Flash Genius Case Study
A case study of my first failed SaaS product I built during college.
Background
During college, I had a friend who did some networking in the tech scene in Seoul. One guy he met built an "MBTI" guesser app and was able to get it to $500 MRR (allegedly).
Inspired by that story, we wanted to try building our own app. Finding the right idea is truly the million-dollar question. Since we both lived in Korea at the time and were language students there, we decided to build a language learning app that had the features we might want to use in a language app.
Building
I made the architectural decisions for the app. We wanted to deploy the app onto mobile app stores, particularly iOS since that gives you access to Apple's walled garden. Plus, stats show that Apple users pay over double what Android users spend on Google Play, so it's the richer market. Considering that, and the fact that our experience was largely in JavaScript, I chose to use React Native with Expo as the main coding framework, since it can transition between web and a native mobile app without needing to write any Swift or Kotlin.
For a couple of weeks, we would meet in Yeouido and code it up. I introduced my friend to Cursor, and it was both of our first times using agentic coding tools. Back then, the models made a lot more mistakes and needed to be hand-held much more to get meaningful work done.
We built a chat window that let the user type in a phrase in their native language and select their target language. The app would take their string, pass it to the OpenAI API, and return the translated sentence as well as a breakdown of the words and their individual parts of speech, i.e., noun, verb, adverb, etc.
We also ran some paid Meta ads to get users. We did get some free users to use the system, but as soon as they hit the paywall, they bounced. The value proposition must not have been good enough, and the pain point not urgent enough to get the wallet to come out.
I even used screen capture to do a walkthrough of the app's main workflow, edited it, and used it as marketing material, but nothing landed.
Summary/Takeaways
All in all, we spent weeks grinding out features that nobody asked for and were not able to get any paid users.
The key takeaway is that with better marketing, I still believe it can work to this day. Not just marketing, but doing market research on your target niche before writing a single line of code to see if there is paying demand, whether there is product-market fit, who the ideal customer profile is, and so much more that I was not taught in my Software Engineering degree.
If you are interested in trying out the app its still up at https://flash-geniusreact.web.app/home.