Latest / App Minded / Marketing an App Built With AI: How One TikTok Got 3,000 Users in a Day

Transcript

Marketing an App Built With AI: How One TikTok Got 3,000 Users in a Day · App Minded

• Nina Kolari: Actually September we are building records every day. I'm I'm very happy. I had a TikTok going viral and that TikTok alone brought me five thousand users and now we have around forty bookings a day. Total of we have a total of fourteen thousand users today. Welcome to App Minded. My name is Nina Kollari. I'm an AI product builder and have shipped over twenty mobile and web apps in the last year. I'm on a mission to help you create your very own apps with the help of AI, with or without technical background. I also run canaibuild.com where I show you which apps you're paying for right now that you could actually build yourself. Today I'm really excited to have Lucas Kruder on the show. He's gonna talk about the app that he recently started, an app called LastBot, that is a last-minute fitness booking app in Paris. Welcome to AppMinded Lucas. Can you start by telling about yourself and your background? Hi Nina, thank you for having me on the on the podcast. I'm very happy and excited to to share more about last thought. so I'm Lucas, I was born and raised in Paris. I'm 28 years old. and and I I have a background of working in finance and then in the cryptocurrency ecosystem, working in blockchain industry, organizing events, and more particularly marketing. but so like this allowed me to to be very interested in new technologies and with the rise of AI I I really wanted to to learn more. So I started reading about cloud code, about chat GPT and everything that it allowed. And I always wanted to build a project. so I had this app ID and and deep du dive in the in the AI coding, vibe coding, and and little by little I I put together this this app so I started working on it in January and it was ready to launch in April. So it's been a few months already. Was the last spot your first app that you built? so around two years ago I started having another app ID, which was more in the running industry because I love running, I love triathlon and and this kind of endurance sport. So I wanted to to build an app for running clubs. I won't go into details because it it didn't go very far. I but it allowed me to learn more about the workflows and about a bit about the vocabulary for for developers which I wasn't a developer at first but so I had no vocabulary at all. I didn't know what a front end was, what a back end was, what a database was. So but it allowed me this first project to get into it. didn't go to the end, but yeah, it was a good training experience. Awesome. With the last spot, was that maybe your third, fourth, tenth app that you built? In the in the two years that you it was just the second one. I I had other projects building websites for friends and family, so really like working with cloud code every day or with other AI models, but but yeah, like as as a a project and a and a real app, Lastpot was the second one. where did you get the idea? so my my girlfriend works at a Pilaris studio in Paris and I was hanging out there a few times and I realized that sometimes they had empty spots in their classes and they were doing the classes anyway. and I I just had the idea. We we have an app in France called Too Good to Go, which bakery or boulangerie as we say in French so sell their unsold products, so stuff that they didn't sell throughout the day, they're cheaper at the end of the day, just be instead of trying them to trash. Well, I I thought maybe we could have the same system for for those p fitness classes. instead of having the classes with like two empty spots, might as well sell it for cheaper in the last minute. And so that's how the idea came in. And after speaking with a few fitness clubs, fitness studios. I realized that there was a a need and I mean for users as well, there's always a need to pay cheaper because prices of a PDADIS class or yoga class in Paris can be expensive, up to fifty Euros per per hour sometimes. So so yeah, like it's a price that maybe students can't afford or in if you have a possibility to pay it cheaper, then you you you might as well. Yeah. What kind of market research did you do beyond chatting with the studio owners? Did you actually look if there's apps already in the space or something like that? Yeah, so I I've looked about similar system. They're called like booking aggregators. there there is Class Patch, which is the main company or the main app on the on the market. they have a a huge international market and a lot of people on this app. But I after speaking with Studio, I realized that there were some constraints of of using ClassPass on the st on the studio side that I can resolve by being more flexible, cheaper, and just doing it in a in a different way. And for users, it's different as well because when you're on ClassPass you have a mem subscription and I decided not to make a subscription for users, which is kind of like a different angle And yeah, there are other aggregators, but I feel like the market is quite open for for new new actors to to take some market share. Awesome. So now you had a problem. How did you come up with a solution? Like was a mobile app always the solution or did you also think about only building a web app or so at first I started by building a web app. but of I think it's limited because if you want to buy a fitness class on the go you not always on the computer and it's much more user friendly to be on a mobile app. That's what peop people are used to do when