#30 Selling your startup and buying it back with Steve Ridout

[00:00:00] It was out for nine months completely for free. And then I turned on pricing at $10 a year, I think it was.

[00:00:08] And I was so happy that waking up the next morning and getting an email that said somebody had paid for it.

[00:00:14] And then it looked like a bug because it was a thread and there was a duplicate email after it.

[00:00:20] And I was like, oh, something went wrong. But then I realized, no, it was two people.

[00:00:24] Two people had paid and I was so happy. That was great.

[00:00:30] Hi, and welcome to the Code and Conquer podcast. In this episode, we're talking to Steve Ridout,

[00:00:35] founder of Readlang, who turned his own struggle with learning Spanish into a successful language

[00:00:40] learning product. We dive into his journey of building Readlang from a simple translation tool

[00:00:45] to a 10K MRR business, including its unexpected detour through Duolingo, where Steve not only

[00:00:51] sold his company but also helped create Duolingo Stories, before buying Readlang back at a quarter

[00:00:57] of our original price. Other topics include how Steve achieves growth purely through word of mouth,

[00:01:03] his experience with freemium pricing from $10 a year to $48 a year, and how he's managing to thrive in

[00:01:10] an increasingly AI-dominated language learning space. We also talk about the challenges of being a solo

[00:01:16] founder, when to focus on product versus marketing, and the surprising observation that his best revenue

[00:01:21] months coincide with when he works the least. I hope you enjoy this one. Let's jump right in.

[00:01:30] And welcome to the episode 30 of the Code and Conquer podcast, the next big number. We finally reached it.

[00:01:37] And the guest today is Steve Rideout. He comes here with an indie hacker product called Readlang.

[00:01:44] Steve, how did you come up with the idea? What is Readlang? Tell us a little bit about yourself.

[00:01:52] Okay. Hi, Tobias. Thanks for inviting me. So Readlang started 12 years ago now. I'd moved to Spain to live,

[00:02:02] maybe a year previous to that. I was doing some contracting at first, and then my plan was to make some

[00:02:12] kind of online business of my own. I'd tinkered around like for one month with some other idea to teach

[00:02:19] people the technical aspects of photography, like a camera simulator. But after a month, I decided that this

[00:02:26] wasn't going anywhere because it was the kind of thing that I think I would use once I'd learn what

[00:02:34] I needed to learn, and then why would I come back? And at the same time, I had this real problem that I

[00:02:40] didn't speak Spanish hardly at all. I was trying to read stuff because I thought that would be a more

[00:02:47] interesting way of learning the language. And I couldn't find a good way on my iPad to get texts

[00:02:56] there to read that I wanted to read and to translate them quickly. I'm sure it's much better now,

[00:03:01] like the default experience in any books app is probably pretty good and on a Kindle. But back

[00:03:07] then, it just felt very slow and cumbersome to get translations of words. So I just hacked together

[00:03:14] something really quickly, like just one HTML page that I could use myself just to test the idea of

[00:03:21] putting a text. And then you tap on any single word and it just uses Google Translate to translate

[00:03:27] that word. And so within a day, I just had something that wasn't that it was the bare minimum proof of

[00:03:35] concept. And with that, I decided, ah, this kind of feels nice. Like it feels better than anything else

[00:03:43] that I could find on the iPad. So maybe there's something to this idea. And that was in November,

[00:03:49] 2012. I can see because of the very first git commit was back in November, 2012.

[00:03:56] And that's a very long time ago. The first time I actually looked at your product,

[00:03:59] I thought that you are like one of the founders who used the whole AI hype wave

[00:04:05] to then create something because there are a lot of teaching languages, AI rappers and whatever. But

[00:04:12] apparently that product goes way further down the road with that history. I was just wondering,

[00:04:19] because we're going to talk about your career and how it changed a lot with Readlang a little later,

[00:04:23] but apparently a lot of indie hackers moved to Spain. I had a lot of Spanish or at least in Spanish,

[00:04:29] in Spain living indie hackers on here. What was your situation like when you started creating Readlang?

[00:04:36] You already said you were trying to do the freelancer route. Was that straight out of college?

[00:04:42] Or how was your way?

[00:04:43] No, I'd worked at a university in London for a bit as a researcher. And then I went to work at a startup

[00:04:51] in London called Mendeley, which makes software for academic researchers to manage their library of PDF

[00:04:58] papers. So I did that for a couple of years. And then I met my girlfriend there who's half Spanish.

[00:05:07] And she wanted to finish a PhD that she was doing in Madrid. So I tagged along thinking,

[00:05:13] this would be a cool adventure to have to live in another country. Like I'd lived with people in London

[00:05:18] who were having that adventure, like from all different parts of the world who'd moved to London.

[00:05:24] And I thought that would be a cool thing to try. I was up for quitting my job and trying to do

[00:05:30] something of my own. And then it just so happened that Mendeley were looking for someone to work as

[00:05:38] a contractor to do this one job, to build an editor, a web editor. And I wasn't even a web developer.

[00:05:45] I was a desktop, like C plus I was doing, but there was this, they had this deal. What was it? Like a

[00:05:53] sponsorship or something from the Sloan Foundation, I think it was in the US. And yeah. And I was like,

[00:06:00] in a, I knew I wasn't a web developer, but I knew a lot around the topic. So they got me to do that.

[00:06:09] And so that funded me for the first few months living in Madrid. I worked remotely doing contracting for them.

[00:06:15] And was that also like how you financed the first few months of Readlang then? Correct? You already

[00:06:21] started there. It was quite lucky because it was perfect because it even paid a lot more than they

[00:06:26] were paying me as a full-time employee. So it's like living in a cheaper country. And then I'd already

[00:06:32] saved up a bit before and this just added. So I had a good buffer and I was used to living very frugally.

[00:06:38] I lived all through my twenties, like very cheaply in London. And so I didn't feel scared of spending a

[00:06:46] couple of years working on something, even if it didn't amount to anything. I felt I've got enough

[00:06:51] of a cushion and I could easily go and get a job if I needed to. So yeah, that helped fund the

[00:06:57] beginning of Readlang basically. And when you started, that was 2012, right? That was a whole

[00:07:03] different landscape to how Indie hacking or bootstrapping probably was the name in 2012,

[00:07:09] right? That was a whole different landscape. How did the first few months or even, I know that

[00:07:15] the first few years were hard money-wise for Readlang. How was that all coming together at the start?

[00:07:20] Yeah, it was, it didn't feel hard because it was just an exciting adventure for me. I was listening to

[00:07:27] things like, there was some podcasts that there was a podcast by Marco Arment, the guy who made

[00:07:35] Instapaper and now Overcast, the podcast app. He was talking about Instapaper in a podcast back then. And

[00:07:41] he'd, he was like an indie hacker making his own iOS app. And so that was inspiring. And reading things like

[00:07:48] Paul Graham's essays, learning to make something users want and all this stuff. And Joel Spolsky as well,

[00:07:56] and all things like that. And yeah, so I just tried to get something out there as soon as I could, even

[00:08:05] before it felt fully baked, just to get feedback from people. And yeah, like I said, I made a thing for

[00:08:11] myself in a day. And then it took, I think it took about three months before I had something ready, which had

[00:08:18] like user authentication and I could share it on Reddit, like that, that I felt comfortable that other

[00:08:25] people could use. It took me a while because I'd never, that project that I did for Mendeley, my first

[00:08:32] web project was only front end. And this was the first time I'd used any backend stuff. So I was

[00:08:37] learning as I went, I picked MongoDB and Node.js, partly because Trello was using it. And I was following

[00:08:43] Joel Spolsky and Jeff Atwood. I don't know if Jeff Atwood was involved with that, but anyway,

[00:08:49] it was, I thought, oh, if it's good enough for them, it'll be good enough for me. And it hasn't

[00:08:55] caused me any serious issues, but if I was picking today, I'd probably pick Postgres.

