I’ve made the difficult decision to shut down my SaaS Content Goblin. It’s no longer financially viable for me to keep running the software.
Content Goblin was my tool to generate image based listicles, recipes, and Pinterest pins for the articles.
Why I created Content Goblin #
For about 25 years I’ve been into creating and monetizing websites and content. I’ve run affiliate ad campaigns, built content based websites monetized with ads, and software. In 2024, I started to dive into Pinterest monetization. Pinterest was growing fast and sending massive amounts of traffic to publisher websites.
The basic model was:
- Create a listicle
- Pin a bunch of pins to Pinterest
- Send traffic to your website
- Monetize via display ads
I found was having a hard time keeping up with the posting volume necessary to keep traffic coming to the website regularly. Then I discovered a tool called Pin Generator that helped me schedule the pins in bulk, but I still had to manually create the articles.
So I started making my own scripts to help with the article generation part. Initially it was a fairly simple Jupyter notebook that evolved into a web UI.
I joined a community of Pinterest creators called Maverick Mastermind forum and started sharing my method for creating articles and pins. I gave out a lot of free advice. One of the forum partners, Jesse Cunningham, reached out to me because he thought what I was working on was cool.
I gave him a demo of my tool and his initial reaction was “I could totally sell this to my audience.” I had never thought about creating a SaaS tool before this, but I thought I could probably pull it off with my background in tech. I decided to go for it.
The name comes from an article that I believe was posted on TheVerge, but I cannot located it now. The article was about how SEO was ruining the internet. In the article, the author not so lovingly referred to the people ruining the internet as “Content Goblins.” I immediately filed a trademark to use the term Content Goblin in software. So Content Goblin was born.
Building the Tool #
I initially coded the prototype by hand using ChatGPT as a guide to learn Python/Django. Eventually I started using Cursor to help code the site, but the coding models were still not where they are today, so it still took a lot of manual work.
I had experience in Ruby and PHP, band I was not a fan of JavaScript, so I made the decision to go with Python. I learned enough to make a working application that launched with a barebones AI listicle generator. I eventually added AI Pinterest Pin, recipe, and image generation.
Eventually AI coding models got much better and I was able to add new features quickly. I even hired a developer to help me add some much needed features like a Pinterest scheduler.
I had recently left a very demanding job at a fast growing startup called Moveworks to go to another very demanding job at Meta (who later laid me off). I wasn’t sure if I would be able to support the product given that I was working 10+ hours per day.
Marketing and the Initial High #
After meeting Jesse Cunningham, I decided to partner with him to promote Content Goblin to his large YouTube following. I gave Jesse a large chunk of the company to promote the tool. I’ve never done this before, but it seemed like having 100k+ subscribers seeing my tool would be a fair trade for the equity.
I setup a landing page and started collecting emails. Jesse cranked out some videos and I saw a wave of people sign up for the waiting list.
I finalized the MPV and we soft launched the initial version of Content Goblin in November of 2024.
I believe the official launch was around the beginning of January 2025. We quickly hit $15k MRR with the help of Jesse’s YouTube channel. I was riding a high and thinking we were sure to hit at least $30k MRR. We talked with some other SaaS partners and they thought we could get to $100k MRR.
I also setup an affiliate program so that Content Goblin users could promote the tool and earn commision each month.
The Downward Spiral #
We peaked at $15k MRR and had some churn, but that is expected in SaaS, especially AI driven SaaS. The churn never stopped though.
*One interesting note is that Jesse predicted his audience would tap out around $15k MRR. *
As the sole developer and support person, I was a little bit overwhelmed by support issues given that I also was working 10+ hours per day at my day job. I kept up, but I was burning out.
Massive Fraud Attack
In March 2025, we got hit with a massive fraud attack. Scammers joined my affiliate program then used fraudulent credit cards to charge about $50k in subscriptions in a few days. This was due to me not requiring a valid email address during sign up. I did that to remove friction for users.
I caught this rather quickly and contacted Stripe to see what to do about it. They actually told me to wait for the chargebacks, which I thought was absolutely insane. Not only would it be unethical, but the chargebacks could get my Stripe account banned.
So I did the right thing and canceled and refunded all of the fraud charges I could identify. Then refunded any one that contacted our support email claiming they did not sign up for the service.
One absolutely absurd thing that happened was that the affiliate that committed the fraud emailed me to demand I paid him. I told him to go fuck himself.
Stagnation
Users continued to use Content Goblin, but I let it stagnate due to the demands of my job at Meta. The tool was getting stale. At the same time competitors were also launching. I think I was one of the first to market in this niche, but I no longer had an advantage.
Ad Networks Clamp Down on AI Content
Around August of 2025, one of the most popular ad networks that my users used to monetize their websites, Mediavine, started to crack down on AI content and specifically listicles.
Mediavine started banning websites from their tier 1 ad network and their Journey starter network. My own websites were caught in the ban.
Mediavine actually made efforts with a company that I can no longer recall to create an industry wide ban list for these websites. They really wanted to hurt these creators and take away any chance of them making money. I don’t believe many other networks agreed to this ban list, thankfully.
