lose this kind of bigger goal that people might have in using a product. And to offer early onboarding services and create just great experiences for initial customers, you need to understand more than just the steps to integrate your product. You need to step away from the technical. You also need to help them contextualize it within their existing job. It's becoming like a processual thing.
You need to understand the process and how it's being described, what the words are, so you can explain it to your customers how and where this fits. Onboarding in SaaS should be tailored to more than just the SaaS. Often, using software products, which are software products of all kinds, is mostly-- you start with onboarding. You have this onboarding screen, and it shows you, click here for this, click here for that. Then something pops up. Enter this now and try to play around with it.
The onboarding flow is kind of an in-product thing. But that's just learning about the product itself. It doesn't say anything about the context. It needs to go beyond that. Instead of walking people just through features in 10 steps, focus on integrating the product into their existing workflows. I know this is hard, and it's easily said and done in a way that is very unique and specific to each product. But that's kind of where the white glove service approach comes in.
It means stepping back from your product and looking at people's tasks and goals, and then projecting your product and all its parts into those tasks and goals. And in every industry, this looks different. So I can't really give you a tip here. But what I would do is, if you don't know the specific job to be done that your product serves, you need to ask and figure out why your product fits in this bigger picture. You need to talk to people.
And again, I guess one of the things that we lose if we're so technical, and we love building, building, building, as indie hackers, where we love the code, we love the setup, the DevOps, the backend, the automations. But very often, we do not really get to see the specifics of the process in which we help people. So you can't just really expect people to do this work for you. You need to do it for them, really. And there might already be a solution in place for their needs, very likely.
Because if they have a well-defined job to be done-- and by well-defined, I don't mean that they see it as a job to be done. I think this framework is something that not everybody understands. But they certainly understand that they have a task that they need to accomplish. And if that task is well outlined, then there is a solution often in place. Most of the time, it's just doing it manually, which is the part where SAS comes in. And it automates it away.
But there could also be another software tool, or this amalgamation of different software tools. The little node here, an Excel sheet there. Maybe there's a database with some really, really basic wrapper around it that somebody in the company built five years ago. All these things exist. And the existence of these things means that you have something that is already there, that already takes data and already produces data.
And you have to understand what you're replacing in terms of a tool or who you're replacing in terms of the person that might be doing it manually. So figure out who they report to and what tools they use. That's one of the first steps that I would do. Figure out what the people that depend on them expect from them. And on the other side, I guess on the input side, what the data is that they depend on to do their job to be done.
Learn about any automations that they have for their input and output data, what comes in for their job and what needs to be produced during their job. And figure out what is handed off to them and what they will hand off to somebody else with a different task. Maybe if you really want to get it, understand the jobs to be done for these people that are doing the input for your business and the output after people use your product as well.
Understanding these things is pretty crucial to really get where the workflow is that you're in. And if you get these things right, your early customers will build massive trust with you that can jumpstart further growth just from word of mouth because you really knew what you were talking about. That's the idea. You want to be so deeply versed in the issues that these people have that they see you as one of them.
And you talk so much more easily in a beneficial way about the people that think like you, that act like you, that use the same vocabulary you do, that really understand your process just like you do, that even may have the same motivation to work in this field as you do.
And if that happens, your early initial customers that you help more than maybe your later more automated customers, the low touch customers you hope to get in your SaaS in the future, these early people, they will become evangelists for you as the person doing it and the business that you're building. So White Glove Service is all about knowing these little details and really truly understanding them and then guiding your customers through this experience in a way that leaves them just wowed.
That's what you want. All right. That is onboarding. Let me share a quick built-in public update that is both Podline related very strongly, I guess, and goes beyond it. Because I've been working on something over the last couple of weeks, and I'm very excited to talk about this. I shared a couple of things on Twitter about this already, but it's really a marketing effort for Podline because I said that I needed to find podcasts that already are established.
That's one of the ideas behind Podline is if you have an audience already, it's so much more useful to have a voicemail system instead of way still fighting for your initial couple listeners. It is unlikely that you get a lot of them to communicate with you because most people just don't really want to chat. So the bigger the audience, the more likely it is that voicemails make sense. And the more established a podcast is, the stronger the community.
And the stronger a community, the more likely is that in the community, people will communicate with each other and want to communicate with the hosts. Again, assumptions, but pretty strongly proven because the more I see older podcasts, the stronger their connection with their communities just from listening to these shows. I listen to a couple of podcasts that are like 10, 15 years old, which is kind of crazy if you think about it. They are in four figures terms of episodes.
And the community involvement in these shows, mostly around literature, I guess, or science fiction or fantasy and stuff, it's pretty strong. People don't lose this kind of hobby. So there is a very strong community there. So what I want is something that helps me find these things outside of my own sphere of knowing. I know a lot of podcasts, but I certainly don't know all of them.
And I wanted to build a tool that automates that for me because I'm a software engineer and I will build automations if I can. That's really what it is for my marketing. A kind of lead generator for podcasts that exist out there that are active and have a community. And they talk to their community in a certain way, and they talk about certain things on their show.