[00:09:00] Oh, as a database, because I was thinking as 2012 was pretty early for Node.js, right? I don't

[00:09:06] want to go into the tech details too much, but that was, I think, still a risky call, right?

[00:09:12] Maybe. It was early, but like I said, I think Trello was using Node.js. I would have to double

[00:09:18] check that. Yeah, I guess it was a bit early, but I liked the idea and I still like the idea

[00:09:23] of using the same language on the front end and the back end, of being able to share code

[00:09:27] if I want. And just to, there's enough other things that you have to worry about in running

[00:09:32] a business. And like just being an expert in one programming language instead of two sounds

[00:09:38] like a win. Yeah. And also I think that Postgres has pretty much been a good pick for 20 years at

[00:09:45] this point. I'm not sure how long they even exist. And I've always used Postgres as a database. A lot

[00:09:51] of things have changed in the whole landscape, but there are some technologies where if I use this,

[00:09:57] it will just work. And that's okay for, especially for indie hacker projects where you're like,

[00:10:02] I have so many other things I need to like think about. Having a boring stack is actually a very

[00:10:07] good thing.

[00:10:09] A hundred percent agree. And I think it was maybe, I, maybe I got lucky that Node.js continues to be

[00:10:16] very popular, but it hadn't been popular for that long. So maybe in retrospect, it was a risky call.

[00:10:22] Like I would recommend people pick a boring stack that's been in use for 10 years.

[00:10:29] In 2012, I still wrote Java code. I think it's good idea to not have gone with Java,

[00:10:35] even though I haven't used it in 10 years now, but yeah, maybe not. I think you went the right track

[00:10:40] down. I think that was a good idea. You already mentioned, like the next question I wanted to ask

[00:10:45] was something like, how did you start off? How did you market it in the first like months or weeks?

[00:10:51] But you already mentioned Reddit. Was that like the way to go for you? And what was the feedback

[00:10:56] like there?

[00:10:57] Actually, now that I think about it, I jumped the gun a bit because just before sharing it on Reddit,

[00:11:02] there was another really niche forum called howtolearnanylanguage.com. I don't know if it

[00:11:08] was very specific.

[00:11:09] Yeah, it was super specific. So that was the first place I shared it. And it was really cool because

[00:11:17] I got, I can't remember the numbers, maybe in the order of 20 to 50 users or something,

[00:11:23] signups. And a bunch of people, a bunch of discussion on there, like some early feedback

[00:11:28] telling me that they thought it was cool on the whole and giving me some constructive feedback.

[00:11:34] And then it was a bit later that I shared it on Reddit. Like I said, I think it was three months

[00:11:41] after I could probably research, I won't research it now. It'd be interesting to go and see those early

[00:11:46] threads.

[00:11:47] But Reddit is always like the place where every indie hacker you talk to me pretty much is,

[00:11:52] I posted on Reddit and got roasted to.

[00:11:54] Oh, really?

[00:11:56] Did you have another experience there?

[00:11:58] No, mine was positive. It was on the language learning subreddit.

[00:12:04] And I think you've got to come in with some humility and just be honest and say, I've spent

[00:12:11] the last three months working on this to help me learn a language the way I'd like to learn it.

[00:12:16] And I would love if you would try it out and leave some feedback that would be really useful,

[00:12:21] something like that. And I guess they could see that I put some work into it. And I don't know,

[00:12:26] I got a good response. I'm not sure if things have changed, if it's maybe become more difficult

[00:12:31] over the years.

[00:12:33] I think it's just that people are very critical. I just talked to Lukas Hermann,

[00:12:38] people that listened to this episode will have just listened, hopefully, to the one with Lukas,

[00:12:42] where we just mentioned that Reddit and also Hacker News can be a pretty hard place to promote

[00:12:49] pretty much anything because all people go, oh, I could build this on the weekend. I don't need your

[00:12:54] product. I could build this myself in like Friday afternoon. And I think that's something that

[00:12:58] happens a lot of times when you go there.

[00:13:00] Yeah, that's true. There's the classic one for Dropbox on Hacker News, like the top comment was

[00:13:06] like, who needs this? It's so easy to set up a shared folder with Linux or something or other.

[00:13:11] Yeah, for some reason, not everyone does that. I'm not sure why, but I always build like all my

[00:13:15] software myself. Yeah, exactly.

[00:13:18] Exactly. So in your progression from how Readlang started, we have some kind of career changes that

[00:13:25] you did. I want to go there, but maybe talk once more about the whole stage where you were on your own.

[00:13:31] So you still did the work for Mendeley or did that stop at some point?

[00:13:36] No, I pretty much did one thing at a time. It was a fixed project. And once it was over,

[00:13:41] and I'm generally not very good at trying to do multiple things at the same time.

[00:13:47] I prefer to get obsessed with something and just work full on on that thing.

[00:13:52] So basically, I think that Mendeley had been over like a couple of months when I started Readlang.

[00:14:00] Like I said, I tried that other camera thing in the meantime. And then I worked for the next three

[00:14:05] years, except there was a break where in 2014. So at that point, after I'd worked on Readlang a couple

[00:14:14] of years, another thing did come up on Mendeley. And I worked there just for a few months, like maybe

[00:14:20] three months over the summer. And then I went back to Readlang again. I can't remember how I think at

[00:14:28] that point, Readlang was probably making like about $500 a month or something like that.

[00:14:33] That's why I was asking because as I've read through some of your blog posts, and one mentioned that like

[00:14:39] for the first three years, you basically got less money out of it than from what's it called an

[00:14:45] entry level job, minimum wage job, exactly. So I was just wondering how you like, because that's

[00:14:52] like my fear scenario. So I'm wondering how you got around and not worry all the time about the

[00:14:59] growth not being enough or not growing fast enough stuff like that.

[00:15:04] Yeah, the good thing was that even though it was slow, there were always signs that it was

[00:15:10] growing. So that gave me the motivation to continue. But yeah, I guess at that point,

[00:15:18] I was happy to go and earn some extra money and approve to myself that I didn't, I wasn't desperate

[00:15:24] because I still had some savings. It was nice to know that I couldn't continue doing that just in

[00:15:32] case I did need to go back to employment or contracting in the future. I thought that would be

[00:15:38] a prudent thing to do to keep to leave that door open. But then after I went back to Readlang,

[00:15:46] I kept working on it. And it grew to about 3000 a month by the time of March 2016, which was when

[00:15:58] Duolingo got in touch with me.

[00:16:00] The big change in the trajectory of both your career and Readlang, right? Where you then took,

[00:16:06] I think a full time job with Duolingo, right? Why did you decide to take that job? Why not just

[00:16:12] continue with Readlang? What was the story there?

[00:16:14] It was way good enough to get an opportunity to turn down during all this time. I was excited to

[00:16:21] work on my own thing. But I was also really curious about what it would be like to work at a really

[00:16:29] well-run tech company, like something like Google or one of these big ones that weren't as big as they

[00:16:36] are now. And Duolingo was a really impressive product. I tried it, I think it was just after

[00:16:43] I started Readlang, tried Duolingo and I thought, oh, this is really cool. This is fun. So I just

[00:16:47] thought they were the most impressive company in the space. Really cool team. The guy who,

[00:16:53] the CEO, was an inspiring guy as well. He previously sold two companies to Google and I'd seen his,

[00:17:01] some of his talks and I just, and they offered like, for me, what was to me a really good salary

[00:17:08] because they were offering me like an American Silicon Valley style salary.

