Hey Warrior,

In this post, I'll cover everything you need to know to come up with an app idea, build it with AI, and market it to start making money on the internet.

I've just grown my app CopyCoder to $10k monthly recurring revenue in around 2 months. But if I'm being honest, it didn't really take me two months. It took me about 2 years to learn all the lessons I needed to get to this point.

In this post, I'll try to compress that knowledge for you and provide you with a full roadmap.

If you prefer watching over reading, check out this my new video:

Here's what we'll cover in this post:

• Step 1: How to come up with an app idea: 3 approaches

• Step 2: How to build your app with AI: full tech stack

• Step 3: How to market your app and make money: step by step approach

Read time: 8 minutes

💡 Step 1: Coming up with an app idea

There are several ways to come up with app ideas, but here are my three favourite ones:

1. Copying something that’s already working

2. Solving your own problems

3. Doing research on social media

And this ranking goes from easiest to hardest method and also most validated to most unvalidated. Let me explain…

Let's start with the first: Copying something that's already working

For most people, especially if you're starting out, this is the best approach.

Most people think that if you build an app, it needs to be totally unique and original and you cannot have any competitors. I think it's a big mistake.

Because here's how I think about it: Imagine the entire existing space of marketplaces. Within that there are validated problems and there are unvalidated problems.

If a problem is validated, it's very likely that several businesses already exist that are providing a solution to solve this problem and getting customers that pay for the solution. Actually, the fact that these businesses exist proves that this problem is validated!

And then there are a bunch of problems that maybe exist and aren't validated yet.

So as a total newbie, is it easier to just go with a problem that is already validated, or trying to find a completely unvalidated problem and building a solution for that?

Well, I'd argue it's a much safer bet to just take a validated problem and build a solution for that.

Because you'll be surprised how big markets usually are and how unsaturated they are. It's almost never the case that markets are so competitive and so crowded that there's absolutely no space and no customers left in that market.

A good way to think about it is, let's say you wanted to start a restaurant in your local town… 

Is it smarter to start an Italian restaurant, or is it smarter to start a restaurant with some totally unique cuisine that this town has never heard of before?

Well, if you start an Italian restaurant, you already know that people like it, and your job will just be about making the food in that Italian restaurant not totally suck. Whereas if you're starting a restaurant with a totally unique cuisine, it's a much bigger gamble, and your business can be a big success, but it can also be a total failure.

The second approach is to solve your own problems

A lot of people recommend this approach because there are several benefits to it. If you encounter a problem yourself, YOU are an expert in it. And therefore you have a much better perspective on building the product that you're about to build.

Also, because of the scale of the Internet, if you have this problem, then it becomes extremely likely that there are enough other people that have this problem too.

The only downside of this approach is that you have to be careful of biases.

You have to make sure that the problem you're trying to solve is actually a problem. Sometimes, this kind of gets distorted in your mind because of the overoptimistic entrepreneur mindset of wanting to build an app to make money.

So, ideally when you are using this approach to come up with ideas, be very honest and ask yourself, "If this problem existed for me, would I actually pay money for this?"

The more painful the problem is, the more likely it is that you're going to be able to charge money for this.

And finally, the third method is doing research on social media

I think this approach can be a bit harder than the other two approaches, but if you do it right and find something, then it can really be a big hit.

Social media platforms are basically a place for a bunch of people to talk. And when people talk, they talk about their problems, so all you have to do is listen.

Don’t just browse. Make sure you go into the comments, because that's where usually most of the conversation happens and where the gems are hidden. A great platform to do this is Twitter, and also Reddit.

Another thing you should do is to be hyper-aware whenever you see viral content (this is how we came up with CopyCoder).

Because whenever something went viral, there is usually a deeper reason for it. And an entire business can be built on top of a viral concept.

🔧 Step 2: Building your app with AI

In 2025, you don't need years of programming experience to build an app. AI can do most of the heavy lifting, and you can learn the rest quickly using AI tools.

Building is getting easier thanks to AI. The real challenge is coming up with a great idea and bringing it to market.