So in having built a podcast tool with Podline that kind of transcribes and summarizes voice memos, which is just a couple of minutes worth of audio, I kind of have the expertise at this point to build something that can transcribe and summarize a full podcast episode, right? Because it's just a slightly longer voice memo where somebody talks about a certain topic and mentions certain words.
And what I wanted to build was a tool that effectively-- and this is kind of the delusion of grandeur that I had here-- scans all the podcasts everywhere, transcribes every single podcast that has been released out there immediately as much as I can, and then searches those transcripts for certain phrases, for mentions of competitor products, or for phrases like send us a voicemail, or leave a message here, or that kind of stuff, where I could figure out, are there
podcasts out there that already have the need to communicate with their audience? And if that's the case, well, where are they? Who is talking? Where are they located? What's their website? And in podcasts, there are lots of repositories out there, Apple Podcasts and Spotify. Every podcast has an image, a podcast thumbnail, or something like this. And that thumbnail has a color scheme.
So I can just pull that thumbnail if I know that the podcast talks about topics that I care about and automatically create a potline theme, and then automatically create a potline for these things and just mail it to them. That's the idea, right? The idea for this lead generator is to find podcasts that talk about things that I care about, then automatically give me all the data that I can quickly build something for them. The freemium version of potline is just very easily used.
I sent them an invite link. They claim their podcast, and boom, they immediately have their voice memo. Everything is set up already. That's the idea. So I built this tool. And in building this, I was surprised by how much I can do on my own computer. So I don't know how deep I should dive into the technology of transcription right now. But transcription used to be around for ages. And I guess tools like Descript have been around for many years.
Otter.ai is one of the tools that most recently came out that I've been using a lot. There are a lot of podcasts or just voice transcript tools. I think Zoom has a lot of tools that just transcribe as you speak in a meeting. It's really cool. And the technology that lies underneath this used to be a bit more arcane. I didn't really understand what it was. Until a couple of years ago, OpenAI released Whisper, which is their pre-trained model.
It trained on millions of hours of people talking and having that transcribed. And it exists in multiple different formats. There's a small, there's a medium, there's a tiny, and there's a large version of this, each of which runs with a particular kind of speed. And it's an LLM, kind of, like an AI model that you can run on your own computer if you have a GPU. And there is a version that runs on your CPU that is kind of also sped up in a certain way. That's pretty fast.
On my Mac, I have a Mac Studio M1, I think an M1 Max. So that's kind of the next to last generation of Apple hardware, right? M3 is around the corner, M3 Ultra is around the corner. So this is going to be even crazier. So this is pretty much not the most recent Mac, but a fairly recent Mac. I installed Whisper.cpp, which is the CPU-bound Whisper inference model. So you can use your own computer to run this Whisper model, the AI, and give it an audio file and gives you a transcript.
The benefit of Whisper.cpp is if you compile it in a way that you use the, what is it called, Core ML or something, like the Apple's own machine learning architecture. There are chips in this, like the neural engine or whatever they call it, chips that are effectively GPUs, machine learning chips.
If you compile it, plus with the software that is around it and all that kind of stuff, in a way that uses all these optimizations, you can-- and listen to this, because it's crazy-- transcribe a 60-minute audio file. 60 minutes, right? 60 minutes, that's 3,660 seconds in 40 seconds. That is bizarre, because there used to be a time where even just real-time audio transcription transcribing one minute of audio in one minute of computer time was hard. And now you get an hour in under a minute.
And once I figured this out and I ran my tests and it actually did this thing-- and depending on which model you choose, you get either a very good transcription that takes a bit longer, or you get a 80% good enough transcription where a couple words may be off, but you get it done in half the time. So you can optimize for that depending on what you need is. I thought, if I can do an hour in a minute, that means I can do 60 podcasts an hour.
And that means that I can do over 1,400 podcasts in a day. And that blew my mind, because as much as I know that there are a lot of podcasts out there being released, I knew that it can't be that many. At least, I thought it couldn't. So I checked it out. I looked into the stats.
And there's a website called, I think, podcastindex.org, which has a wonderful API that provides you with trending podcasts, the most recent trending podcasts that a lot of people download, the most recently published episodes of podcasts that you can download. It has an API that kind of oversees the whole podcasting field. They also have a stats endpoint.
And they give you the number of podcasts that are being released over the last three days, that have been released over the last three, the last six, the last 90 and 30, and all that kind of stuff. And think over the last three days, 30,000 podcasts have pushed an episode. That is a lot of podcasts. But you have to think about the fact that on my one computer, I could do 1.5 thousand per day. So if you have 30,000 in three days, that's an average 10,000 in a day.
Half of them aren't even English. So that's 5,000, 6,000 maybe in English in a day. So my 1.5 thousand, so my 1,500 out of 6,000 on my one computer means that I need like three more MacStudio computers to be able to accurately and in almost sub-real time transcribe every single podcast that is being released as it goes out. And that blew my mind. Because I can literally figure out which podcasts over the last two hours have talked about a certain topic.