[00:17:14] After coming from working in the UK, it was, yeah, it's too good to turn down. It was very good.

[00:17:21] And yet I'd read enough online. I think Patrick McKenzie, patio 11 had written this whole blog,

[00:17:29] long blog post about how you should negotiate. I thought, I can't just accept I've got to at least

[00:17:35] ask for a bit more, even though it was already like way more than I, than I would have accepted

[00:17:40] for less. But I, so I got a little bit more even and it was really, but it was a really awesome

[00:17:46] experience. They were really, it was really, it was five years I worked for them and it was a really

[00:17:52] good team. I really enjoyed working for them. And I think I read that you actually at one point

[00:17:57] sold Readlang to Duolingo and also back again, you bought it back at some other point, but how did

[00:18:05] it come to, to be that you then started working for them and even sold them your product essentially?

[00:18:13] Yeah. So what it was, we talked about me selling to them at first from our discussions. It seemed like

[00:18:21] they didn't have big plans to do anything with Readlang. They were more interested in me.

[00:18:26] And so I thought this is a bit of a risk if I sell it to them and this, the sale would have come with

[00:18:32] a condition of needing to work for them for X years to get all the money and stock options.

[00:18:38] And so I thought, why don't I just work for you for a bit? And then it's low risk. I can walk away

[00:18:46] if things aren't going well. And yeah, that just seemed like a sensible thing to do. And then I think

[00:18:53] it was about nine months in, I'd been working on this project just as an engineer. It was a chat bot,

[00:19:00] which was in retrospect, a bit too early. The tech hadn't quite gotten good enough to pull off a

[00:19:07] convincing chat bot. And I wanted to switch projects after a while, or actually I talked about just

[00:19:15] going back to Readlang, like after nine months. And then they said, I don't leave. I just start

[00:19:21] something new. We want to do something around listening practice. So I hacked away. I thought,

[00:19:29] oh, this would be cool. I'll just, yeah, you pay me and I'll just hack around on a new project.

[00:19:34] That sounds fun. So I came up with what became Duolingo Stories. It's like a feature of the app

[00:19:40] where you have these around two minute audio stories with a transcript synchronized and some

[00:19:48] challenges interspersed. And then we launched that. And then again, I was like, all right,

[00:19:55] I've done that. That's cool. I think I'll go back to Readlang maybe now. Oh, what was it? That? No,

[00:20:01] hang on. It was, yeah, I wasn't sure what to do. I was like, that was my plan was to go back to

[00:20:08] Readlang. But the launch went well. And so it was exciting. And at that point we talked again about,

[00:20:15] how about we, you sell Readlang to us now. And yeah. So we talked about that again. We agreed on

[00:20:22] a figure and I carried on working for them and grew the Stories feature within Duolingo.

[00:20:29] But it wasn't the plan that you would then work on Readlang while being on Duolingo, right?

[00:20:34] No, no. It was basically what I figured was I'll work on Stories for a while and then I'll maintain

[00:20:40] Readlang, keep it running. And at some point my plan was to do something with Readlang in Duolingo

[00:20:48] to like maybe integrate it or try to grow it. But every time I thought about that idea,

[00:20:56] about pitching that, it never seemed to make, it seemed really risky. It seemed like this is so tiny.

[00:21:04] If I say I want to work on this and build a team around it, it's going to have to grow so big to be

[00:21:10] interesting to them. Otherwise, the chances are is it'll just be shut down or abandoned. And it felt

[00:21:18] like an uphill battle to try to do that. And especially since the code was old and a bit hacky,

[00:21:26] it would have been felt like I need time just to refactor it. And then if you're just working for

[00:21:34] yourself, you can just do that. But if you're working for a boss and a company, it's like they

[00:21:41] would understand, I guess, but you wouldn't be showing much visible progress.

[00:21:44] But I probably like, why is he working on something that makes

[00:21:47] exactly how much money instead of doing something where they make a lot of money instead?

[00:21:54] Yeah, it wasn't. They were more focused on user growth and stuff like that rather than directly

[00:22:00] about money at that point. But yes, exactly. It would have needed to grow a ton. And I'm not sure

[00:22:10] if it would have. It's much more niche. It's for people who are further along in their learning

[00:22:15] journey than Duolingo. Duolingo attracts people who like day one, like beginner language learners,

[00:22:22] and they aspire to teach further. And they get into the point where they can be interesting for

[00:22:28] people at B1 level, say. But Readlang, I think it's most useful when you are at B1 level or above,

[00:22:35] probably. Maybe A2 as well, if you get the right texts to read. But it's pyramid where there's tons of

[00:22:45] beginners at the bottom, but the further you go up, the less there are. So I think it's difficult.

[00:22:50] Interesting, right?

[00:22:51] If they're just looking at numbers, it wouldn't be that interesting. There's other ways you could

[00:22:54] look at it though. It still may be interesting for them, even if the numbers are less, to have a

[00:23:00] really good product in there that appeals to experienced people.

[00:23:04] But then you didn't, like the product was then sold to them and just stayed there at Duolingo with

[00:23:11] you. But at some point, you apparently didn't want to keep working for Duolingo at any point.

[00:23:18] And then there came the discussion of getting Readlang out with you.

[00:23:23] Well, it didn't quite work exactly like that. Yeah, it's true. I wanted to leave at some point.

[00:23:30] Even though it is a great company and I really love people, but I just wanted to do my own thing again.

[00:23:37] And I didn't, it might've been different if I lived there and working with them. I did feel a bit

[00:23:43] disconnected too, because I was like remote in Madrid and most of the other people were in Pittsburgh.

[00:23:49] I used to enjoy my visits there, but I hadn't been in like over a year because of COVID.

[00:23:55] And yeah, so it was, they'd also had the IPO recently. So it was like a convenient time for me to leave.

[00:24:02] And I left and just left Readlang with them. But I said to them, if please, if you ever think about

[00:24:10] shutting it down or anything, please get in touch with me first. And I knew they weren't going to

[00:24:15] do anything with it because I was handing it over to the CTO who had way bigger things to worry about

[00:24:21] than my tiny little project. So it kept running without, I guess, much work on their part for a

[00:24:28] year or so. And then they started having problems with card testing fraud, which is where people

[00:24:38] who've got stolen credit card numbers want to test them on websites that offer very low priced purchases.

[00:24:47] So they were lots of, they were processing lots of these payments and I, people started emailing me,

[00:24:55] even though I wasn't connected to it anymore because Readlang was taking their money and they didn't

[00:25:01] know what it was. And so I got back in touch with Duolingo and I ended up helping them fix that problem.

[00:25:09] And they just asked me, do you want it back? And how much would you pay for it? So I ended up buying it

[00:25:16] back off them.

[00:25:17] I'm interested, just you don't have to give like certain total numbers, but was it like way more

[00:25:23] expensive to buy it back from them? Or did you basically get the price that you, that they paid

[00:25:28] you back?

[00:25:29] Oh, it was way cheaper to buy it back.

[00:25:31] Oh, okay. That's awesome.

[00:25:33] So yeah, it was a, I'd say about a quarter of the price to buy it back.

[00:25:39] But if you actually consider that when they paid me, it was a chunk of it was in stock options and

[00:25:47] those had gone up a lot. So really it's much less than a quarter. If you consider how much those

[00:25:53] had increased.

[00:25:54] Perfect deal for you and get getting the company back.

[00:25:57] Yeah, it was nice. And also I'd spent the last year doing tinkering around on some projects,

[00:26:03] but not really getting anywhere with them. It was nice to have something that had

[00:26:08] users. It was making a bit of revenue and it was something that I could work on

[00:26:15] and actually be benefiting people instead of just tinkering around on projects of my own

[00:26:20] that weren't, weren't getting much traction.