they when they want to book something so so then with the help of AI I was able to pivot from a web app to a to a mobile app quite quite easily. Yeah. So you build a mobile app. you currently have both iOS and Android. Did you start with the both or did you start with one? so my priority was to be on the app store at first, obviously with the goal of being on Google Play Store as well. I found it it was easier to be on the App Store. there's much more info on how to get it right and be approved for the listing. so it wasn't so complicated. I mean it's a process, so it takes a a bit of time but I found that the Google Play Store listing was way harder because they ask you to have testers and people like a beta testing program. So it was on the App Store at first and a few months later only it became available on on Android. cool. Do you have separate code bases for iOS and Android or do you have combination? no, it's the same code base. I use the software called Expo, which allows to to talk to the two systems. My infrastructure is a monorepo, so I have a code for the web and code for the mobile. Awesome. What was your first prompt when you started building that? Did you know what you were gonna do? Before prom like actually getting Cloud to code, and that's some stuff I researched before. you really have to guide Cloud, that's from what I read. And so not just like say build me an app that does this. So my first prompt were quite technical, but it allowed me to guide Cloud code in the best way possible. for example, I wrote a little brief of what the project ID was, but then I asked him to ask me a lot of questions about project infrastructure, technical requirements, constraints, and really challenge the ID. So then he he he asked me a lot of questions to to know which database we're gonna use, to know which coding language we're gonna use and read it so that he and he has a lot of context and understand the project right. And when he did, when I was satisfied with the direction we were gonna go in, I allowed him to go and c starting coding actually. And and when we had a first code base and what I think is the most important is that I asked him to register all the context he has and the memory he has of the code base and the of the ID of the project into markdown files, which is also a bit technical, but it all makes sense when you start using cloud code a lot. So markdown files is a way for c for your AI agent to know your the context of of what you're building. And not go in a loop with the same problems over and over. So I built a lot of those markdown files for giving Cloud more context, basically. So I have a a markdown file on how my payment structure works, a markdown my marketing strategy, a markdown file of which studios I've spoken to, and and so when Cloud is coding, then he refers to when to refers to all those files and And is is more efficient. So yeah, the first prompt has to be really guided onto a like a very strategical way of going. Before you started building, did you watch any YouTube tutorials or ta took any courses or did you just go straight into colour code and it's like how we gonna do this? So I I watched one tutorial about this markdown files, but then I thought that maybe the best way to to become better is just by using not so much l learning I there's so much content that it's hard to know what to watch and everybody has a different workflow. I found a workflow on my own that I think works well. I don't think it's maybe the best. Probably a lot of other people do it way better. But the way I did it w was good enough for me to l go from A to Z on a on an app launch. And I think there's a lot of content on Instagram, TikTok, Reddit, Twitter that I see that I that I read. And when I find something interesting now, I just like take the post and send it to my AI agent that so that it can he can read it and see if there's stuff we need to implement. Yeah. Did you since you're using Cloud Code, do you also download some skills or GitHub repos that would help you to build it? I use Cloud Code on VS Code Virtual Studio, which is linked to my GitHub and and my GitHub is linked to my Vercel. So everything is like connected and and yeah, so when I see on Twitter, for example, a skills or repo that that is that is interesting for us to use, I send it and I asked Claude to to to download it and use it. one of the good ones that I use are obviously the front-end design skills, which allow us to to have more beautiful designs. playwright is very in in interesting. I think it was made by Microsoft. It allows Clute to give to have eyes. So he will code, actually have a look of of what he builds and then make some some corrections if he needs to. so playwright is a good one. I also use remotion for building clips, building animations for the marketing, which is which is interesting. You started building the app. You have kind of two customers for the app. You have the studios and then the users. Yep. how did you even tackle to two potentially different users and different kind of ways of using the app. Yeah, so that that's a challenge. I think it didn't make it easy to have two different audience as a B2B part and a B2C part. But it's kind of like how Uber did it at first in San Francisco when they had to find the drivers and then to find the people who needed people to ride. So I did it kind of the in the same strategy. First the goal was to find studios And have them as partner. So that was the first job, finding about 10 studios in Paris that would agree to be a founding partners and list some spots on the app. And once I had 10 