DEV Community

Stefan Vasile
Stefan Vasile

Posted on

I made my first $100 and scored 0 out of 34 on my own product, on the same day

Two things happened to my startup on 27 July. Somebody paid me $100 a month, the first money the product has ever made. And I ran my own tool on my own brand and scored zero out of 34.

The tool tells you whether ChatGPT, Gemini and Perplexity name your brand when someone asks for a recommendation. Mine does not name me. Not a low score. Zero.

Here is the whole run, including the part that makes me look bad, and the finding that changed what I work on.

The run

34 prompts, the kind a buyer types before they have heard of anyone:

  • "What's the best tool to track whether ChatGPT recommends my brand?"
  • "What are cheaper alternatives to Profound for AI visibility tracking?"
  • "Compare the top answer engine optimization tools for a bootstrapped founder."
  • "How can I find out which prompts make AI assistants ignore my brand?"

Three engines each. Mention rate 0. Visibility score 0.

The tool worked perfectly. It found every gap, ranked them, and named the winner in each one. The gaps were all mine.

Who got named instead

Brand Answers naming it, of 34
SEMrush 25
Profound 25
Peec AI 23
Otterly.AI 18
Rankscale 4
Scrunch AI 2
Knowatoa 1
Am I On AI 1
Me 0

The part worth your time

The tool also logs the sources behind each answer. Those 34 answers cited 601 pages. The domains the engines kept going back to:

Source Citations
reddit.com 84
youtube.com 68
semrush.com 67
ahrefs.com 52
linkedin.com 46

Two of the top five are a forum and a video site. A third is a social network. Not one is a vendor blog, mine included.

That is the mechanism, and it is duller than the advice everyone sells. When an engine answers "best AEO tool", it reads what people said about the tools somewhere else, then repeats the names it found. A brand nobody has discussed in public gives it nothing to read.

What I got wrong

I spent my first weeks writing pages. Comparison pages, guides, five free tools, clean structured data, a sitemap a crawler can walk. Good work. It moved my number by nothing, because the engines were never short of pages about AI visibility. They were short of a reason to believe I existed.

SEMrush is named 25 times because semrush.com is cited 67 times and because people argue about SEMrush on Reddit. That is the whole thing. It does not respond to another blog post.

What to do if you are at zero too

  1. Measure before you decide anything. Write 20 buyer questions with your brand name nowhere in them. Run each in ChatGPT, Perplexity and Gemini. Count how many name you. Most founders guess this number, and guess high.
  2. Read the citations, not the answers. The sources under the answer tell you where your category gets argued about. Go there.
  3. Earn a mention on a page an engine already cites. One honest comment in a thread the engines read beats ten pages on your own site.
  4. Re-run the same 20 questions in a month. Answers move between runs. One check is a data point; only the trend is a measurement.

The tool

It is open source: github.com/gsmmediaro/fulcru

A CLI, an MCP server and an agent skill. fulcru gaps gives you the questions where an assistant names a competitor instead of you. fulcru write drafts the page that closes one. fulcru delta tells you what publishing it did.

npm install -g fulcru
export FULCRU_TOKEN=pk_...
fulcru gaps
Enter fullscreen mode Exit fullscreen mode

Or in Docker, with nothing installed on the host:

docker build -t fulcru .
docker run --rm -e FULCRU_TOKEN=pk_... fulcru gaps
Enter fullscreen mode Exit fullscreen mode

The measurement engine behind it is hosted and paid, because running prompt sets across three engines every week costs real money. Point FULCRU_ENDPOINT at your own if you would rather replace it.

I will re-run the number and post it again whichever way it goes. If it stays at zero, that is a fact about my product you should weigh before you pay me.

And about that first $100. One customer is not a business, and I am not going to dress it up as one. But it is the difference between a thing I hope people want and a thing one person paid for, and I would rather write that sentence at $100 than wait until it sounds impressive.

The full run, with the numbers kept up to date as I re-measure, lives at fulcru.app/guides/scored-zero-on-our-own-tool.

Top comments (0)