Starting a (generative-AI) web app is easier than ever. You can pretty much download a whole framework as open source, start adding something valuable and launch it within a week. Add a website and you can start selling online. Well… kind of. Making small changes can take longer than expected, adding ways for customers to trial or pay is complexer than we thought but nothing compares to marketing.
Marketing is everything. It is the whole thing. It’s the difference between a leather bag and a Louis Vuitton bag and is also the difference between MRR and no revenue. Everything else has a cost to ROI ratio. Need a new feature? Go on fiverr and have it added, whether it’s a ChatGPT function or a way to add coupons for payments, you can find an expert and it will get done, and it will work. But marketing? You can spend 5 million on ads and get nothing in return. It’s the key thing to get right.
In this blog we’ll cover how we are doing it, hint: through with a lot of trial and error.
Where to Start
We knew building everything from scratch would be insane so luckily we had the insight to look for complete components and frameworks. We were actually amazed at how much is freely available online. We settled on Siman CMS to get going because it has a lot of features built in and runs on PHP. We had a good idea of what we wanted to launch so getting an MVP up was not that hard. But it felt like we had built a small shop in the middle of a corn field in the middle of Nebraska… Nobody would ever find us, let alone pay for any features (not that they could).
What was built: an app that lets anyone pull up to 60 good business leads in one go, including any publicly available info like emails, social media links, address. Basically what’s available via Google Maps but easier and faster.
The whole thing is built in PHP, runs on Ubuntu Linux with Apache as web server and mysql as database. Hosted on Digital Ocean. The web site is WordPress.
Connecting with Customers
Forget customers, try users. We hooked up the app to Stripe but there was no trial or free tier so people had to basically bet on the app being useful. This way we were never going to get any MRR. We got some early adopters to try it, mostly by asking for feedback via reddit and from friends and family. But to get more users, we had to build a much more thoughtful onboarding, starting with a free trier of the App.