Here's a simple tech stack I recommend for building a web app as a beginner:

• Cursor: AI-powered code editor

• Next.js & shadcn: web app framework

• Supabase: backend and auth provider

• Vercel: app deployment

• Stripe: payment processing

These are popular, easy-to-use tools. You can optimize later as you gain experience.

Cursor

Cursor is an IDE with powerful AI features. It's more flexible than beginner-friendly tools like Bolt or Replit, making it better for complex apps with integrations.

Next.js & Shadcn

Use Next.js with shadcn UI components as a "template" to quickly build your app's foundation.

Next.js + shadcn is currently the most popular stack for web apps. Next.js also provides many templates to help you start fast.

Or use my tool CopyCoder to quickly recreate app designs from screenshots.

The easiest way to deploy Next.js apps is with Vercel.

Supabase

Supabase is an easy-to-integrate, popular backend provider that includes user authentication. Just ask AI for help with integration.

Payments

Use Stripe to easily accept payments. Lemon Squeezy or Paddle are also good options if you want to avoid dealing with taxes.

One tip: don't offer a free plan at first. Charging from the start validates that people will pay for your product. You can add a free plan later when focusing on growth.

📣 Step 3: How to Market Your Product

Finally, we get to the most important part: how to market your product and get paying customers.

This is the part most people skip. They think if you build a good product, people will just come magically. But doing this well is what will decide whether your app succeeds or not.

I even recommend thinking about this before you build. Go with a marketing-first approach, rather than building-first, and you'll be way more successful.

There are essentially four categories of marketing, based on the framework by Alex Hormozi:

1. Warm outreach (1:1)

2. Cold outreach (1:1)

3. Posting content (1:many)

4. Running ads (1:many)

For marketing your app, let's ignore the 1:1 categories, as they don't scale. Let's also ignore ads for now. (Not ideal to focus on in the beginning, unless you have a lot of experience or a lot of money to invest).

We'll focus on 1:many marketing to warm prospects, AKA creating content.

And here's the order I recommend:

Step 1: Join communities, start commenting

In the beginning, you must do some unscalable things to gain traction.

The best way is to find where your target audience hangs out. If you're building in the health space, find a relevant subreddit. Join the conversation, occasionally mentioning your app.

Building something for expats in Japan? Find YouTube channels that talk to this audience and engage in the comments.

Step 2: Build an audience, leverage viral content

Next, start creating content yourself. This is the step I spent two years mastering.

You could say CopyCoder had an advantage since I'd already built an audience. That's partly true. But nowadays, follower count isn't that important on most platforms. They're moving to a TikTok model where each piece of content is judged independently and can go viral regardless of audience size.

So the key going forward is simply learning to create good content.

And the best way to learn is to study what others are making.

If you see a viral video, dissect it word-by-word. Recreate it in your own style. Same with tweets or LinkedIn posts.

TikTok is probably the easiest place to go viral from zero right now. My first video there got 100K views, the second 800K, all with no followers. It can be similar on Instagram Reels if you find the right format.

YouTube, Twitter, and LinkedIn can be harder to get impressions in the beginning.

Interestingly, once you create viral content, other creators will copy your video and talk about you for free, just because it went viral - this is the power of organic content marketing.

Figure out how to make content go viral and then let others do the work for you.

Step 3: Build an affiliate network

To continue scaling your organic content, build an affiliate network.

This aligns your incentives with other content creators even more: Others create content promoting your product, and you share some revenue with them.

Once you know how to create viral content, building an affiliate network is easier. Your affiliates are incentivized to grow their audiences and make money while promoting you.

Step 4: Optimize SEO

Finally, tap into SEO optimization.

Some might think SEO will be outdated with AI coming. Yes and no. Right now, LLMs mostly learn from SEO results. So there's a high correlation between SEO performance and what shows up high when you ask your LLM.

Also, the reality is Google still drives a lot of traffic. If something works now, I wouldn’t ignore it just because you think it won't work in the future.

The good news is steps 1-3 actually contribute to SEO. You'll get good SEO just by doing those steps well. And if you then want to take it to the next level, you can do keyword analysis, write optimized blog posts or even try pSEO.

Thanks for reading!

If you enjoyed this, consider forwarding this newsletter to a friend or colleague.

What would you like me to write about in the next post? Reply to this email and let me know.

Moritz