Latest / App Minded / How to Win a Hackathon: I Scrapped My Idea at Hour 36 and Won Lovable SheBuilds
Transcript
- Nina Kolari: Hello everybody, this is another solo episode today. There's a reason why I'm going solo today is because just a couple of days ago I heard I want she builds this year. She builds is lovable hackathon for women. And in this year, over 3,400 women around the world applied. It was over nine hundred women around the world who were building for seventy-two hours. And out of those, six hundred and ninety-five actually submitted, which is phenomenal result, because usually in hackathons, it's a very small number that actually go all the way to submission. So in this episode I'm gonna talk about the hackathon experience itself, why I think I may have won, and why I think you should also consider joining Hackathons. And at the end, I'll also talk about that. I actually won with an idea that I started it in the last half of the hackathon. So what I actually started the hackathon with was not the app that I ended it with. I'm gonna talk about that as well. So let's start with number one. Why did I even join, or why did I apply for She-Bills? One of the reasons is I like competition. They give me a hard deadline to actually execute my ideas. It's a huge help in procrastination because you have a hard deadline that the app needs to be out. I also really wanted to win this. I'm not going to be shy about it. I won two hackathons previously. They were both for iPhone apps and I kinda am known for iPhone apps, but I actually do quite a lot of web apps as well. This allowed me to actually create a web app and showcase my web app building capabilities and you know, broadened scope I have online of what people know of me. and one of the other reasons that I really wanted to join is if you remember Christina a few episodes back, she went to SheBuild's and got honorary mention for her app. So from her, I heard that the the community. that comes together for those seventy two hours is amazing. And she was not wrong. And if you don't know, SheBuilds is application-only hackathon. So you first need to submit an idea that you go with. You kind of give a rough idea what are you going to build, and they'll have a look at it, evaluate it, and then decide if it's go or no go. I've applied twice before, I did not get in. This time I did. And I actually joined with an idea that for an app that I really wanted to do so for quite some time, but I never had the time to do it. it's an app that for a book that I've been dying to write for quite some time. And I thought that maybe if I built the app now, the book actually kind of it forces me to write the book that is in my head. And before I applied, I did not really think about is this a good idea as an app? I just thought it probably will work, but I didn't know the mechanics, how is it gonna be an app? But, anyways, I joined with that. I got in, and when the building started, I was like, Well, I'm locked into this idea, so I just gotta figure it out. So for 36 hours out of those 72. I was hammering this idea, changing it, trying to figure out a way how I can make this book into an app. And it was not working. I kept changing the angle I think four or five times. And I was trying to make it into an app. My lovable account actually has four apps that are kind of started, but I just did not know how I can make it into an app that potentially could win. So I fought the idea for 36 hours. Then I asked, can I change my idea? And Lovable said, Yes, you can, but be mindful, you were accepted based on the idea you had before. I was like, okay, that's not great, but I'm just gonna go go ahead and forget this idea and do something else. One of the biggest reasons why I wanted to change the idea is that I learned that my web app on Lovable can have MCP. And MCP allows my app to be available for agents. So you can connect my to my app with Claude ChatGPT without actually having to go into my app, And once I realized the capabilities of an MCP, because I don't know how to build an MCP and Lovable could do that for me, I knew instantly what product I wanted to create. So at the 36th hour, I took a huge gamble and just scratched all the other ideas and started fresh. Since I knew I have 36 hours left to build this, I have to be super clear what I want, what I don't want. I gotta be strict with the Lovable what it is allowed to do and what is it not, so I can build super fast. And what I built was a basically based on my checklist I have for my apps. I have this massive Google Sheet of all the things that I need to verify and check. and remind me of things that I need to do once an app is published. There's like onboarding, there's marketing, there's whole slew of things that one can check. It is super valuable resource, but it's also Google Sheet. And I'm like, I don't want to use Google Sheet anymore because all my context and my brain is already in these AI tools. Why do I have to jump into Google Sheet? Claude or ChatGPT they can access it, but it's a bit different. As soon as I realized that Lovable can actually create MCP server for me, I was like, I'm gonna use this Google Sheet and make it into an app. So it was basically my knowledge system turned into an app format and cherry on top, I can build an MCP so I don't even have to open the app to use it. And since I also started from scratch, I did not have that couple of weeks of prep time to think about branding and the colors and the UI. I had to really think fast on my feet that what is it gonna be? And I I consulted my AI team on name and the branding and everything else, but I wasn't vibing with anything. I got like growth kit and streamline and all that with random words. It was like nothing was sticking. so while Lovable was building the skeleton of the app, I was thinking about the branding. And the name karrot actually comes from my own marketing experience. If you've done any marketing or growth, you always have that karrot dangling ahead of you because there's always one more marketing thing you can do, one more optimization. there is absolutely no finish line when you work in growth or marketing because it always keeps moving forward and forward and forward. So that's what I decided. I'm gonna call it karrot and on a whim