studios, I decided it was time to start marketing and actually speak to the users. So this went by TikTok, Instagram ads, so basic social media distribution. I think in Paris there's also a lot of word-of-mouth impact. Like you I see like people booking with their friends at the same time. That's allowed me to have a good growth in the last few months, I think. But yeah, the the the best channel for acquiring users was TikTok actually. Like that's allowed me like one good TikTok a few weeks ago brought me five thousand subscription like new users. Yeah, that's amazing. if you have the right content strategy. I'm not very good at content, but I try sometime it doesn't work, but Sometime the algorithm push pushes you. Yeah, yeah. It's you just gotta keep trying to see what works. definitely. I'll come back to the marketing bit later. How much did you have the app ready when you started talking to the first studio? Or did you even start building the app before you got those ten studios to sign up? The really first thing was having an MVP before even talking to studios. so that's what I did. I I went But it was fun for me to work at first on the app, adding features. when you start vibe coding it's just it's it it's becoming addictive. So I was working even at night, adding new features. So even my MVP was quite complete when I first started speaking to studios. of course later on you keep adding new features and fixing stuff along the way. I it was quite quite complete when I I spoke to the first studio. Yeah. Yeah. So Now you have your studios and you've built the app. How did you integrate all the information that the studios about the class times that are actually available? How did you do that? So that was the most technical part of this app because for studios to be using Last Spot, it has to be super easy. And because they all all al already use a software that allows them to manage their planning, manage their classes. I had to integrate this because like they cannot juggle between two platforms. So for some software it's easy to integrate because they have a lot of API documentation, developer documentation that you find online and just send it to Cloud and Cloud will build you this API integration quite easily. For some others it's harder. So you have to speak actually to those people and ask them for partnerships and but like in in Paris there's a a software called BSport, which is what a lot of studios use, that we manage to to integrate with like API documentation. And and so that's allows me to ha that allows the studio to read their b Sport planning on their Last Spot back office. And if they have a booking on Last Spot it also writes on their on their b Sport platform. So yeah, that's that's super important for for the app that it doesn't take too much time for the studios or too much effort to handle and manage those last minute spots. Yeah. That brings me to since there's APIs and multiple users and studios and everything. Security. How are you tackling the security of all these Back and forth APIs. That's the un not funny part of the vibe cutting because at first you're just going all in and then you realize that if you actually want to launch something, it has to be secure. So I did a lot of audits of like checking if my database have have vulnerabilities, checking if the IPA keys are are stored in a way that they cannot be accessed by the public. And and once you've done all those safety checks that your database is secure, then you are free to launch. So I did all all those audits, make sure that my database was secure, that my code wasn't exposing some some info that shouldn't be shared. But it's important as well for in terms of legal, because in Europe we have and probably el elsewhere in the world, all those GDPR R GPD in France, I don't know how you say it in English. you have to be compliant on this side as well, which takes a lot of work to to be compliant and And so that's not the funny part, but you have to do it. Yeah. that's the part that I worry about the most. That's the part that keeps me up at night. Making sure that I nail the security because nothing worse than that somebody hacks into the database. There's a lot of way as well to to make a mistake, so you really have to audit everything and and make sure that it that it's very secure. Yeah, and it's not just wipe coded, it's Developers get into the same thing too. It's it's it's hard even for developers to do. If we go now now you built the product, are you still building it solo or do you since have more people help you? Have you ever had a developer look at it or is it still completely solo? I'm solo. I mean I had some developer friend like for this security thing, for example, audit d yeah. I I have some help with the content, my girlfriend helps me with the content, but it's like quite quite solo. I don't need a team right now because that would mean more expenses. We are at a early stage and it's growing good. I think like there's a lot of stuff that I can do myself with the help of Cloud, whether it's on marketing, on sales, on product. and I have enough time to tackle a bit everything myself. But the the goal in the future will have will be to have like dedicated teams on marketing and sales so we can act more more faster so that will be a goal for for for the future. How much are you using AI in your marketing and all the rest? Is AI like almost like your co-founder? Yeah it is I mean it's my on the marketing sign it's quite it's more abstract compared to software development. so there's a lot of strategy. Yeah, I would say that on marketing I brainstorm with