I've built two apps before, and neither of them gave me the result I wanted. The funny part is that my goal with both of them was ridiculously small: make $1. Now, thanks to Shipaton 2026, I'm building my third app at a much faster pace. The iOS version is already live, and I'm writing this while waiting for the 14-day closed testing period on Google Play. I want this first post to be a kind of postmortem. Why did my previous apps fail? Why did I choose this idea? What am I doing differently this time? And what comes next?First, the two apps that came before1. A Chrome extensionThe first one was a simple Chrome extension for managing draft real estate listing templates. It was the definition of an MVP. There wasn't even a payment flow. I just added a Buy Me a Coffee link. The product was so simple that if you made it compete with a random notes app, I'm honestly not sure it would win. It worked. That was pretty much it.2. A clothing size tracking appThe second app was something I personally needed a lot. The idea was simple: keep track of which clothing sizes fit you across different brands and products. The problem was that I never managed to define the target user properly, so the product stayed in this weird gray area. I didn't want it to become another digital wardrobe app, but at the same time, I didn't build enough around the simple idea of tracking clothing sizes. Users could add the brand, model, and size as free text. There are so many different sizing systems around the world that instead of trying to build a system that supported all of them, I chose the easiest solution.Again, it was a very MVP-ish product. So MVP-ish that it was basically competing with a notes app too.But I didn't give up on it quickly. I shipped nine updates, translated the app into multiple languages, created separate App Store pages for different languages, kept changing the keywords, and experimented with ASO. And after all of that? Not a single person subscribed. My premium model was basically: you can save up to X clothing records for free, and then you have to pay. The problem was that nobody even created one clothing record.That was the real failure. Since the beginning of this year, the app has received only 22 downloads. So yes, nine updates, multiple languages, ASO experiments... 22 downloads.The problem wasn't just marketingAs you can probably tell, neither of these products created enough meaningful value for the user. I still like both apps. I especially think the second one could genuinely be useful to some people. But the product was simply too MVP. There wasn't a particularly good UX, the UI wasn't polished enough, the feature set wasn't carefully designed around the user's experience, and there weren't any extra features that made someone think: “Yeah, I'd actually pay for this.”So for Shipaton, I decided I wanted to change that.I saw two separate problems:Getting traffic to my App Store page through ASO, marketing, or other channels.Making sure that the person who downloads the app is actually the right user and immediately understands the value of the product.Those are two very different problems. And solving only the first one doesn't mean much if the second one is broken.This time, I didn't want to build another MVPBefore choosing an idea for Shipaton, I looked at the winning projects from last year's Shipaton and Shipyard. One app in particular really stuck with me: Preplo, the Shipyard winner. If someone had described the idea to me before they built it, I probably would've said: “There are already plenty of alternatives. I'm not sure you need to build another one.” But then I actually used it, and it was incredibly polished. What impressed me wasn't some giant groundbreaking feature. It was the small details. They had clearly thought about the actual situation the user would be in while using the product. For example, they even considered that someone cooking might not want to touch the screen with dirty hands and designed around that. That kind of detail stayed with me. When I looked at other successful Shipaton and Shipyard projects, I noticed something similar. It wasn't only about having a good idea. The way the idea was executed also made the product stand out. And now that AI makes development dramatically faster, I started thinking that maybe MVP isn't enough anymore. At least not for the kind of products I want to build. So this time, I decided to think in terms of an MLP: Minimum Lovable Product. Not a product with every possible feature. Not something overloaded with technology. Just the smallest version of the product that could still make someone think: “Oh. This is actually nice.”Finding the ideaBetween August 1 and August 10, I tried to generate app ideas from basically everything I came across during the day. Whenever I noticed a problem, an annoying workflow, or something interesting, I wrote it down. By the end, I had around 20–25 ideas. Most of them were eliminated pretty quickly. Shipaton's categories were obviously one of the filters. But there was one idea that kept coming back, and I kept avoiding it. I liked the idea, but I thought it would take too much effort to build properly, so I didn't want to commit to it. Still, somehow, it kept showing up again. Then I started talking to people on Reddit about building apps and marketing. One thing became very clear: how marketable the product is should be part of the product decision itself. That sounds obvious now. But considering that one of the biggest problems with my previous apps was getting people to them in the first place, it mattered a lot. And that's when the idea that I kept rejecting came back again: Video Cue: Shot Planner.Why Video Cue?I first discovered CapCut because of the little CapCut branding at the end of videos created with the app. I had seen videos made with it before I ever intentionally searched for CapCut. That made me think about Video Cue differently. What if the product itself naturally created marketing? If someone used Video Cue to create a video and that video performed well, the result itself could lead other people back to the product. One good video could become distribution. That was something my previous apps never had. So eventually, I stopped fighting the idea and started building it.Building the MLPThe idea behind Video Cue is pretty simple: help someone pick up their phone and create a decent piece of video content within a few minutes, using simple but effective guidance. While building it, I had plenty of ideas for things I could add: augmented reality, local LLMs, more AI, more complicated editing, and plenty of technically cool things.And I deliberately didn't add most of them. That's another thing I'm trying to do differently this time. MLP doesn't mean adding more features. For me, it means spending the limited development time on the things that create the biggest difference for the user. So instead of choosing the most impressive technologies, I tried to choose features that were relatively cheap to build but genuinely useful in the actual workflow. The user shouldn't care that there's some cool technology underneath. They should just feel like the app understands what they're trying to do.Where I am nowA few days ago, Video Cue went live on the App Store. The Android version has also entered Google's 14-day closed testing period. And no, I still haven't received that magical first subscription notification. So this isn't a success story. At least not yet. But for the first time, I feel like I've built something that provides significantly more value than my previous apps. I've already used Video Cue to create UGC for a small business, and the finished video was published on the business's own account. That alone was a nice moment. Someone actually used the result of the product for something real.Right now, I'm fixing bugs, looking through UGC creator communities, trying to understand how I can get the product in front of the right people, and figuring out what I can do with Noise. Maybe I'll be back here in a few weeks saying this one failed too. That's completely possible. But at least this time, the question isn't: “Does the app work?” The question is: “Did I build something people can actually care about?”That's what I'm trying to find out.See you in the next post, hopefully with another problem solved.