 |
| Research the pattern, not the product. Compare multiple sellers, study search trends and buyer language, look for gaps, then turn what you learn into an original t-shirt idea instead of a copy. |
There’s a useful way to research successful t-shirts, and there’s a lazy way.
The lazy version is finding a shirt that appears to be selling well, changing the font, moving the flowers around and producing your own slightly altered version.
Apart from the obvious intellectual-property problems, that doesn’t teach you very much. You’ve learned what somebody else made. You haven’t learned why shoppers wanted it.
The better approach is to research the pattern behind successful shirts: the subject, buyer, occasion, mood, wording style, colors, layout, product type and search language. Then you take those clues somewhere different.
That’s the kind of research I want to use while developing Public Nuisance Apparel. I’m much more interested in figuring out why several shirts seem to work than recreating the one sitting at the top of a marketplace page.
First, Don't Confuse “Visible” With “Best-Selling”
This is an important distinction because marketplaces don't always tell us how many units an individual shirt has sold.
A product appearing near the top of search results may be there because of relevance, personalization, marketplace promotion, advertising, recent activity or other ranking factors. Even on Etsy, which now gives sellers access to Marketplace Insights, Etsy explains that marketplace search ranking uses numerous signals and that results can be personalized to individual shoppers. (Etsy)
So when I browse Zazzle, Etsy or another marketplace, I don't automatically label the first ten products best sellers unless the platform actually gives me evidence that they are.
What I can say is:
These designs are prominent.
These themes appear repeatedly.
These products have visible engagement.
This keyword seems to have demand.
Several successful-looking sellers are approaching the same audience in different ways.
That's enough to start useful research without inventing sales data we don't have.
Research the Market, Not One Shirt
Looking at one successful design is where copying starts to become tempting.
Looking at fifty designs turns it into research.
Suppose you're considering a book-lover shirt. Instead of finding one shirt that says something clever about reading and using it as your model, look across many results.
You might begin noticing patterns:
vintage typography appears frequently
dark green, burgundy and cream recur
some designs target romance readers while others target general book lovers
stacks of books and botanical illustrations show up often
short phrases dominate over long quotes
certain designs look more like gifts than everyday fashion
Now you have useful information.
None of those observations requires copying somebody's illustration.
For Public Nuisance Apparel, that kind of research could inform future products for the Book Lovers collection without telling us what the actual artwork should look like.
The market gives us the question.
We still have to create our own answer.
Break Successful Designs Into Components
One trick I've found useful is mentally removing the artwork altogether.
Instead of asking, How can I make something like this?, ask what the product is doing.
A shirt can usually be broken down into several decisions:
Audience: Who is it for?
Subject: What is it about?
Emotion: Funny, nostalgic, sarcastic, sentimental, proud, spooky?
Occasion: Everyday wear, birthday, Christmas, Halloween, Mother's Day?
Style: Retro, minimalist, distressed, botanical, gothic, preppy?
Composition: Big central graphic, typography only, badge, illustration plus phrase?
Color palette: Bright, muted, monochrome, autumnal?
Search intent: What might somebody type to find it?
That's enough information to learn from the marketplace without borrowing the protected expression itself.
If I saw ten popular-looking black-cat shirts, for example, I might conclude that black cats combined with vintage Halloween imagery are worth investigating for Black Cat or Halloween.
That does not mean copying someone's black cat pose, wording, border treatment or composition.
Look for Repetition Across Different Sellers
One seller can have an unusual hit.
When unrelated sellers independently keep targeting the same idea, it becomes more interesting.
Maybe you repeatedly see:
retro botanical graphics
bookish humor
martini illustrations
witchy celestial artwork
bow-heavy feminine designs
pickleball jokes
grandma gifts
birth-year nostalgia
That repetition suggests a broader demand pattern rather than one designer's private success.
This is also where official Zazzle material can help. The current Creator Hub includes 2026 guides built around trends, color palettes and major occasions, while Zazzle Ideas regularly publishes current seasonal and design inspiration. (Zazzle)
I wouldn't treat an official trend guide as an instruction sheet for what to design.
I'd treat it as another piece of evidence.
If Zazzle is highlighting a theme, Google searches are increasing, and I keep seeing related products across several marketplaces, now the idea has more support behind it.
Use Search Suggestions as Research
Marketplace search boxes are useful because they often expose the language shoppers use.
Start typing a broad subject.
Instead of simply searching:
cat shirt
try variations such as:
black cat shirtvintage black cat shirtfunny black cat shirtblack cat halloween shirtblack cat mom shirt
Then watch what related language appears.