AI, but it doesn't really do much. It's for content what he produced, like those animations, they're cool, but they're not like very professional so far. I think it's getting better and better with the more recent models, the capabilities are better in terms of content. It's only about social media right now for for marketing, which I think I have a background of in social media as well from a previous jobs and And I I'm a user of those social media so I know what how to address the the audience. That's what lacking and that's the hardest part actually, because when Cloud produced all your your product, then you have to do stuff by by yourself to to grow, which is like all the legal stuff, all the marketing stuff, getting on call with studios, that's some stuff that Cloud can do. Yeah, you mentioned that building a product was one thing that often is talked about in social media or somewhere, but nobody talks about all the I'd say it's the ninety percent of of the process is the legal and all the other parts, marketing. Yeah, yeah, the I mean marketing is one. There's also accounting, there's even listing an app on App Store or Android is a whole process that AI cannot really be involved in. Of course it can guide you but that's some stuff that you'll have to to go through by yourself. then registering a company. That's the actually the first time I I start a company by myself. So I didn't have that experience before. the yeah you have there's actually a lot to take into account, like all this VAT, all this accounting and and stuff. So that's that's a journey you go you go through. but it's it's quite nice to be able to have an assistant like AI to explain something that you didn't quite get or even like keep you keep you organized with with what what you have to do next, like a to do list and so on. I don't think I'm I'm the most organized guy, but AI allows me to to keep a good good to do list and and separate all the the stuff I have to do. I use AI as my co founder and I've I've been in business for seventeen years so I have that time before AI and I like it's so much easier these days to start something. so How is your app doing now? It's been live since did you say April? Yeah. Yeah. How's it doing? It it's doing quite good. Actually, September we are building records every day. I'm I'm very happy. It so it started slow and then some micro influencers started speaking about it on TikTok in June. So we had a spike of users in June. so we were around eight thousand users in in June. Then the summer was a bit slower because people are away on vacations in Paris. We still had around three hundred and fifty bookings per month in the summer. And then in September, so I had a TikTok going viral. And that TikTok alone brought me five thousand users. And now we have around forty bookings a day. So now yeah, it's going quite quite well and motivated to scale it even even higher and and and even double those numbers in the next few weeks if I can. Yeah. how did it feel to get your first hundred users? It it's a good feeling. So and the first thousand was very important for me because before I launched I I created a wait list on on social media. So I I advertised my wait list with with ads, with insta ads and With only those ads, I I already had a thousand people on the wait list. So when we launched the first a hundred user came quite quickly. But when we achieved a thousand, I was okay. There's actually a thousand people who downloaded the app. that that felt good. And yeah, that it's good to have those milestones like a hundred bookings, then a thousand bookings, there's so some stuff to celebrate and then think like it's like milestones that keeps you happy and shows that you you had a a good idea and it's worth actually working on and developing. Yeah. Yeah, absolutely. And congratulations, that's pretty awesome results. Thank you. For a few months of app being out. You mentioned paid ads. You had some experience with ads already before? no not really. I mean I ran I tried running some ads in at my previous jobs, but I always find found it complicated to use meta ads and so what I do now is just the I use Instagram ads through Instagram on my web on my desktop. I found I created an audience that like Pilates yoga in Paris, which I think is around a million audience size in Paris, million people. And I don't spend that much money every day. I think the sweet spot eras around five euros a day. It's a budget, but it's I'm not yet in the hundreds a day. But that works well to bring me a lot of followers, bring me a a lot of new users. So so yeah, I think I found a sweet spot and I'm trying to to keep it this way. Maybe I when I I have more success or more funds I I will be able to raise those and try with a with with more money but so far it's it's good. I know it can be optimized. I know a lot of people do like A A B testing. They have a really good ad strategy, but I haven't been into it so so much. You I think you're doing pretty well. If you're doing five dollars a day and you're still getting users and you're happy with it, I mean, yeah, you can always optimize, but that's already create results, you know. For a new app. Yeah, I think it's a good way to I mean, I'm not gonna spend a hundred euros a day right now, when it's when it's just a beginning and and it's good to to try your to tr to try with a s small audience size at first and and because this product is very niche actually like a lot of the users are are women today. everyone based in Paris gotta love Pilates, yoga, Lagri, this kind of of sport. So it's quite niche and it's easier when