I decided to use K versus C. I don't know where it came from, but that's what I went with. Once I had the name, I got ChatGPT to actually come up with a couple of ideas what the UI, the landing page should look like. And in one of those, there was a picture of a hand holding a karrot and this cutest bunny trying to catch it. And I thought it was absolutely amazing. once I saw it, I knew that was gonna be my hero image. Although it's nothing like the usual hero image you see online. But I thought, what the hell, let's just go with it. If you would like to see the hero image, you can go to heycarrot.com and remember it's Carrot with a K. So what is Carrot? Carrot is really my checklist of things that I do to optimize my apps and optimize my marketing. I built this up really for myself so I can use use the knowledge base and the sheet with an MCP so I don't have to open the sheet. It tells me what's next. I have multiple apps. It's really hard to keep track. Even if it's just one app, there's so many things one can do after it's been shipped. The real work really only starts when you ship an app. but tracking those was getting challenging. so I created this Google Sheet that has all my knowledge base, all the things that I think I should be checking on on with every app. And it solves the problem for me of keeping track of everything I've done and everything I could be and probably should be doing and keeps reminding me of the things that I need to get done. Why is it different from having a ChatGPT to help you? It is not. You could totally do this probably with ChatGPT, but the thing is, you might not know what to actually ask. I've been building and helping others build products online for what over a decade now. I know what you should be asking at what stage. So that is everything that's been mapped out in this Google Sheet and now in karrot So that is my knowledge base in an app format, basically. And another thing is that a lot of people never talk about what happens after the shipping. It's all this talk about I build an app and it took this long and I do that, but nobody really talks about all the work that happens after you ship an app, because that's the never ending loop you will go through. And karrot allows me to go through those loops again and again, one app at a time, and keep track when was the last time maybe I did some edits on the paywall. So it really helps me to keep track because the only way you can make money with apps is by constantly testing what works. And that's what I say to my mentoring clients, I can guarantee you will make success with your app, but that means that you can never stop testing. And by testing, I really mean you test the paywall, you test your onboarding, you might test your branding, you might need to change your name, you might need to set change features. It is a never-ending cycle to make an app. What? Anyways, that's enough about Carrot. Let's talk about how it was to build this. Since I had 36 hours left of this hackathon, I did not have luxury of time to think things through. If a decision had to be made, I just had to do it immediately and move on to the next thing. That also meant because I had limited time, I did not think about doing Supabase or Stripe or any external services. I built everything within Lovable My payments are with Paddle, which is great integration. My database is with a lovable cloud. And I understand that totally locks me in with lovable, but I think twenty dollars is a pretty good investment to have fully functioning web app with MCP server, mind you, to publish to the world. And of course if I wanted to I can always move into Supabase or move it in off lovable completely, but I honestly don't see any reason. lovable helps me to kind of push the technical not responsibility but expertise to lovable. Since I'm using their cloud service, I'm using integrations that they offer, I push a little bit of that responsibility to them. This doesn't mean that I'm still hundred percent liable what happens in the app and what happens with my user data. I think it's not important what tool do you use to build it. All of these tools can pretty much do anything you want. The most important thing is you're using a tool that is created specifically for the purpose of doing that one thing. So you want to find expertise in that tool. And then whatever stops you procrastinating and actually getting the app into the world, use that tool. For many of my mentoring clients, I say pick the tool that you can pick up the fastest and worry about the tech decisions later. and this is particularly for completely non-technical people, have no background of any kind, they've used Claude and ChatGPT as a chat, and now they will build an app. I actually do not recommend going with Codex or Cloud Code or even Cursor And the reason is pretty simple. you're not just learning about how to build an app for the first time. You also are forced to make technical decisions that you have absolutely no knowledge. And we know AI can make mistakes, it can make huge mistakes, and especially on the database, is why I say don't learn. coding off how to create a landing page or something. Focus on learning database and security. When you build with Claude Code and Codex they're phenomenal on that. But they also can make mistakes and if you don't know how to ask, they might not help you. To sum it up, it really doesn't matter what tool you choose to use. They're all phenomenal, they all get to your to a result. The biggest part is that you actually decide and execute. And the faster you start actually building and faster you get into the user's hand, the better. But please be mindful of security and database particularly. How I use AI during this hackathon. whenever I build a new app, I don't just chat to Cursor if I'm doing iPhone app. I chat to Claude and ChatGPT and quite often Gemini and kind of get my counsel to pitch in on what am I missing, what is required, what is not needed. I ask them to do research. look at branding, the whole slew. So whenever I start a project, I start with a discovery process that I built To be used with Claude or ChatGPT. And it's really in-depth interview process where the AI interviews me and tries to figure out my thinking of this app. And at the end of it spits