This eliminated the source of income for many of my users, causing them to cancel their Content Goblin subscriptions.
Making the Decision to Shut Down #
Around the time of the initial decline, I did make an attempt to sell the company. It was hard though because we didn’t have a long running history of steady MRR and we were clearly in decline. The site did not sell.
I kept running it and watching the churn month over month. This is the point where I hired a developer to introduce new features. I optimized the landing page from feature base to problem based. I wanted to focus more on the pain the users were having and how we could solve this.
Jesse made the decision to pivot his YouTube channel away from Pinterest as he could see the death sipral also. So I lost that marketing channel. I don’t blame him for making that pivot at all, but I was stuck in Pinterest.
Then I started running Facebook ads to attempt to drive some growth. I got a few conversion but CAC was way too high to make up on the backend. I needed users to stay subscribed much longer than they would to be profitable.
I took a look at my monthly profit and we were barely breaking even. Because I relied on so many other AI model providers, my profit margin was less than 40% which is actually pretty low for SaaS from what I understand.
I had usage based credit system, so users would always pay that margin when they generated an image or article, but I also had fixed costs for the server, database, and image storage. Once the MRR was 1k, I could not longer sustain these costs. We were still slightly profitable, but it was no longer worth my time.
I decided to shutdown Content Goblin in July of 2025.
How I am Shutting it Down #
Shutting down a product that people are paying for is not something I have had to do before, so I needed to think about this carefully. I really cared about the users, but it has to be done.
So my plan is:
-
Remove the option to sign up from the website and landing pages.
-
Immediately cancel all stripe subscriptions
-
Allow the users to continue to use the application until the end of their billing period, at which time their subscription would not renew.
-
I will shut down the main server after all of the users have reached the end of their billing period.
-
This will cause some users to lose images that were initially hosted on Content Goblin’s server but were hot linked on the user’s website.
-
Images stored on Cloudflare Images will continue to be hosted for at least 1 year. I pivoted from local storage to Cloudflare images after about the 4th month.
Then my little buddy Content Goblin will be buried.
Post Mortem #
I wanted to take some time to reflect on what I learned, what went right, and what went wrong.
What Went Right
- I learned that I was able to actually build and support a working SaaS application with paid users.
- Partnering with an influencer that already had an audience was the absolute best thing that I did to market the product. Without Jesse’s audience we would have never hit the peak MRR.
- The affiliate program was the second largest driver of new user acquisition.
- After the initial growth, I decided to raise prices and it did not initially impact conversions.
What went wrong
- I chose a niche that was too narrow and more of a “prosumer” user base. I say that because I don’t believe many of the users were actually making money from their websites or even treating Pinterest as if it was a real business opportunity. They signed up and didn’t give it enough time to see a return before they canceled.
- I made a huge mistake at one point and let one of the early GPT models connect to product and it decided to delete all of the locally user images andI didn’t have a backup of some of the images. I had to email the users letting them know what was lost. - I initially had a hard paywall, but then I decided to add a free trial. Once I added the free trial, conversions to paid account went way down.
- I had an email list, but I didn’t have an onboarding funnel to guide users to use the production.
- I found that the lowest paying users were the most difficult to support. For example, an image with my cheapest model would cost 1 cent. I would have users on the cheapest plan emailing me constantly to request a 1 cent refund because an image wasn’t exactly perfect.
- ChatGPT and other tools caught up. My value prop was really time savings. A user could make an article in 20 seconds instead of a few hours. They could make the same article in 10-20 minutes with ChatGPT. The users didn’t seem to value their time at all.
What I learned
- I overcame imposter syndrome, or maybe not… You could still claim I completely failed here.
- I would target a larger, more B2B oriented niche. Though I do have a half built tool for Facebook monetization that I probably won’t launch.
- I would focus on email list sign up and build and optimize a funnel to drive user conversions and actual use. I saw that some people signed up and didn’t really use the tool.
- I would do more marketing on my own instead of relying on a single channel. With AI given anyone the ability to make software, distribution is truly the moat.
- I would start with higher prices. I would price out the lowest tier / highest support users and focus on trying to acquire serious businesses. This would give me fewer users to support. - I would hire someone to handle the support function. The reason I did not do this with Content Goblin is that we never really had enough profit to do this and I didn’t want to invest the time to build tooling for support and train a support person.
- Many people from lower income countries emailed me asking for discounts. I never gave anyone I discount, ever. I understand they don’t as much money, but my costs are the same no matter where they live. I still would not change prices for other countries.
Conclusion #
Content Goblin was a fun ride and I learned a lot. I stepped out of my comfort zone and learned that it was tough but it didn’t kill me.
I saw all the stories of solopeneurs working from the beach and was totally bought in to that lifestyle. I learned it isn’t as glorious as these influencers make it seem. In the future I would have to think about whether or not I even enjoy running a SaaS or have the time to support one with a demanding job.
I want to thank the users for making Content Goblin an initial success. I also want to thank Jesse Cunningham for believe in me and Tony Hill for sharing all of his Pinterest expertise with me.