[00:26:23] And I think at the point that you bought it, that's at least what I saw in the blog post is

[00:26:26] that it was at around the 5,000 MRR mark.

[00:26:31] Yeah, it was a bit less. I think it was probably about 4k.

[00:26:35] Yeah. You then, I think we can celebrate together because you, you, you hit the 10k mark,

[00:26:40] I think in end of October, we're recording this in the middle of November.

[00:26:44] Yeah.

[00:26:44] So how was like, how long was that journey after you got it back and how did you

[00:26:48] manage to grow it that way?

[00:26:50] Yeah. So I got it back in February last year.

[00:26:53] Mm-hmm .

[00:26:54] So it's basically about a year and a half. And oh, just a tiny sort of caveat,

[00:27:00] which is that it made 10,000 revenue last month, but on the Stripe dashboard,

[00:27:07] it doesn't show the MRR as 10,000 because a lot of that revenue was yearly plans.

[00:27:13] Okay.

[00:27:13] Yeah.

[00:27:13] But anyway, it's headed there. It's increasing. And actually, if I look at the trailing 30 days

[00:27:18] from now, it's hit 11,000 in US dollars. So it's going up.

[00:27:24] I think it did count.

[00:27:25] Yeah. And so the cool thing is to get there and it's happened faster than I thought. I thought it

[00:27:33] would take a bit longer to get there. I've just been improving, fixing bugs and improving the product.

[00:27:39] I haven't done any marketing really. I've been actually responding to people which had fallen

[00:27:46] by the wayside while it was at Duolingo. And yeah, just making sure that the product runs well.

[00:27:53] I set up a forum a few months ago and I had a user voice forum before that, but I've got my own hosted

[00:27:59] forum now. So I've been interacting with people on that and adding features. I think the key features

[00:28:06] that have probably helped contribute to this AI powered now. The first one was AI explanations.

[00:28:17] So in the past, it would just be, you click on a word or drag across a phrase and just that word

[00:28:23] or phrase would be sent to Google translate to give a Google translate translation. But now you can get

[00:28:28] like a whole paragraph of explanation of what that means in the context it's in, in that sentence.

[00:28:35] And that's super useful. And then there's a bunch of other features. The most recent one,

[00:28:41] which again is pretty, makes a big difference to the product is instead of sending it to Google

[00:28:47] translate, if you're on the premium plan now, it can be sent to chat GPT with a prompt that now gives you

[00:28:55] the translation of what that word or phrase means to display above it. But again, taking the context

[00:29:00] sentence into account. So even without having to look at this lengthy explanation, you just get the

[00:29:04] quick translation. But this time you're almost guaranteed that it's going to be the right one for

[00:29:12] the sentence. And apparently you're living in the hacker stream where you just build features and people

[00:29:18] come. Exactly. Build it and they will come.

[00:29:21] But with the caveat that you have to have a user base first and a product that grows organically

[00:29:28] through word of mouth, and then you can live the indie hacker's dream. Yeah.

[00:29:32] But is it actually mostly word of mouth? It's funny because the last product that we had on was

[00:29:37] Stage Timer by Lucas, who also says that word of mouth is more than 50% of their growth. Is it the same

[00:29:45] for readlang in that scope that it's mostly people just telling each other?

[00:29:49] It depends how you classify it. I don't know if it's people telling each other in real life.

[00:29:57] To be honest, I'm not 100% sure how to break this down. Because in the early days,

[00:30:04] I would get a good sense of where the waves of traffic were coming from. Because I would see

[00:30:10] like a big spike happened and it turned out somebody written about it in Lifehacker or BuzzFeed,

[00:30:16] or there was some different online publications or like on Reddit, like there's a Reddit post and

[00:30:21] something. But now most people almost, I'm making these numbers up a bit, but it's 90%

[00:30:29] come from a Google search, which is just readlang. Just somebody searches for readlang and comes to me.

[00:30:34] So it's hard to know exactly where they came from. And now I've tracked down some of these waves to

[00:30:41] things like a TikTok video or an Instagram reel that somebody's made. I've got nothing to do with it.

[00:30:46] Somebody's just made it and it's in some other language. And so I get people talking about it online

[00:30:53] without me doing anything. So it's cool. And now I would love a way to grease the wheels and

[00:30:59] encourage more of this. And I'm not sure how to do that. I like it. Like I know people do influencer

[00:31:07] marketing where they pay people to do this, which could be worth exploring. I don't know. It could

[00:31:13] be worth talking to somebody who's done something like this. Yeah, I was just thinking that something

[00:31:17] like an affiliate program or even paying directly to the influencers you pick might be an idea. Yeah,

[00:31:24] maybe I did try an affiliate program before Duolingo and I got, I may have set it up for this

[00:31:32] collaboration. I can't remember the order exactly, but I got this guy called Benny Lewis who ran a pop.

[00:31:39] It was like the most popular blog of language learning at the time that I was aware of anyway.

[00:31:44] I know. I think it was called learning language in three months or something like that. And he

[00:31:48] featured it on his newsletter and it got like a wave of users to sign up. Not as big as I had

[00:31:56] from say Lifehacker, but a good amount. But then at the time my conversion rate between somebody signing

[00:32:05] up for free and then paying was about 1.8% of users. And the amount that actually paid was pretty small.

[00:32:14] And I got back to him and told him, oh, given the deal that we set up, I owe you this much.

[00:32:19] And he wasn't even interested in receiving the money because it was such a small amount.

[00:32:24] And so it didn't work basically. But that's, I guess that's a different kind of affiliate. You

[00:32:31] could think of a different kind of affiliate scheme that's more just peer to peer, like a user

[00:32:36] recommending it to their family and that kind of thing. I've considered actually everyone who signs up

[00:32:42] gets a gift coupon that they can just give a couple of months free to somebody else

[00:32:46] as a way to encourage viral growth. That could be a fun thing to experiment with.

[00:32:51] Yeah, we just talked about this in the other day on the podcast, doing something where

[00:32:55] the user gets something and also the one that they invite. So they get

[00:32:59] 10 free credits, they get a free month, they get something like that for recommending the app to

[00:33:05] somebody else. So it's like a game to them, right? You can maybe invite max 10 people and for every

[00:33:10] people you invite and they sign up, you get a free month, 10 credits, 10 tokens, whatever you're

[00:33:17] selling on top of it for yourself. And that seems to work for a lot of people. So that might be

[00:33:22] something worth considering.

[00:33:24] Yeah, maybe it worked well for Dropbox.

[00:33:26] It was an example.

[00:33:28] Yeah, exactly.

[00:33:29] True.

[00:33:30] But the one thing that I'm really, I think someone in your position who hits 10k a month

[00:33:37] needs to be commended a little bit because you're in a very hard B2C market, which is not our favorite,

[00:33:43] at least. And also your price points are really low. I got to talk, I got to let you talk to the

[00:33:50] other Steve that we had on Steve McCloud, who's always raised your prices. But I think you have a

[00:33:55] very hard combination to get something to the like this to 10k as your prices are zero to $15.

[00:34:02] And also in the B2C market. But what was your, but what's your thoughts on that?

[00:34:07] My initial thoughts, when I started back 12 years ago, I think there's even some comments from me

[00:34:13] on a forum saying that I didn't like the way people were pricing things that what, why can't it just be

[00:34:20] that everyone has to pay, but just a small amount. So my, like, why can't it just be like $2 a month

[00:34:26] or something? And obviously I've changed my tune because I think there's like such a difference

[00:34:33] between free and two, the difference between free and $2 a month is like way bigger than $2 versus $10

[00:34:43] a month. Like if you're going to get someone to pay $2 a month, a huge percentage of them are also going

[00:34:48] to just pay $10 a month. So I, but getting them from zero to two was, would probably be really hard.