about eight to ten documents depending on what I'm building and those are the documents I start my project in cursor is it's iPhone app or if I'm doing web app it's lovable or something else. And that is really one of the most important things I can recommend to builders is not to start with the building tool initially. Go to ChatGPT and Claude and Gemini and have them interview. what is the things that I'm missing? Can you do some market research? So do all that work before you use credits on your main coding tool. So during this hackathon, I also used Grok bot Or actually used it before to map out my initial idea for the hackathon. So in my Grok bot I have a group chat with my individual bots that come together and help me build this up. So in my SheBuild's group chat, I had my technical lead who is kind of overseeing any kind of technical decisions. I have engineer who is kind of looking at the code. I also have designer and I have product designer and then I had just one bot that was Shebuild's expert. So this group chat I used extensively prior I started building. Of course, since I changed my idea halfway, it was less involved it at this point, but it did do some research for me on the branding and what can I what kind of features I should add on the version one. In that chat, I also discussed the branding, and so I used it quite extensively while I was building just to see. I hate to say opinion, but it is kind of an opinion. on, for example, my hero section or my logo, and I asked the designer in that group chat, like, do you think this is on brand or is this off? How does it align with everything else? I did quite a bit of work with the product designer in that chat as well on the features, how some my admin use user interface actually look like. What is the user journey? So they were mapping those out And if you have not used Grok bot before, I highly, highly recommend it. It's absolutely phenomenal. Now I actually believe on the agents and what they can do for me. And I have about 20 different bots. They're all specialists in their own field. And then I have a chief of staff who's the boss of everybody, including me at this point, and keeps telling me what needs to be done and hiring. each bot for the purpose of the task, whatever that is. If you're interested in learning more about my Grok bot setup, I'm happy to do an episode that please let me know. And I'll get that recorded as soon as possible. what was the actual experience of SheBuilds like? I have to say, with all the rumors of the community being great in these events, it generally was true. I think the season four is by far the largest they've had. I've heard the numbers are quite small in others. We had over 900 women around the world building. That didn't stop us actually connecting on LinkedIn or Instagram or testing our tools or helping each other as we needed some assistance or just genuine support because it is not easy to build an app in 74 hours. lovable host many webinars prior to the event. I was not able to attend because of my time zone, but all the recordings have been available right after the event itself. We also had dedicated hub for all the QA, all the webinar links and everything. It was really well designed and executed event for over 900 people. So I can imagine the team that were responsible, they took a really deep breath after the hackathon was finished. The only event I actually attended live was the final call. I decided to stay awake for that and to hear if I won or not. and in that call there was a sentence that really hit home. It was closest to the problem should be building the solution. And it that's not exact quote, that's what I wrote down after. And it really highlights my thinking as well and what I've been preaching that now because the building or the barrier to build has collapsed. Anybody can build. We are seeing solutions that could have never ever happened, even two years ago. Even one year ago things have changed so fast. and that is also what Lovable would know from the apps that were built. They were built quite often to fix the women's own problem, whether it was life, work, or something else. They were really close to the problem. And I think those are one of the best builders. and those stories you can hear also in the podcast. It is usually builder has an issue or problem and nobody has fixed it and now they decide maybe I can give a crack and do that and there they go. And the same thing is also evident in people I've interviewed past and will be interviewed in the future on this podcast is that people build from their own struggles, which I think is phenomenal. We are gonna get so many interesting apps in the next year or two when more and more people start building from their struggles and from their problems and creating solutions. That could have never existed before. And let's talk a little bit about the competition and how the winner was determined. So with over 900 women actually building, 695 submitted by the deadline. So once our work was done, it was up to Lovable to judge all those six hundred and ninety-five apps. That is not easy beast. I've been judge on a hackathon and I only had seven to evaluate. Imagine having six hundred and ninety-five. Not that everybody watched everybody's apps. I'm sure they were split into groups and everybody checked few apps. I don't know the numbers, but It is a massive undertaking to do. And that had to be done pretty quickly. The hackathon ended on Sunday night. We think got the results, I think on Thursday, of the top thirty apps in this in the competition. And this year lovable had six different categories. So each categories had top five. So I got into the top five in my category along with four other amazing apps. and then once we knew the top five, it was actually up to the community or builders itself to vote who should win each category. So the community voting was for three days. And anybody who participated in She builds this year were allowed to vote, excluding the top thirty themselves. So I did not cast any votes. I was able to only see who are in the top thirty. And this was nerve-wracking time. I was in a category with four other apps. I could not know what people liked, what did they did, what they didn't like. a hackathon judge and