And that reminded me of something, which is why I know it's part-line related, right? Because this is being built as a marketing tool to find podcasts that talk about the thing. But I was like, hey, this is something that probably other people might also be interested in. So now I'm at a point where I'm really struggling. Because I love part-line. I love the idea of building community and voice messaging with another podcast.
And in many ways, the tool itself, the part-line tool, is technically feature complete. Obviously, it's not completely bug free. I'm still working on that kind of stuff. But the voice messaging works, the back and forth, the notifications work, the upload works, the transcription works, the summarization work, all of these things are in the bag. So feature-wise, it's kind of at least feature parity with the existing things in the world, the competing products.
I've built way more than most of the competitors' products offer in terms of theming and in terms of customization and that kind of stuff. Also, transcription and summarization are a thing that most of them do not support. So that one is feature complete, which kind of allows me to spend a lot of time on this new kind of podcast scanning thing that I'm building. And it's kind of tearing me apart because I want to focus on marketing part-line.
But this tool itself is such an interesting marketing tool that I might want to build this into a business.
So maybe just really, if you're listening right now and you understand the potential of a tool that-- if I build it right, if I find a solution to kind of scale this in a financially feasible way as an indie hacker without having to buy 10 more Mac Studio computers just to be able to scan all the podcasts, but put it in the cloud, pay for some GPU-based Amazon instance or whatever, and run it there and reliably scan every podcast everywhere, if I would do this, what do you think?
Do you think this is too much of a deviation from the product that I'm currently working on? Do you think I should look into this more as a marketing tool, kind of a Google alert for podcasts where you can put your phrase or your brand's name in there and it can report it back to you like minutes after a podcast is released anywhere in the world that you're not even aware of? What do you think? Is that a good idea or is it not? Because I'm kind of struggling with-- I want to focus.
I don't want to dilute my focus. The question is, what should I be focusing on? So yeah, again, indie hacking is sometimes such a hilarious thing. You think things are not even possible, then you build them. You know this is possible. And then all of a sudden, this world of potential, you'd be standing right in the middle of it. You didn't expect to go there. You didn't want to go for the trip. You just wanted to build something cool. And now you're wondering, should I do this instead?
Man, it's such a blast and such a struggle at all times. Just wanted to share this. I hope this is interesting. If you want to let me know about this, let's just use the same way as we always do. Go to podline.fm/arvid, A-R-V-I-D, and just tell me about it. If you send me a message that is funny enough, I'm going to put it on the show next week. We can talk back and forth. That'd be fun. But yeah, that's where I'm at. So I guess that's it for today.
I'm just going to briefly thank my sponsor, Aquia.com, right now, because I regularly check the products that are listed there, because it's just Revel in the side of profitable businesses and that are looking for new owners, pretty much. And it's really always a surprise just how niche you can get and still build something worth like five, six, seven, or eight figures that other people would love to own. I see a lot of products in the podcast space.
There's also dog magazines and e-commerce things. There's everything there, which is really cool. I just really get inspired just even to follow my own choices by seeing how other people followed their own ideas of what they wanted to build and then built that and turned into something meaningful. And the more people I talk to, the more I see that they just all sell for different reasons. Some are just done with that particular business.
Others want to refocus on something new, which is kind of my problem right now. And then there are some that just need a quick cash infusion, turn what they have into money. And that's kind of what unites them. They all build something for themselves and now want to turn it into financial stability. That's the idea. So if you're on track to getting to that point or you're already there-- congratulations, I guess-- go check out acquire.com. I think it's a perfect place to just check it out.
You don't need to make a choice, but checking it out definitely is an idea. Their team is super helpful. They make sure you're getting a really good price for your business, and they know what they're doing. But talk about White Glove Service. This is a team of people that have your best interest in mind because the more you sell for, the better for the community and further people that then later want to sell for. Hey, if you're interested, go to try.acquire.com/arvid and check it out.
It doesn't hurt to be prepared for that happy acquisition day. And yeah, just see if it is for you and your project at this point in time. Thank you for listening to The Bootstrap Founder today. That was a fun little conversation. You can find me on Twitter @arvidcarl, A-R-V-A-D-K-A-H-L. And you'll find my books and my Twitter course there too. If you want to support me and the show, please subscribe to my YouTube channel.
Get this podcast in your player of choice and leave a five-star rating and a review by going to ratethispodcast.com/founder, which is also an indie project, ratethispodcast.com. It's always kind of cool to support your fellow indies. Makes a massive difference if you show up there and give me a five-star rating and a review because this will make the podcast show up in other people's feeds. It will make it interesting to them.
And everybody benefits from learning more about being a business builder and bootstrapping businesses and creating something worth creating. So any of this helps the show, so feel free to do it. Obviously, no obligation. I would just really, really enjoy it. Thank you so much for listening. Have a wonderful day and bye-bye.