[00:34:54] I think what I'm my thinking, I don't know. It's just, so I started with that frame of mind. Like,

[00:35:02] I think I launched it for free because I wanted feedback, but my, my thinking was everyone should,

[00:35:08] I didn't want it to be ad supported. So I figured like the free plan is just like a trial thing,

[00:35:16] just so you could try it out and have some confidence that the software works. And then

[00:35:21] almost everyone, if they want to use it should pay. That was my initial thinking, but I guess not,

[00:35:28] it didn't work out like that. And I don't know, but I just followed like what I saw other freemium

[00:35:33] people doing then. And now the biggest example in this space is Duolingo now, and they have grown

[00:35:42] as big as they have because they have a free plan. They didn't do marketing. They didn't do paid

[00:35:49] marketing for the first, I don't want to say almost 10 years, maybe I'm not sure the exact numbers,

[00:35:56] but, but they did have a really good social media game, which I don't have. So they did put some work

[00:36:02] into it, but they weren't paying for ads and they grew from, from being free.

[00:36:08] And I, so I, and they're doing really well. So it seems like a good approach.

[00:36:14] Have you ever experimented with the price? Did you do any A-B testing with the prices and also with

[00:36:21] the free tier? I know that from Mark Kohlbrugge, which I would love to have on the podcast, but he

[00:36:26] hasn't been, that he did an experiment where he just removed the free tier and saw how the conversion

[00:36:33] changes and stuff like that. Did you ever try to experiment with either pricing or removing the free tier?

[00:36:38] I haven't done any A-B tests on pricing, but I have gradually increased the price over the years.

[00:36:46] I think if I remember correctly, my first, so I, it was out for six to nine months completely for free.

[00:36:57] And then I turned on pricing at $10 a year, I think it was. And I was so happy that waking up

[00:37:05] the next morning and get an email that said somebody had paid for it. And then it looked like a bug

[00:37:11] because like, it was a thread and there was like a duplicate email after it. And I was like, oh,

[00:37:16] something went wrong. But then I realized, no, it was two people, two people had paid. And I was like,

[00:37:21] so happy. That was a great...

[00:37:23] It's just like the first dollar is the best one, right?

[00:37:25] Yeah, totally. Totally. It's just like, oh, this could actually work. It's a great feeling.

[00:37:31] So I started there and then I went to $5 for three months. And then it just gradually increased it

[00:37:40] until I went to what it is now, which is $6 per month or $48 per year. And just two months ago,

[00:37:50] one or two months ago, I launched an extra tier. So it's more expensive and it uses a more expensive

[00:37:56] AI model.

[00:37:58] Did you ever... Did you see you always like a lot of people are using like a bigger or

[00:38:03] an unlimited plan or luxury plan that has more stuff in it for an anchor. So people are like,

[00:38:09] oh, it's free. It's either $6 or $15, I think is the other price. Did you notice that the $6

[00:38:14] amount grew as well when you introduced the higher price?

[00:38:19] It might have, but I can't say that with any certainty because I don't think I have enough

[00:38:25] data and it wasn't run as an A-B test and all that. So I'm just going by my feeling,

[00:38:31] looking at the numbers and it might've gone up a little bit. I was aware of that theory,

[00:38:37] the price anchoring, so that I'd always thought that it would be nice to have a more expensive plan,

[00:38:42] both because there are some users who are just not price sensitive at all at these price points.

[00:38:47] They'll just, I just want the best and it doesn't matter to them. So it's like silly not to have

[00:38:52] something that they can buy. And yeah, it makes the normal plan then feel like a good deal. So I'm

[00:39:01] glad that I finally got an extra tier. Yeah. And since I've added that it has hit the 10K, I think it

[00:39:08] was the month after. So it seems to work. Yeah. Yeah. Yeah. That's basically also the only reason

[00:39:15] we have a 247 plan. We don't have a subscription plan, so it's a little different, but we have a big

[00:39:21] credits plan that's like much has a way higher amount that you get in credits than the other two

[00:39:28] plans have just, I'm recording it. I think the biggest one is called studio. So it's has to appeal

[00:39:33] for like bigger companies. Yeah. Like, oh yeah, we need this. So we just take the biggest one

[00:39:38] and anchor the other two for the remaining people. And we're not live yet with it. We are live,

[00:39:43] but I haven't launched it officially, but we'll see how that works out. Cool. I'm not sure anyone

[00:39:50] will ever buy the biggest studio plan, but we'll see. Cool. I think it has an extra effect apart from

[00:39:56] the price anchoring. I think it makes you look more serious too. It's like people, if people believe

[00:40:03] that people are paying you that much money for that plan, then they're going to think that, oh,

[00:40:10] this company must know what it's doing. If it, if people are using it for studios, for buying at

[00:40:16] this price point. Yeah. Yeah. And also solves a very important problem for us because before we did that,

[00:40:23] we also had subscription plans, but here's the issue for a product. So this product is for podcasts

[00:40:30] and there might be a podcast that's just starting out and they have two episodes that they want to

[00:40:34] then convert into other stuff with repodcasted. But what if the Joe Rogan podcast signs up and they

[00:40:41] have 3000 episodes and do they then have to subscribe to the biggest tier for four months that they need to

[00:40:48] reiterate through all of the episodes and then go to the lowest plan because they're just releasing two

[00:40:52] months, two episodes a month. So we had this issue where subscriptions really didn't work for our

[00:40:58] model and switching to credits, you now get like, you can show the user, okay, if you're a small user,

[00:41:04] you get a very cheap price and you can buy the credits when they're over and you can just see how it goes.

[00:41:10] And you only have to pay once. And if you don't like it, that's it. But have a lower per hour price,

[00:41:16] the higher you go. So if you're very convinced, it actually makes sense to purchase a higher price because

[00:41:22] you get more credits for your money. So that's why we moved away from subscription totally because it

[00:41:28] just wasn't fair to users and not to us really. And it just made things more complicated.

[00:41:34] That's interesting. It's even without the subscription, it's naturally recurring anyway,

[00:41:40] because people publish podcasts on a schedule normally. So that's, that makes sense.

[00:41:46] The plan is to then return maybe later to return to a subscription where they get like 10,000 credits

[00:41:52] a month and they can just use them mostly for people that go, I don't want to buy these manually

[00:41:58] every three months. I just want them to be there. But we'll see after launch how that works.

[00:42:03] Yeah. Yeah. Because it's what reminds me of audible where you can buy books one off,

[00:42:08] but it's so much cheaper if you buy the subscription and then you just get one credit per month.

[00:42:13] Yeah, exactly. Yeah.

[00:42:14] That why wouldn't you just do that? And then you end up six months

[00:42:18] and you've got all these credits.

[00:42:20] Yeah, exactly. And that's, and then you run into the issue of,

[00:42:23] do I renew credits that some people have paid? But then I have the issue, like they pay for the

[00:42:30] subscription for a year and they don't use it. And you already have taken out the profit or you have

[00:42:35] done something with the revenue ads, whatever. And then somebody starts using the tokens and is

[00:42:41] occurring the AI costs that we don't have. Do you drop the tokens after a month? Do you give them

[00:42:47] the new ones?

[00:42:47] You need to think about that.

