This Self-Taught Programmer Built A $360K/Year GPT-3 App

Published: January 14th, 2022
Malcolm Tyson
Founder, Nichesss
$30K
revenue/mo
1
Founders
1
Employees
Nichesss
from Rio de Janeiro
started December 2020
$30,000
revenue/mo
1
Founders
1
Employees
Discover what tools Malcolm recommends to grow your business!
reviews
other
Discover what books Malcolm recommends to grow your business!
Want more updates on Nichesss? Check out these stories:

Hello! Who are you and what business did you start?

Hey, I’m Malcolm Tyson and I started nichesss, a tool that does 2 things:

  1. Lets people get ideas for businesses to start, and then
  2. Uses A.I.to help you create content for your business, anything from blog posts, ads, and social media posts.

The business brings in ~ $30,000 a month. Our users are made up of copywriters, small businesses, and marketing agencies who need to be able to write content quickly.

this-self-taught-programmer-built-a-360k-year-gpt-3-app

What's your backstory and how did you come up with the idea?

In July of 2020, a company called OpenAI released an API-based AI engine they called GPT-3, which I was eager to try. After finally getting access in November of 2020 by tweeting the co-founder of OpenAI, I knew I needed to make an easy-to-use app that other people could use and take advantage of this game-changing technology.

Having a computer generate text that looks like it was written by a human has major implications for the future, making it a lot easier to get things done.

ALWAYS and CONSTANTLY be testing. Testing messaging, whether it’s in ads or in the application itself. Testing pricing (am I charging too much or too little?)

There were a few other competitors in the space, but I felt the pricing was out of reach for the people I felt were my target demographic. So I set out to create an affordable product while not compromising on quality.

I’m a self-taught programmer and designer, so there was no need to outsource or build a team -- which means I could compete on price. I also recently relocated to Brazil, where the cost of living is relatively low compared to the US, so I didn’t need to price the product outrageously.

I started writing the app in November of 2020, and by December I had a fully functioning MVP. I then set out to get my first customers. I relied on Facebook and Google Ads to get those first few sales. Then in January of 2021 nichesss was listed on AppSumo where it sold $100,000 in the first month.

Take us through the process of designing and prototyping the app

Designing nichesss was a really fun process. The goal was to create an app that requires the user to provide a little bit of input, and then we give you content back. The big challenges were making the ability to talk to GPT-3 easy, and finding users that could benefit from this type of service.

The GPT-3 Artificial Intelligence API asks for instructions and examples. In the nichesss code, we have hundreds of examples of copies we deem ‘good’ for the various tools we have in our app. We then ask our users for information about the content they want to be generated. This coupled with our examples of ‘good copy’ allows the GPT-3 AI to get really creative and generate text that our users can use wherever they see fit.

this-self-taught-programmer-built-a-360k-year-gpt-3-app

With the initial prototype, users would fill out these fields and then the app would generate content. Since writing blog posts are the most popular use case, I added an AI tool that was trained on the Blog Name, Blog Category, Blog Topic, and Blog Post Title. Before I committed to building a full app, I wanted to see if traffic I sent over would be able to understand the app, and ultimately generate content.

It ended up working. So then I added trial limits and a payment form. I sent paid traffic via Facebook Ads, targeted people that liked AppSumo and Noah Kagan (lifetime deal buyers) and lots of people ended up buying it.

this-self-taught-programmer-built-a-360k-year-gpt-3-app

Because I built the app myself, there really weren’t any startup costs, just the standard domain name + hosting costs.

Describe the process of launching the business.

Once I was done testing the prototype and was able to get ‘test’ traffic to convert, I was ready to launch.

Original MVP

this-self-taught-programmer-built-a-360k-year-gpt-3-app

Product Today

this-self-taught-programmer-built-a-360k-year-gpt-3-app

In December of 2020, in preparation for our January AppSumo launch, I started running Facebook Ads which were getting me about $3 back for every $1 I put in.

The biggest lesson I have learned between starting and launching is to ALWAYS and CONSTANTLY be testing. Testing messaging, whether it’s in ads or in the application itself. Testing pricing (am I charging too much or too little?) People were sensitive to the price initially, and comments on Facebook groups complained about how expensive the competition was. So I sought out to charge 80% of what they charge, and bank on word of mouth and eventually – having a large volume of customers.

The way I typically test is to change the text (whether it’s an ad or landing page copy) and monitor the results. What I'm looking for mainly is if the conversation rates change (more sales with the same amount of traffic).

Another lesson I learned is to create a community. Create a place for you to talk with your users, as well as a place for them to interact amongst themselves. For nichesss, that ended up being a Facebook Group. The nichesss Facebook Group has been the #1 place for me to know what features I need to build next to make nichesss a killer AI content generation app.

Having the first-mover advantage is a real thing. Being the first to enter any space has huge implications.

this-self-taught-programmer-built-a-360k-year-gpt-3-app

Since launch, what has worked to attract and retain customers?

Since my product is software-based, my number 1 driver of users has been AppSumo. Since launching and prototyping I haven’t used Facebook or Google Ads, but come next year I will begin using them to scale the business.