You're not only looking for an artwork idea. You're looking for buyer intent.
A person searching black cat Halloween shirt is probably shopping differently from someone searching minimalist cat shirt.
That distinction can affect the illustration, product title, keywords and even the shirt color.
It's also much more useful than staring at somebody else's artwork and wondering which part to imitate.
Google Trends Can Tell You Whether an Idea Has Legs
Google Trends is one of the better free checks because it lets you compare search interest over time, by location and across related searches. Google also distinguishes between exact search terms and broader topics, which is useful when you're trying to determine whether you're studying one phrase or a larger subject. (Google Help)
Its related-search section is particularly handy. Google separates related searches into Top and Rising, and a “Breakout” label means the search has grown by more than 5,000% compared with the previous period. (Google Help)
That doesn't mean every rising search should become a shirt.
Plenty of trending searches are news stories, people or events that either make terrible merchandise or create intellectual-property problems.
But Trends can help answer questions like:
Is this interest seasonal?
Is it growing or fading?
Which wording do people use?
Does one version of the idea have considerably more interest than another?
Is the supposed “trend” actually just something POD sellers keep talking about?
That last question is worth asking more often.
Etsy Can Be Useful Even If You Sell on Zazzle
You don't need to sell on Etsy to learn something from Etsy's marketplace.
Etsy introduced Marketplace Insights using real Etsy search data. Sellers can see search activity, related terms and the number of competing listings over the previous 30 days. Etsy specifically describes it as a tool for spotting buyer demand, trends and areas where strong search interest may exist alongside lower competition. (Etsy)
That is much more useful than assuming a style is popular because you've seen it repeatedly on Pinterest.
The markets aren't identical, of course. A phrase that performs well on Etsy won't automatically sell on Zazzle.
But if you're researching a subject for Graphic Shirts & Sweatshirts, seeing real shopper search behavior somewhere else can help determine whether an idea deserves further investigation.
Think of it as supporting evidence rather than proof.
Pinterest and Social Media Are Good for Style Signals, Not Sales Proof
Pinterest, Instagram and TikTok can tell you a lot about visual direction.
They can show you that:
a color combination is spreading
a particular illustration style is becoming common
a decade is being revived
certain hobbies are having a moment
a seasonal look is developing earlier than expected
What they generally cannot tell you is that somebody wants that trend printed on a t-shirt.
That's where POD advice sometimes goes wrong.
Something can be huge aesthetically and still make mediocre merchandise.
I'd rather use social media to discover a visual direction, then check marketplaces and search data to see whether there is actual shopping interest behind it.
Read the Words Before Studying the Pictures
Typography shirts make copying particularly easy because sometimes the “design” is largely a phrase.
That is also an area where you need to be careful.
Instead of collecting exact sayings, classify the type of saying.
For example:
not this specific martini joke
but dry cocktail humor aimed at women
not this particular reading slogan
but introverted book-lover humor
not this exact Christmas phrase
but anti-cheerful holiday sarcasm
That is far more useful creatively.
For Cocktail & Martini, I might research what kind of humor seems common around cocktail apparel without taking somebody's exact line.
Then we can write our own concept from scratch.
Trend Research Is Not a License to Use Somebody Else's Intellectual Property
“Everyone else is selling it” is not a rights-clearing process.
Zazzle says it respects third-party intellectual property and may remove content or terminate users who infringe copyright or other rights. Its AI rules go even further in stating that AI-generated content should be novel and not knowingly copied from existing designs. (Zazzle)
And there's an important distinction in U.S. copyright law that is helpful when researching merchandise.
Copyright generally protects original creative expression, but not the underlying idea itself. The Copyright Office gives the basic example that ideas aren't protected merely as ideas, while the particular creative way someone expresses one can be. (U.S. Copyright Office)
So:
“A retro shirt about gardening” is an idea.
Someone's specific flower illustration, arrangement, lettering and artwork are their expression.
That's one reason researching broad patterns is so much safer — and creatively more useful — than remodeling a competitor's finished design.
Trademark is another issue entirely, particularly with brand names, franchises, celebrities, teams and phrases being used as source identifiers. That's something to check separately before publishing.
Make a Research Sheet, Not a Swipe File
I don't particularly like the idea of keeping hundreds of competitor designs as “inspiration.”
It's too easy to absorb them more literally than you realize.
A research sheet is better.