[00:42:49] My mobile provider in Spain carries over the data you don't use from one month to the next,

[00:42:55] I guess it's a bit like that. But I think there's a limit. I think it wouldn't,

[00:42:59] not everything accumulates. So yeah, you have to maybe set some limits, but maybe it'll all average

[00:43:05] out anyway. And so you don't need to worry about it too much because you'll have new people paying and

[00:43:11] some people will use them. Some people won't. And I don't know, you need to see if it's a real problem

[00:43:18] or not.

[00:43:19] I think that's the number one tip for me to not worry that much. But it's also the biggest issue

[00:43:23] I have with all of my India Aga projects, I think, is that I have the security guy in mind who tries

[00:43:31] to like hack every route that we have. And I have the guy in mind that upgrades their plan to get the

[00:43:37] credits and then downgrades, but still has the credits. So you have to pay more for them. So yeah,

[00:43:41] we're trying to fill all the holes.

[00:43:43] I think try it and see is a good approach, but maybe guilty of this line of thinking when it comes to

[00:43:49] affiliate schemes or anything that tries to achieve virality without paying anything like the way

[00:43:55] Dropbox did. Because I do have a lot of users who use it for free. And there's certain countries that

[00:44:03] people come from that for some reason, I don't know why exactly, Readlang has really taken off

[00:44:08] in North Africa and the Middle East, like people learn in English. And many of these countries

[00:44:14] aren't monetizing that well, compared to the USA. And so I'd worry about people taking advantage of

[00:44:21] anything that I offer for free there, affiliate wise. But that's what led me to think, ah, but if I

[00:44:27] do it for the people who've paid, they can give a gift to somebody else that could work and not have

[00:44:32] that problem. But in the end, I think you just got to look at what's the downside. You can try

[00:44:38] something, it doesn't work and you just roll it back. It's not a big deal. So I would advise you

[00:44:42] just to how long have you been working on your thing?

[00:44:46] David I hate this question because everyone on Twitter has launched it already like for half an

[00:44:50] hour for half a year now. I think every guest has said just fucking launch it. It's it's life. You could

[00:44:56] probably go in there and buy credits and use it. It's not in a, it's not in a state that I want to

[00:45:03] launch it on, let's say product hunt, something like that. We still need to like haul over some

[00:45:08] design decisions. I've done some stupid things. I've done, I've built it for my podcast and at

[00:45:13] some points in the app that I've built it for me. And that's not a great thing in that case,

[00:45:18] because there are some...

[00:45:19] David It doesn't matter. If it's like you need to get your first five, 10 users, who cares?

[00:45:25] David Yeah, probably.

[00:45:26] David Who cares?

[00:45:27] David It's always the issue of indie hackers building

[00:45:31] product instead of marketing, I think. And in the end, maybe nobody buys it.

[00:45:35] David Yeah, but building product, to build product well, I think you do need some feedback

[00:45:38] from people. It depends on the product, right? But it's always good to get feedback as early as

[00:45:45] possible. I would just get it, get other people using it. Are other people using it yet?

[00:45:50] David Yeah. And that's the thing, right? My first product, I never launched it and I worked on it for

[00:45:54] two and a half years. Now for Repodcasted, I think we're working on it for less than a year now. And I

[00:46:00] want to have it launched before it is a year. But I think we already got learnings from it again,

[00:46:06] that we need to be faster than that. We need to be like, we don't want to be like, I am doing this

[00:46:11] with my significant other, who is a UX designer. So we have both these covered, which is awesome.

[00:46:17] But I think we both already learned, we never want to be the people that launch in a week.

[00:46:23] But we need to be able to be the people that launch in two months. That's, I think,

[00:46:28] something that I can manage. I don't want to put something out there that's so unfinished that I'm

[00:46:34] not proud of it. But I want to be able to show the world something before it is perfect,

[00:46:39] because it will never be perfect. And then it will never be launched. We need to get better at

[00:46:45] finding that, that angle. Yeah. Yeah. And I'm aware of that. Yeah.

[00:46:50] I think it helps to get stuff out there, to get the feedback. And also, for me anyway,

[00:46:57] and I heard the guy, I think Paul Buchheit, the guy who made Gmail said this, that he likes getting,

[00:47:03] like releasing things really quickly, because that kind of motivates him. And he'd rather

[00:47:10] get something out there and get the feedback and fix problems out in the wild. Like I'm a bit like

[00:47:18] that. Like I, it's motivates me. And I don't mind putting something out there. And I feel really excited

[00:47:25] for people to use it. And I'm still like checking if it works properly or not.

[00:47:29] I don't know if you heard the podcast that Lex Friedman did with Peter levels recently,

[00:47:35] where he talked about, he doesn't even test stuff locally. He just pushes the change

[00:47:39] and then tested in production. Yeah, that's maybe a little, you need to move not 100% there,

[00:47:47] but more in his direction. Yeah, I for sure need to be deprogrammed. I think that's maybe because

[00:47:52] I've just been doing enterprise software for maybe more than a decade at this point.

[00:47:58] And it's just like testing. I think that's something you need to get out of your head when

[00:48:02] you are in the indie hacker space or the bootstrap way. Yeah, I think so.

[00:48:07] Yeah. So hopefully when this, I think this episode will probably come out in December.

[00:48:13] So hopefully we have something going on till then. I hope so. For my own sake.

[00:48:17] Yeah, yeah. That'd be cool. Getting back to ReadLang a little bit. Obviously now, since the AI hype has

[00:48:26] been going from 2022 and onwards, a lot of people have been doing language stuff, be it language learning

[00:48:33] apps, apps where you can talk to an AI in the language that you want to learn. What's the deal for

[00:48:39] you there? Obviously you still have grown a lot in that time period, but what's your thoughts on that?

[00:48:45] How much do you see those people as competitors? I don't think about it much, especially say they're

[00:48:52] doing an app that lets you talk to an AI. That's not my niche within language learning. So that doesn't

[00:48:59] bother me at all. I guess if someone came along and did something that was much more directly

[00:49:05] competitive with ReadLang, which basically enabled you to read stuff and provided tools to help you read

[00:49:14] and improve your skills by reading, then yes, I guess that might concern me. But I don't know,

[00:49:21] things like this have come and gone over the years. The one that seems to be the biggest one

[00:49:29] that's close to ReadLang and ReadLang gets compared with it a lot is Link or LinkQ. And every now and

[00:49:36] then I do a Google Trends search to see how ReadLang compares with LinkQ. And I'm slowly catching up.

[00:49:42] They're ahead of me and they predate ReadLang. They've been around a long time and I wasn't aware

[00:49:46] of them when I started ReadLang, but I think within a few months of started ReadLang, I found it,

[00:49:52] or people mentioned it. And at the time, I found it confusing. The UI, I wasn't that worried about it

[00:50:01] because I thought, but they have clearly been doing something right because they are growing

[00:50:05] and they have a team as opposed to just one person. So they must be making a decent amount of money to

[00:50:13] pay a whole team of people. So to be honest though, that just makes me feel more confident that

[00:50:22] ReadLang has room to grow. The fact that there's this other competitor that can support a whole team.

[00:50:29] And when I ask people to compare the two, there are things that Link does that ReadLang doesn't.

[00:50:36] But in general, people seem to prefer ReadLang. They say it's simpler and easier to use.

[00:50:41] So it makes me feel good that there's a competitor that proves there's a market there that I can now

[00:50:49] work towards growing as big or if not bigger than them, maybe.

[00:50:55] Do you already have the SEO standard trick of having a comparison page with them on your site?

[00:51:02] The closest I've got to that is on my own ReadLang forum. I asked people to compare who else here,

[00:51:08] who here uses Link? I'd love to hear about how you think ReadLang compares to Link.