average app builder might look at different things. So I had no clue if I was gonna do well or if I was not gonna do well. It was a very interesting time to just see what people were saying. I did get few DMs or people sending me messages that I voted for you, I loved your app, but I was still I d did not know if other people were also getting those. top thirty in She builds really kinda had a couple of things in common. These apps were all quite specific who is the user, what is the user gonna use the app for, and what is the outcome? And I think that is one of the key reasons of winning a hackathon or actually winning customers or clients online. the clearer you are on the target customer, the problem you're solving, and the solution, why you are the right solution for that person, the easier it is to sell online. And I used that thinking when I created my app and particularly on the landing page. I want it to be instantly clear who is it for and how is it gonna help and what might be the end result. and in those top thirty, there was both mobile and web apps. I really love that people also use Lovable to build a mobile app, which is not the usual thing you build with Lovable. but it was really great to see variety. And those top thirty are genuinely phenomenal. And I will see if I can find a list of top thirty publicly and I will list that in the show notes as well. So what would I do differently in a she builds if I was going back again, knowing what I know now? Well, number one would be find a better idea from the get-go so I don't have to change it in the halfway through. I am known to change my ideas in hackathon. so I was not surprised that it happened also at this time, but The more you can prep beforehand, it also of obviously helps. My case was really unique. I I changed my ideas so late that I just thought it in my feet and had to execute like crazy. And that actually also works beautifully because it removes any kind of procrastination. You don't have time to think if I should go orange or purple, you just pick a color and go with it and that's it. And you don't even think about it again. So it is that's why I love hackathons because it puts a hard deadline for you to execute. There's no luxury of time to rethink and reconsider and procrastinate a little bit more what the app should be. If I was doing it again, I would would have asked way sooner if I'm allowed to change the idea. And that would have given me more time. So I of course don't know if my app would have been better if I had the full 72 hours and the week or two weeks before to actually prep everything. Probably the UI would have been a little more polished, but other than that, I'm not really that keen on going improving the visual itself. I have full admin dashboards, I have user dashboards, I have all the data there, and I am not looking to change too many things. I'll have to go through the onboarding and things like that and making sure those are correct. But other than that, I'm not quite sure how much I'm actually changing when it goes live. time will tell how much actually gonna change before it goes live. And the final question is what happens with karrot after SheBuilds? So since I won my category, I will definitely go live with the app. I already have a wait list of people wanting to use the app. So in the next week or two it will go live public for everybody to use. The moment it is being beta tested, which I highly recommend always doing with your apps. So I have complete marketing newbies going through everything in the app and seeing what things they immediately understand and especially what they don't. Because just because I know what something means doesn't mean that the user base actually understands what I'm trying to achieve. So I have complete newbies. tearing my app and trying to break it and find all inconsistencies and things that are not working, not making any sense. So there's still work to be done But having said that, that work is never-ending. There's always the dangling karrot that one has of things that can be and should be improved. it's gonna be a never-ending cycle. I just now have an app that I can use with MCP, with my knowledge base already in place in Claude and ChatGPT to do all those marketing things one one. and of course I'm gonna use my own checklist to keep track of things that I've done, things that I need to be doing, and things that I should be redoing over again. And lastly, should you join a hackathon if you're a non-technical builder or even if you're not a builder at all? I highly recommend you join because We don't have enough non-technical people in hackathons. They are increasing by numbers, which is fantastic. But the number that is not changing enough is women in general hackathons. She builds is, I think it's the largest women-only hackathon, and it does capture thousands and thousands of people every time it runs. But the problem is that we still don't have the floodgates open in general hackathons. You still walk into a hackathon and it's often as a woman you are in a small percentage of the participants. But more women participate in hackathons, the more women will win hackathons, the more apps they're gonna build and launch. And that whole thing snowballs. the more women we get as builders and not just using AI, the better it is going forward. And I do wish if you do build an app, you join a hackathon, talk about it publicly. We really need women's voices in this space. And of course you have to let me know if you join Hackathon. I think that's enough for me recapping SheBuilds It was a phenomenal experience. I highly, highly recommend it. Mine was season four, so when the season five comes along, follow Lovable on LinkedIn or any other socials and I guarantee you'll hear about it and you have enough time to apply. If you have any questions about Hey karrot my app, or any questions about hackathons or anything else, you know where to find me. I'll add all the links to the show notes and if you have any questions about the hackathon experience how to build an app or you have a question that you want me to answer I'm gonna do another solo episode for QA for questions that my listeners submit so I'll add a link how you can do that as well and until then Keep building and keep shipping.