Instead of saving the design, write down what you learned from it:
| What I noticed | What I might investigate |
|---|---|
| Several retro book shirts use dark academia colors | Test whether bookish designs work better in dark muted palettes |
| Black-cat shirts rise around Halloween | Compare evergreen cat designs with seasonal Halloween versions |
| Martini graphics often use mid-century illustration | Research whether retro cocktail art is still growing |
| Short phrases dominate | Test short copy against illustration-only designs |
| Cream shirts appear often with vintage graphics | Compare dark versus natural garment colors |
Now you've separated the observation from the artwork.
That's exactly what you want.
Look for Gaps, Not Just Winners
The most interesting research question isn't always:
What's selling?
Sometimes it's:
What's missing?
Suppose you search for a topic and find hundreds of similar designs.
That tells you there is interest.
It also tells you the obvious version has already been done repeatedly.
Look for what those listings aren't doing.
Maybe everything is:
very feminine
extremely colorful
typography-heavy
aimed at younger buyers
focused on one joke
using the same retro palette
That creates possible directions.
Could the subject work in a darker style?
Could it work without words?
Could it target a different recipient?
Could it combine naturally with another interest?
Could it work better on a tote or Case-Mate phone case than another shirt?
That is where research starts generating original ideas instead of clones.
Combine Two Signals Instead of Following One
I like this approach because it creates distance from the source material.
Suppose the research tells us:
Signal 1: vintage botanical artwork keeps appearing.
Signal 2: book-lover merchandise has steady demand.
Instead of copying a popular botanical book shirt, we could ask:
What could botanical + bookish look like in the visual language of Public Nuisance Apparel?
Or:
Signal 1: black cats remain popular.
Signal 2: retro celestial graphics keep appearing.
That might suggest an original direction for Black Cat.
You're building from market information rather than rebuilding somebody's product.
Don't Copy the Layout Either
People tend to think copying only means tracing artwork or stealing a phrase.
There's a subtler version.
Same phrase structure.
Same font style.
Same arch.
Same flower placement.
Same colors.
Same central illustration.
Same tiny stars around it.
No individual piece may look dramatic on its own, but the finished design can still feel suspiciously close.
When research starts influencing too many individual decisions in the same direction, stop looking at the reference.
Sketch from memory.
Change the composition.
Change the focal point.
Introduce a different visual idea.
Better yet, research several unrelated references before designing anything.
AI Doesn't Make Copying More Original
This is especially relevant now.
Taking somebody else's shirt, describing it to an image generator and asking for “something similar but different” doesn't magically turn the concept into original research.
The AI may change the flowers and redraw the cat, but you still started by trying to reproduce somebody's finished solution.
A better AI prompt starts from your research notes:
Retro black-cat artwork with a 1970s storybook feel, autumn palette, bold silhouette, designed for a dark sweatshirt. Avoid existing slogans. Create an original composition.
That's a creative brief.
Very different thing.
And for anything eventually headed into Public Nuisance Apparel, we'd still inspect and rework the generated result rather than assume AI automatically made it original or production-ready.
What I Would Research Before Designing a New Shirt
If I were starting a shirt idea today, my process would be roughly:
Pick a broad niche or theme.
Browse enough marketplace results to identify recurring patterns.
Write observations instead of saving designs.
Check Zazzle Creator Hub and Zazzle Ideas for related current themes. (Zazzle)
Check Google Trends for seasonality, wording and related searches. (Google Help)
Look for evidence of buyer searches elsewhere, such as Etsy Marketplace Insights when available. (Etsy Help)
Identify an underserved angle or an interesting combination.
Stop looking at competitors.
Create the design from the research brief.
Check intellectual-property issues before publishing.
The important part is step eight.
At some point, research has to stop.
Otherwise you can spend so long looking at what's already selling that it becomes difficult to imagine anything else.
How We'll Use This at Public Nuisance Apparel
This gives us a much better way to decide what Public Nuisance Apparel should create next.
Instead of asking me to find a best-selling shirt and make our version, we can pick one of the existing collection areas — Book Lovers, Black Cat, Cocktail & Martini, Mahjong, Halloween or Vintage & Retro Style — and research the surrounding market.
Then we can build a short creative brief from what we find.
Not:
Here's the shirt we're going to copy.
But:
Here's the audience. Here's what they seem to search for. Here are the styles appearing repeatedly. Here's what's overcrowded. Here's something I don't see enough of. Here's what we're going to test.
That's a much more useful experiment.
And eventually it gives Flowshirt something better to report than “we followed a trend.”
We can show the whole chain:
Research → Idea → Design → Listing → Test → Result → Adjust.
That's where this gets interesting.