[00:51:14] And that does come up in Google searches, but it's not that... I should probably work on SEO,

[00:51:22] things like comparison pages with Link, even with Duolingo, why not with different language learning

[00:51:28] apps and explain who should use this one, who should use ReadLang. There's probably low hanging

[00:51:34] fruit there. The problem is... Well, it's not a problem. The great thing is I have so many people

[00:51:40] that come to ReadLang searching for ReadLang directly that it just feels like any other channel

[00:51:49] has a high bar to hit to become meaningful in impacted traffic to ReadLang. But it doesn't

[00:51:56] mean it's not worth trying and I should try that, but I haven't got around to it. A million other things.

[00:52:02] Yeah, it's always like I have a list of so many ideas that I want to do, but we have to do other

[00:52:06] things first and stuff like that. Two people... I went to MicroConf recently and there were two people there

[00:52:12] who were really insisting that I do SEO, that I do this kind of thing. So I probably should...

[00:52:18] I might know one or two of those. Yeah. And also one thing is like how much time do you really want

[00:52:23] to invest versus doing other experiments, doing other software maybe, or doing nothing at all,

[00:52:29] which is also fine if you have hit some market, some revenue numbers that you want to hit and

[00:52:35] need to hit to sustain yourself. Because I've seen that you wrote like the second year in a row revenue

[00:52:39] hits an all-time high on the month I worked the least.

[00:52:44] Yeah, that's nice. That was in August when the kids are off school. I've got one kid and we go on holiday.

[00:52:51] And so I don't have time to do hardly anything that month. And yeah, for two months in a row,

[00:52:57] that was the month that it hit the all-time highs. So that's pretty sweet.

[00:53:02] I think we need to get together all the founders that have talked about getting most of the users

[00:53:07] from word of mouth and find the secret sauce you guys are using to lend that out.

[00:53:12] Well, I'm sure freemium has a lot to do with it in my case and being around a long time.

[00:53:19] So I started 12 years ago. So that really, I have a lot of benefit from a lot of links built up from

[00:53:27] different blogs, forums and stuff online. So I guess Google values my domain quite highly,

[00:53:33] probably because of that. Yeah. Have you learned anything from talking to other people about

[00:53:40] what works well? What I've seen now in the recent past is Stefan Wirt, who has done a lot of stuff in

[00:53:47] SEO and has taken a lot of time to find a way that it works. But now he has, I think, two products

[00:53:54] adjacent to the actual SaaS product he's building, who makes him enough money to now be pretty much

[00:54:01] independent from any other source of income. So he has done a lot of SEO work. So if you want to check

[00:54:07] that out, the listeners as well, go over to Stefan's Twitter page. He writes extensively about this.

[00:54:13] You find him on YouTube as well. And then another guest that I had on the show.

[00:54:18] Stefan, what was his name? Stefan Wirt, which his last name is W-I-R-T-H. He was not on a podcast yet,

[00:54:25] but I would love to have him. He has done extensive work on how and explaining how the SEO stuff can work,

[00:54:31] how you find the right keywords. And the thing that I discovered a few months ago, which I really like,

[00:54:38] one, I'm not affiliated by any of these people. One is another guest who has been on is Yossi.

[00:54:44] On Twitter, his name is Hi, I'm Yossi or Hey, I'm Yossi. Sorry. And he's done the website SEOstuff.com,

[00:54:50] which is like Ahrefs or SEMrush. But I got the credits thing that we just talked about from him.

[00:54:56] You only pay for credits and then use them when you need them. So it's a very expensive subscription.

[00:55:02] So that's one thing I'm doing, looking up keywords there. And then another free thing you can do,

[00:55:09] which is awesome, is the website called alsoasked.com. So let's say you have a keyword that

[00:55:15] you want to do something about. In my case, repodcasted, it would be audio repurposing,

[00:55:20] for example, or turn podcasts into LinkedIn posts, something like that. And then I would go with a

[00:55:27] short keyword, I would go into alsoasked.com. And you can put that keyword in there. And it gives you

[00:55:32] like, a hundred questions that people have asked on Google that are connected to that keyword. So you

[00:55:39] get a lot of long tail keywords you can feed with blog posts. So this website gives you like three,

[00:55:44] I think three searches every 24 hours. So it's basically free for what I need it for.

[00:55:50] And then you have 500 ideas for blog posts you can add to your site. And that's one way I would try to

[00:55:56] do it. I think if I would go full all in on the SEO part. Yeah. Yeah. This stuff would make a lot of

[00:56:02] sense if I hired somebody to write, I think. Yeah. That's the thing. That's why I haven't done it yet.

[00:56:10] Because I don't know, it's hard. Like, I've got this channel that works basically, which is like this

[00:56:17] mysterious organic spreading word of mouth thing. And I attribute it to making the product better.

[00:56:25] So I guess I'm just continuing to ride that train. It's like why? And that is the best long term

[00:56:34] investment surely is improving the product, especially if people are actually using it.

[00:56:39] I think it's an undervalued strategy as well. I think a lot of people on IndieHacker bubbles are like,

[00:56:45] yeah, you only need to have good marketing and a bad product will always win if it has the better

[00:56:50] marketing. I'm not sure about that. You might get the first month of a subscription, but you won't get

[00:56:56] the second one. And I think... Right. I find some of this advice, I agree, is a bit short-sighted.

[00:57:02] And the one about, ah, remove your free plan and charge everyone. I did that and it increased

[00:57:07] my revenue by 50% or whatever. And yeah, that's not that surprising that it's going to lead to a

[00:57:13] one-off increase of that. But then where's your future growth coming from? So I do feel that

[00:57:20] advice like that can be optimizing the short term.

[00:57:24] Yeah.

[00:57:25] Too much.

[00:57:25] I think if I didn't have an AI application that has like growing costs, the more the people are

[00:57:31] using it. So it's very hard to have a free tier there. If I do another project where I don't have

[00:57:37] any large costs when the user uses my product more, I think I would 100% go the way of tally,

[00:57:44] of stage timer, of also your application and have a free tier that works for people and then have some

[00:57:50] features locked in or some usage point where they are like, okay, now it's basically okay that you

[00:57:57] pay for this. But yeah, if I wouldn't have AI costs, I would always at least think about a free tier for

[00:58:03] having users on a platform and getting feedback.

[00:58:06] Yeah. It does make it more scary when you have AI costs, but you probably can get away with

[00:58:14] still giving some stuff away for free. You could try, you could still try it. Like

[00:58:21] my guess is, I don't, I don't know exactly what yours does. If you're dealing with processing whole

[00:58:27] podcasts, maybe it is quite expensive.

[00:58:29] The transcription stuff is like having very big context windows. So they're very expensive

[00:58:34] token wise. Yeah. That's the issue. Otherwise we have thought about stuff like giving them

[00:58:39] or even building a free tool where we just take the first 10 minutes in the transcript and not the

[00:58:45] whole thing. And then they can extract some value out of that. We're not sure how yet, but that might

[00:58:50] be something we're doing where we can provide some free stuff, but having like very large transcripts,

[00:58:56] at least I'm testing this with this podcast, which is always running like an hour longer.

[00:59:01] And the transcript alone is like 20, 30, 40,000 tokens. And then it gets very expensive, very fast.

[00:59:08] That's the issue. Yeah. Otherwise we would try that. Yeah.

[00:59:12] Yeah. That makes sense. What's the, what does your product do then? So does it generate?

