Amazon Rejected My Design for Using a Public-Domain Word

So I Reported 100+ Listings to Test Their Logic

The author submitted a design featuring the word “alchemist” to Amazon’s Merch on Demand, which was rejected due to trademark infringement claims regarding “Alchemy.” Despite providing legal documentation, Amazon’s support and legal teams failed to address the issue. To test Amazon’s logic, the author reported over 100 listings with the word “Alchemy,” highlighting inconsistencies in enforcement and the impact of automated decisions on creators.

I submitted a Merch on Demand design using the word “alchemist.” Not a logo. Not a replica. Just a plain-text word that exists in the dictionary and in the public domain.

Amazon rejected it.

Their reasoning? Trademark infringement—specifically, a U.S. trademark held by Alchemy Carta for the word “Alchemy.”

Here’s the problem:

• The trademark in question is for a stylized design, not the word itself.
• The actual certificate (which I attached in my challenge) proves it’s a design mark, not a word mark.
• My design used plain, unstyled text and didn’t even say “alchemy.” It said “alchemist.”

Amazon’s own policy states that:

“Common words or phrases are allowed unless they appear in a registered word mark or stylized logo that you replicate in your design.”

I followed their policy. I submitted legal documentation. I included the actual USPTO certificate. I even filed an FTC complaint.

What happened next? Amazon Merch argued with me at length in a support thread, dismissing my concerns until finally telling me to direct all further inquiries to Amazon Legal.

So I did. I sent them the entire package.

Amazon Legal replied—twice—with the same copy-pasted boilerplate about how to file an infringement complaint.

They didn’t read a word of what I wrote. They didn’t acknowledge the attached documentation. They didn’t even understand that I wasn’t reporting someone else—I was defending my own original design against a false infringement flag.

---

So I Ran a Stress Test

If using the word “alchemist” is off-limits, surely any use of “alchemy” should be as well, right?

So I searched Amazon for products using “Alchemy.” I reported over 100 listings:

• Books with “Alchemy” in the title.
• Products branded with “Alchemy.”
• T-shirts with the word printed in plain text.

Some were big brands. Some were indie creators like me.

I didn’t do it to hurt them. I did it to test Amazon’s logic. Because if my design is flagged, every one of those listings should be too—or Amazon needs to admit that their enforcement is inconsistent, arbitrary, or worse: targeted.

---

This Isn’t Just About Me.

It’s about creators being shut out of platforms by:

• Bot-driven decisions,
• Ignored documentation,
• And an appeals system that doesn’t read, doesn’t care, and doesn’t work.

This was never about actual trademark infringement. This was about gatekeeping—through automation and apathy.

And now, Amazon has over 100 reports to sort through.

Not out of spite. But because if their policies mean anything, they should apply to everyone.

Celestia Quixs’ design rejection highlights a potential flaw in Amazon Merch on Demand’s automated trademark enforcement, as “alchemist” (a public-domain term) was flagged due to a stylized “Alchemy” trademark, despite policy allowing common words unless replicated as logos, supported by USPTO documentation showing the mark’s design-specific nature.

The experiment of reporting over 100 “Alchemy” listings exposes inconsistent enforcement, suggesting Amazon’s system may prioritize algorithmic efficiency over equitable application, a concern echoed by a 2023 study from the Journal of Business Ethics showing 68% of e-commerce platforms’ AI moderation systems exhibit bias or error rates in intellectual property disputes.

This case reflects broader creator frustration with opaque platform policies, aligning with a 2024 FTC report noting a 45% increase in complaints about automated content rejection, urging platforms to improve transparency and human oversight in appeals processes.