Why has AppSumo worked so well to attract users for me?

this-self-taught-programmer-built-a-360k-year-gpt-3-app

I attribute the AppSumo success to the type of customers AppSumo has spent the last couple of years targeting. These users know the value that software brings to their business, so they are more willing to take a chance on an unknown piece of software -- if it delivers on what it promises. Many people ask why I have relied so heavily on AppSumo, even in the face of steep discounts, and the answer is simple. It's a steady predictable stream of new users.

The positioning and ranking of the nichesss AppSumo listing is a function of the various products AppSumo offers. Nichesss is a marketplace deal which means they do limited promotion. So for 30% of sales, I land where I land, but if I give 50% then I will receive a better placement.

Once the AppSumo users have converted, then it becomes retaining and upselling those users. And that’s easily done by keeping an open line of communication via emailing product updates to our list of over 30,000 users and keeping users updated via our Facebook Group.

Our group currently is almost at 10k, and I link to it in each email I send out. I can’t recommend Facebook Groups enough as a way to learn more about your users’ wants and needs, it should be a requirement for every SaaS.

Example of an email update

this-self-taught-programmer-built-a-360k-year-gpt-3-app

How are you doing today and what does the future look like?

Nichesss has been profitable since its inception. The operating costs are fairly simple:

  • Hosting: $200/mo
  • AI Generation Costs: $2,000/mo

Since I’m a solo founder, there are no employees or contractors to pay. We aren’t using paid ads right now, so we don’t have any customer acquisition costs.

On any given month we have about 12,000 users in the app. We get about 100 signups a day and about 16 people a day convert to a paid account. Users spend about 8 minutes per session in the app generating content.

this-self-taught-programmer-built-a-360k-year-gpt-3-app

We have users on every continent. The US is our biggest market, followed by India.

this-self-taught-programmer-built-a-360k-year-gpt-3-app

AppSumo is pretty much the bulk of our distribution. In the coming months, I hope to launch our affiliate program as well as start running Facebook and Google ads to scale the business.

If nobody knows about your amazing product or service, you aren’t going to make any money.

In the future, I’d like to focus on creating tools that let our users create content that puts money in their pocket, like Courses or E-books. This can present another revenue opportunity for us if we bundle Content Creation (what we currently do) with selling content (taking a percentage of the sale of the content we helped them generate).

Through starting the business, have you learned anything particularly helpful or advantageous?

With nichesss, I learned having a first-mover advantage is a real thing. Being the first to enter any space has huge implications. If you can be the first to release a product or release it at a price point that’s very attractive and affordable, try to! I believe that having my offer on AppSumo, at an irresistible price point made all the difference. It got people talking on various Facebook Groups (word of mouth) and I had a huge first month because of it.

Building a community is another thing that I learned is huge. Getting followers really doesn’t mean much to me, it’s the community that matters (Facebook Group + Email List) I was able to rely on that community to get me #2 on Product Hunt and whenever I launch new products within nichesss, I’ve been able to get a good amount of sales because of the community I’ve built.

What platform/tools do you use for your business?

My favorite tools for my business are Google Analytics, Segment, June.so, and Facebook Groups.

Google Analytics is great because it gives me a high-level overview of my users. Where they are from and where they are coming from. With an international app like nichesss, where they are from is important -- I want to be able to have content in my customer's language. Knowing where they are coming from is also equally important. I want to know traffic sources so I can double down on what's working to bring new people to the app.

Segment is great because when users in my app do certain things, I can send that data to other places, like Facebook or Google, for more targeted ads.

June.so is great because I can see who my power users are, where my users are getting stuck, and what features in my app they like the most. It works for hand and hand with Segment.

Facebook Groups are amazing because I can have a conversation with my users, and these conversations appear naturally in the Facebook Newsfeed. It helps build my community and keeps users engaged outside of the app. The nichesss Facebook Group is 9,000 people strong!

What have been the most influential books, podcasts, or other resources?

There are a few books that I think every entrepreneur should read. The Lean Startup by Eric Ries and Traction by Gabriel Weinburg Justin Mares.

“The Lean Startup” is valuable because it really brings home the idea that a startup is really something where you are constantly testing hypotheses. Do people want this? How can I validate that they do? A lot of people waste time building things that nobody wants or that there is no market for.

“Traction” is valuable because it breaks down the various ways you can get users. Building something and thinking people will come is not how it works. Once you have validated that people want your product/app/service, you need to go out there and get users. Traction really breaks down the paid and free methods of getting users and I enjoyed that.

Advice for other entrepreneurs who want to get started or are just starting out?

Nichesss is an app, so my advice will be more geared towards apps. But my biggest piece of advice is getting distribution down. If nobody knows about your amazing product or service, you aren’t going to make any money.

The biggest mistake I see people make is thinking that building an app automatically means you’ll get users or millions of dollars of venture capital money. It’s really important to have a system in place for getting users whenever you want.

The most popular ways to get the word out are Facebook Ads or Google Ads. There is a stat rolling around that says 40% of all Venture Capital money goes towards Facebook and Google Ads, and I believe it. If you can’t convert Facebook or Google traffic, you are going to have a tough time building an online business.

One distribution channel for software that I think doesn’t get enough praise is AppSumo. Many people make a fuss about the cut that AppSumo takes, and don’t take into account how many new customers you may get as a result of a campaign with them. I listed nichesss on AppSumo and in the first month I made $100,000 AFTER their cut.

Where can we go to learn more?

If you have any questions or comments, drop a comment below!