[00:59:18] Basically I solved my own problem that I have one and a half hours of podcast audio,

[00:59:24] and now I want to post a tweet about it, but how do I get the value out of it without having to

[00:59:30] listen to one and a half hours? So there might be some awesome insight in there, but it's very hard

[00:59:34] to extract. And what we podcast does then is it turns the whole transcript. First, it gives you

[00:59:41] the transcript with speaker labels, which is also cool. So people could then copy it and do something

[00:59:46] else with it, but it also gives you a summaries, chapters. It gives you the biggest insights of the

[00:59:52] podcast episode. We're working on a blog post feature where we are not writing the blog posts

[00:59:57] because I think AI is not that great at it, but it's giving you an agenda with quotes.

[01:00:01] So you can say, okay, there is like three ideas for blog posts on base of this episode. Let's say

[01:00:08] we talked a lot about pricing. Then I would get something about pricing back as an idea for blog

[01:00:13] posts with quotes from you and me, and also with an agenda that might be useful to write the blog

[01:00:20] posts on them. And then you have to just write the blog posts and, or you could give it to a writer

[01:00:24] and already have quotes and chapters for the blog posts, but not the blog posts itself. And that's

[01:00:31] where we are now. And we try to add more features after we launch. We've seen a lot of people who want

[01:00:38] easy shorts to come out of this. I would love that feature. I think it's very hard to do very good.

[01:00:44] And we have very good competitors there, other indie hackers as well. So I'm not sure we go into that

[01:00:50] route, but taking the podcast audio, which is awesome and full of cool knowledge and extracting

[01:00:57] that in any way we can think of to get that knowledge to more venues, to be able as a podcaster to host

[01:01:05] it on Twitter, on LinkedIn, in your newsletter and make more of the audio you're doing because you're

[01:01:11] already doing the audio. You can then extract more stuff out of it. That's the idea.

[01:01:15] Yeah. It sounds very cool. Yeah, it sounds so like this. There must be so many ideas like this now that

[01:01:21] are opening up with AI. And that sounds like a pretty compelling use. My only doubt would be like how many

[01:01:29] other people are doing the exact same thing. But other than that, it sounds like, yeah, that's a no brainer.

[01:01:34] Yeah, we have some competitors, I think. One of them says that they have over 1000 users, and they only have

[01:01:41] paid plans. Yeah, I'm skeptical of that number. Let's say that but if that's correct, it's the same

[01:01:50] for you as for me, as you mentioned earlier, then there is a market for that. And I've tried that

[01:01:56] product. And I think we can do better, especially since I have a UX designer who's very keen on making

[01:02:01] the software easy to use and awesome, which I can't do on my own. So we'll see where it goes. But that's the

[01:02:07] plan. Yeah, make it better, make it a credit system. So you don't have a subscription if you

[01:02:12] don't want to make it fair to the user as they only pay what they actually use. And also make the

[01:02:18] interface cool, make them have the say that the right value in it. And we'll see what the users want

[01:02:25] after that. Yeah. Yeah. Yeah. Get it out ASAP. Don't worry if it's not super doesn't look super cool

[01:02:34] for v1. Yeah, I think it looks okay. Now I think I've fucked up some stuff that needs to be corrected

[01:02:41] by my designer now. But we'll see how it goes. But I hopefully we can make an official launch.

[01:02:47] It also took forever to found the company that needs to have that asset. It's a German thing.

[01:02:53] We talked about this in length in the last podcast episode, we needed a company, we didn't want to

[01:02:58] found it, but we needed one because of German tax law. And yeah, that took six months. So we couldn't

[01:03:03] launch for six months, even if you wanted to. I started Readlang in as a UK company, because

[01:03:09] I was in Spain, but we were like planning to go back to the UK. And it just we didn't end we

[01:03:15] ended up staying in Spain. But it's so much easier in the UK than it was in Spain, like to comply with

[01:03:21] all the tax regulations and like accounting requirements and stuff, especially when you're

[01:03:27] really small. Like the UK makes it really easy if you are not earning hardly anything. They

[01:03:32] it's super cheap. You pay like about 15 pounds a month in national insurance. And my account,

[01:03:39] I just did one accountant once a year to do the annual reports and I paid them maybe it was almost

[01:03:44] a thousand pounds, but once a year, that was it. We had around like 450 a month.

[01:03:51] Yeah. Okay, that's expensive because we just paid 1000 euros to found the company and put another two,

[01:03:58] I think in it. So it doesn't go bankrupt. That's the moment it starts. Yeah, yeah.

[01:04:02] But yeah, it's not fast. It took six months and it's not cheap. So that's just a German way of doing

[01:04:09] things slow, inefficient and very expensive. Yeah, we got to wrap up in a second. But I have one

[01:04:16] question I want to ask you that I asked all the guests in the end, which is if you would have started

[01:04:21] now, if you have an indie hacker asking you, what should I do now? What's the one advice you would

[01:04:26] give someone that's just started out now? Launch.

[01:04:35] I mean, basically, yeah.

[01:04:40] You need to get feedback as soon as possible, I think. It doesn't matter if it looks like crap,

[01:04:45] we'll just get something out there. I think, yeah, I think that's a good idea. Yeah.

[01:04:49] I think I need to adhere to that more. Talking to 30 guests now, I think that's a good idea.

[01:04:55] There's no downside really, because it's not like you're going to launch and it's going to hit

[01:05:00] the New York Times and the whole world is going to know about it. It's going to launch. And if you're

[01:05:04] lucky, you'll get 10 people to sign up or something. That would be already.

[01:05:09] Yeah. You can already. And then by the time other people hear about it, they won't have any idea

[01:05:14] how crap it was when you launched. But those 10 people may give you some really useful feedback.

[01:05:20] Yeah. So we'll see that when this episode goes live, we have already launched. This is a big

[01:05:24] thing for me to say that even in the podcast, because it binds me to that. But seeing as we don't

[01:05:30] really have that much more to do to be very comfortable in launching, I hope we get that

[01:05:35] done. Yeah. All right. So thank you for being here for doing the podcast. If people are interested in

[01:05:40] readlang or you, how can they find you? So readlang is at readlang.com. Just search for readlang. Or

[01:05:47] you can see me on Twitter or X. I'm not that active, but I read it a lot, but I'm not that active with

[01:05:54] posting, but maybe that'll change at some point. Exactly. And also Steve underscore

[01:05:59] writer, I think is my handle. Yeah. Steve underscore write out. Yeah. I already have it open.

[01:06:04] And there's also your website. I think at steverideout.com without an underscore, right?

[01:06:09] Yeah, exactly. I don't post that often, but I might be tempted to do like an end of year

[01:06:14] blog posts. Yeah. And I think it's also already very interesting to go on Steve's website because there

[01:06:20] are blog posts about all of his journey at different points, which is very interesting

[01:06:25] to go through when you already on the 2024, Ryan, of history and go back to 2022 and further down

[01:06:31] and see what Steve thought about then. So yeah, check that out. You find all of the links as well

[01:06:36] in the description as always. Steve, thank you for being here. And yeah, hopefully when this releases,

[01:06:42] there is also repodcasted live somewhere. Definitely. Yeah, I'll be checking. But thanks

[01:06:50] for Tobias. This was really fun. Yeah. Thank you too. And goodbye.

[01:06:56] And that's our episode. Thank you for sticking with us to the end. You can find me on Twitter

[01:07:00] with the username icebellabs. That's I-C-E-B-E-A-R-L-A-B-S. We have a website you can check

[01:07:05] out. Go to codeandconquer.fm to find out more. You can find this podcast on Twitter, TikTok,

[01:07:11] and Instagram with the handle codeconquerpod and can write us an email at hello at codeandconquer.fm.

[01:07:17] If you're one of the people who listen to this podcast on either the Apple Podcast platform

[01:07:21] or Spotify, please consider rating us. It lets other people discover this podcast and helps

[01:07:26] us grow this even bigger. Thank you so much for listening. Till next time.