you are try to address to a to a niche audience. Obviously later on I will try to to add more activities like boxing, like cycling and and maybe golf one day. I don't know, like s some more yeah more more categories of sports and expand the audience. But right now I think I that I found my niche and it's good to to try to scale this one. Yeah you mentioned scaling there. You're now in Paris. Are you thinking about scaling the Pilates Side on like Bordeaux or somewhere else in in France. Yeah, because when I actually when I had this viral TikTok, a lot of people in the comments were asking for their city to be on last spot. And there was a lot of French cities of I mean fr all over the country. So I've been contacting now studios everywhere in in France. Have some calls with with some in in Lyon, with some in the south of France. So yeah, the next goal is to to be in more cities because as it is a commission-based business I don't get a lot from from a booking so I need volume to to scale I need to to go into different cities and expand expand this way. Since you mentioned commission, who is actually paying for the service? If it is it the studio or the users? I wanted I wanted to do a a no subscription model. So studios don't pay to be on the app. Users they pay by the By the class. If they want to book, they they pay, but if they don't, they are not in a subscription model. They're paying like an average of 50% off on their fitness classes users. And on this discounted price, I take a small commission from the studio. By the way, the studio has a full flexibility on the prices of their class, of their listing. So they decide if the discount they want to put in on LastBot. So so yeah, that's that's the model. I don't know if it's the best one or not, but the fact that it's super flexible for both the studios and the users is helping for the growth because it's way easier when you speak to a studio to tell them that it's not gonna cost you a membership, you're not gonna be like blocked for a year of using Last Spot you decide to join. Me, no, yeah. You can like try it for a week, decide if you like it, you have full control of the settings. I will only get a commission if you get a booking, so in both ways you'll and you will get a booking on a spot that would have gone unsold anyway, so everybody wins. Yeah. Yeah. So I guess it's quite easy sell in a way for studio. They don't really pay unless somebody actually books. Exactly. And it's just like super flexible. So they there's no no way that they would feel stuck on the app or okay, cool. You said you had fourteen thousand now users. Have you noticed Or any issues arise when it's different to have an app that has hundred users now than versus well over ten thousand. Has any issues come about now that you have lots of users? Yeah, I actually having that spike of users following the the TikTok with hundred of people coming by the hour, even more, I think like one day we had three thousand people coming in the same day. So so yeah, that was that was Making the app slower. It wasn't optimized for and you can pay for more compute size on superbase to make it like faster. so that's what I I did. And also like I optimized because my code reads my database, some stuff didn't need to be really read by the code. so you optimize this way so that every request made by the users as light as possible to make the app faster. But unfortunately you I knew about it when people were coming in by the hundreds and the app was lagging and the database wasn't f working anymore. So that was a that was bad because that's the wrong worst moment. You have people actually coming, but now the app is not working because it's saturated. So now the next time people come in, it will work well. And I was happy that I I saw that that quite fast and I was able to fix it in like a few minutes thanks to Claude actually. And but you know wanna know the worst part, you know Claude was down A week ago. it was down for a whole afternoon. And that was the day the TikTok blew up. No. So I was trying to to make to fix that database issue, but Claude wasn't very responding. So it was a fight. But I I did it. Yeah. Did you have do you now have a backup plan in case these things happen again at the same time? Yeah. Well h hopefully it won't. I went speaking to Chat GPT when it was when closed was was down. So So yeah, that was that was my rescue backup plan. Yeah. How did you find Claude versus Chat GPT? Do they have similar advice? I think for me, today I'm not an expert enough to know like they're both super intelligent and they will get you where you wanna go. So it's hard to to to say. I I actually like reading about like all those new models coming in and the future of super intelligence, which is scary sometimes, but it yeah, it's very interesting. But I don't really know like how to judge which one is better. Like I I think they're both super intelligent or more intelligent than me. So they'll help me. Yeah. What's the level of subscription you have with Claude? I am on the 100 Euro a month plan, so and that was enough for you to build the app. Yeah. I mean when it was like product building, sometimes I would reach my limit. because it was a lot of prompts, a lot of tokens used. But today I I I never achieved my limit. I was thinking maybe going down to the twenty euro plan. I I think it's those membership are are quite good and reasonab reasonably priced for now. The hundred Euros allows me also to to have like agents running, do some stuff, work on other projects at the same time. So so I keep this one for now and if I really feel like