I sold a $29 template. It became my agency, nad next APP
One years ago I sold a $29 n8n template on Gumroad. It outsold the SaaS I had spent six months building. That single Gumroad listing turned into a three-brand stack I now run solo from Poland.
Here is the actual sequence, with the numbers I can verify.
THE FAILED PROJECT
I had shipped a B2B tool. Forget the name. It got 11 signups. Two paid. Zero retention. I was burned out and broke.
On the side, for my own sanity, I had built an n8n workflow that researched a topic, drafted a 1500 word blog post, and pushed it to my WordPress draft queue. It was glue code, not a product. I wrote the README in 40 minutes and put it on Gumroad for $29 because I needed coffee money.
THE $29 SALE
It sold 14 copies in the first week. The dead SaaS sold zero in the same week.
The interesting part was the email after the sale. Three different buyers wrote some version of: "I cannot get this running. Will you set it up for me for $500?"
I said yes to all three. That is when I learned the template was not the product. The setup was the product. The template was the lead magnet.
THE SERVICE SPIRAL
Over the next four months I took 9 setup gigs at increasing prices. $500 to $2,800 each. Every client wanted small variations. Different research sources. Different brand tone. Different CMS targets.
I started naming this work Lumizone. One years later Lumizone has shipped 1,000+ articles across 9 active retainer clients. One pest control client got +15% organic traffic in a 6 month window. One peptides e-commerce client got +19%. These are clients in industries nobody wants to write about, which is exactly why the SEO is winnable.
THE 80 PERCENT MOMENT
After the fifth custom build I noticed something. The architecture was the same on every project. Research stage. Outline stage. Per-section briefing. Draft. Score. Publish. Refresh. The only things changing were the brand voice profile and the CMS target.
I had built the same pipeline 5 times.
So I extracted it. Same architecture, no per-client work, one dashboard, one price. I called it Articfly. The first 50 users came from existing Lumizone clients who wanted the cheaper self-serve tier. articfly.com is the product I would have built first if I had known what to build. 50,000+ articles have shipped through it since.
WHAT THE GUMROAD SALE ACTUALLY TAUGHT ME
3 lessons, in the order I learned them. None of them are what I expected.
1. The market tells you which problem is real. You do not get to vote. My polished SaaS got 11 signups. My 40 minute Gumroad listing got buyers asking to pay 100x for setup. I had been working on the wrong thing for 6 months.
2. Productize the bottleneck, not the tool. The tool was easy to copy. The bottleneck was setting it up correctly for a specific brand. That was Lumizone for 2 years. The SaaS version only worked once I had productized the bottleneck itself: brand voice analyzer, automatic CMS adapters, per-section briefs.
3. Service to product is a one way street, but it takes longer than you think. I should have started extracting Articfly after client 3. I waited until client 5. That cost me roughly 8 months of compounding.
The current shape: Lumizone for custom work where the client wants a hand on the wheel, Articfly for everyone else who wants a product they can run themselves. Same engine underneath. Different price points. Different buyers.
I would have laughed in 2024 if you told me a Gumroad listing I wrote during a bad week would split into two businesses.
Question for anyone who has done this loop: at what point did you stop taking service work? I am at the point where Lumizone retainer cash is high but Articfly compounding is real. Pulling the plug on the agency feels both obvious and terrifying.