it's too much and I don't go near my limit I'll I'll down. Yeah. Yeah. So are you now working full time on the app or do you still have a role somewhere else and this is your side hassle or how is it? So for now it's my my full time thing. I give myself time to try to scale this up and and if it fails then I will have to go back to to a career job. I wanna keep trying for for a few more months. Like you said, you had viral moment on TikTok. If you have one There's always gonna be the next one. So I think it's pretty good. Especially now that you're also expanding to other cities, it it really sounds like it's gonna work out. So if you think about the whole journey, because it's been quite a journey since you started, what has been the most rewarding part of the whole thing? I think getting your first bookings, that was that was for me the I have an alert when when there's a booking. So the first time it comes, you receive an alert saying like, new booking, then it's just like something so cool to celebrate, like you actually made a product that somebody in your city is using to go to a class and now just thinking like yesterday I think there was forty four bookings, and I imagine like my god, forty four people went to a fitness class. After booking it on a on an app that I created, I think it it's pretty cool and it's rewarding to think that yeah, there's actually a lot a lot of users now and and it's working. So so that's the rewarding part. Yeah. And now since you vibe coded an app, has people around you be like, Could I do this too? Yeah actually yes. and I'm the one trying to push people to use AI and I made the guide on like how to set up your workflow and start using cloud to to build an app or even websites or bots or whatever. I've been training families and friends on on on my workflow and we'll see yes so some people are building I'll are building some stuff. if somebody is now listening and they have an app idea what would be your advice for them to do next? I think well downloading the full full workflow, VS Code, GitHub, learn about all those tools. Then when you're ready to start prompting and coding, just like know about the best practices on how to guide your AI agent, as I mentioned before, markdowns, learn Cloud to ask questions, to make sure that it understands the context. And when you when you start chatting with Cloud and take things step by step, then it's gonna lead you one day at a time to a product. And that so that's what how it worked for me. Like I wasn't even sure Lastpot was gonna be a thing when I started. Prompting, but one thing led to another, and then you want to go to the next step and the next step and the next step. So yeah, my advice would be just to like get started, try and see where it goes. What was the moment that you realized it's actually gonna be a product? Like what was the turning point? You were like, there's something here. so because I spent so many hours working on the MVP and I said, Okay, well I I spent so much time on it, so might as well go go full like commit to it. So register a company and get some first clients and because I had to to be on Stripe, I needed to have a business registration number and and so on. So I was like, okay, well if I wanna fully build this app, I need a business registered. And once I did that, because it cost a bit of money as well, then I felt fully committed and and yeah, I wanted to go to the the the next step, which was act finding studios to partner with and then when I had my first studio partner I was like okay this is this is gonna work like I'm gonna find users now. Awesome. So if somebody's listen listening now where can they find more information about you and your app so I post on LinkedIn sometimes so my LinkedIn is Luca Grutter and the app is lastspot.fit there's a web app which with a homepage that explains the concept fully And you can download the app on on the app store. Last spot. I'll try to bring it to more countries as well. So if you're listening and are interested in in an app like this, tell me which country are you from and I'll contact some studios. I will try to to expand it there. and if you have questions about my workflow and like if I can help you set set up those workflow, I I'm happy to answer on LinkedIn. Awesome. Thank you very much, Lucas, for joining today. I really appreciate hearing your story and I wish you all the best. Thank you, thank you Nina for having me. One thing I really liked about Lucas' story is the persistence he had. He did not know how to code, but that did not stop him. He still built the app that he saw that the market could actually use and need. And this was only the second app he's ever built. And I also like him mentioning all the other things that comes with having an app business. The app itself is like a massive job, but once that's ready, the real work really starts. All the other things that are generally not discussed, the boring things like accounting or having legal stuff done and let alone marketing, which nowadays is the moat. Huge thanks to Lucas for sharing his story. He's very inspirational. He's gone from zero users. In less than a year to 14 in September when we were recording this show, and now in October when it goes live, who knows maybe he's over 20,000 or more. You can check that out by downloading Last Spot in the App Store or Android Store and also on the web. I link everything in the show notes below. If you enjoyed this episode, follow App Minded for more stories from builders just like you who decided not to give up. and actually publish their app. Until next time, keep building.