 |
| Before you change a design, figure out where the sales process is actually breaking down. This funnel helps you diagnose whether the problem is visibility, clicks, conversion, or simply not enough data yet—so you can test the right thing instead of changing everything at once. |
Publishing a product and getting no sales is frustrating.
What makes it worse is not knowing why.
Was the design wrong?
The niche?
The keywords?
The price?
The product?
The mockup?
Or did hardly anyone see it in the first place?
That’s why I don’t think “this shirt isn’t selling” is a useful diagnosis.
It’s the result.
The useful part is figuring out where the process is breaking down.
Instead of changing everything at once, I’d rather test one part of the listing at a time and use whatever data I have to decide what deserves attention next.
Here’s what I’d look at.
Start With the Most Important Question: Is Anyone Seeing It?
Before redesigning anything, check whether the product is actually getting views.
This is exactly where I am with some of my early Public Nuisance Apparel products.
If only a handful of people have seen the listing, there really isn’t enough information yet to conclude that the design is the problem.
A product with:
10 views and zero sales
is in a completely different situation from one with:
1,000 views and zero sales.
Both have zero sales.
But they’re telling you two very different stories.
If hardly anybody has seen the product, I’d start with visibility.
If plenty of people are seeing it and nobody is buying, then I’d start looking at conversion.
That distinction can save you from redesigning something that hasn’t actually been tested yet.
Test #1: Your Keywords
If the listing isn’t getting traffic, keywords are one of the first places I’d investigate.
Ask yourself whether you used the language a shopper would actually type.
A clever design phrase might be great on the shirt but weak as the main search term.
For example, a product called:
One More for the Afterlife
doesn’t tell the search system much by itself.
A shopper might be looking for:
skeleton martini Halloween shirt
Halloween cocktail shirt
women’s spooky shirt
retro skeleton tee
The branded phrase can still appear in the title.
It just shouldn’t have to carry all of the search work.
If traffic is low, I’d compare the listing language with actual marketplace searches and see whether the product is being described the way shoppers describe it.
Test #2: The Product Title
The title should quickly tell both the shopper and the search system what you’re selling.
That doesn’t mean stuffing every keyword you can think of into it.
It means being specific.
Compare:
Bad Omen Good Taste
with:
Women’s Retro Black Cat Halloween Shirt – Bad Omen Good Taste
The second title immediately gives us:
the audience;
the style;
the subject;
the season;
the product;
and the design phrase.
If a listing has very little visibility, I’d look at whether the title is doing enough descriptive work.
Test #3: Your Tags
Tags can help fill in the search language that doesn’t fit naturally into the title.
If you’re selling a Halloween skeleton martini design, useful tags might cover several related ideas:
skeleton shirt
martini shirt
Halloween cocktail shirt
spooky season
retro Halloween
women’s Halloween apparel
The important part is relevance.
Adding popular terms that have nothing to do with the product may bring the wrong visitors—or no useful traffic at all.
If the tags all repeat the exact same phrase with tiny variations, I’d test a broader set of relevant searches.
Test #4: Whether the Niche Has Real Demand
Sometimes the listing is fine.
The idea just isn’t something enough people want.
This is why research before designing is so useful.
If a product gets very little traffic even after the listing has been optimized, go back and look at the niche.
Are shoppers searching for this type of product?
Do similar products appear to be selling?
Is the niche active?
Or did we create something because it sounded clever without finding much evidence that anyone was buying it?
No amount of SEO can manufacture demand that doesn’t exist.
That doesn’t mean the design is bad.
It may simply mean the audience is too small.
Test #5: Competition
The opposite problem can happen too.
There may be plenty of demand—and thousands of sellers fighting for it.
If you search your main keyword and see pages of extremely similar shirts, your product may be buried.
That’s especially likely with broad terms such as:
Halloween shirt
teacher shirt
mom shirt
cat shirt
Going narrower can help.
Instead of competing for:
Halloween shirt
you may have a better chance targeting:
women’s skeleton martini Halloween shirt
You’re trading some search volume for relevance and lower direct competition.
Sometimes that’s exactly the trade you need.
Test #6: The Thumbnail
If the product gets impressions but very few clicks, I’d look closely at how it appears in the search results.
This is different from asking whether the full-size artwork looks good.
Shoppers usually encounter the product as a much smaller image first.
Can they tell what it is?
Can they read the main phrase?
Does the artwork have enough contrast?
Does it stand out from the surrounding products?
Or does everything turn into one little blob?
A design can be beautiful at full size and weak at thumbnail size.
If impressions are coming in but clicks aren’t, that’s one of the first things I’d examine.
Test #7: The Design Itself
Now we finally get to the artwork.
If people are seeing the listing and clicking on it but not buying, the design becomes a more reasonable suspect.
Ask:
Would I actually wear this?
Is the wording readable?
Does the design feel polished?
Is it too busy?
Does it look like something already available everywhere?
Is the style appropriate for the audience?
Does the design look good on the garment?
Are there obvious AI artifacts?
Would someone understand the idea immediately?
This is where candid critique becomes useful.
Not:
Do I personally like it?
But:
Does it work as a product?
Test #8: The Product Choice
A strong graphic can still be attached to the wrong product.
Maybe the design feels better on a sweatshirt.
Maybe the shirt style looks too fitted.
Maybe the available garment colors don’t suit the artwork.
Maybe the product itself is priced higher than comparable alternatives.
If your platform offers several garment choices under one design, look through them.
Make sure the graphic sits well across the important options.
A design problem and a product problem can look very similar from the sales dashboard.
Test #9: Placement and Scale
This one is easy to overlook.
The artwork may be good, but it could be:
too small;
too large;
too high;
too low;
too close to the neckline;
or visually unbalanced on the shirt.
Mockups can expose those problems quickly.
I like checking the product from a normal shopping distance rather than zooming in.
A customer isn’t going to inspect the artwork at 400%.
They’re going to see the whole shirt.
That’s the view that matters.
Test #10: Price
If shoppers are clicking but not buying, price deserves a look.
Compare the product with similar merchandise on the same platform.
Are you significantly higher?
Lower?
Does the price make sense for the product category?
Could your royalty or markup be pushing the final price beyond what shoppers expect?
I wouldn’t automatically lower the price just because sales are slow.
Cheap isn’t always persuasive.
But if your product is $42 and similar shirts are sitting around $26, you’ve found something worth investigating.
Price is part of the buying decision.
Test #11: The Description
A weak description probably won’t destroy a fantastic product by itself.
But a generic, confusing, or inaccurate one certainly doesn’t help.
Does the description explain the design?
Does it sound like the product?
Does it naturally include useful search language?
Does it help the shopper understand who the shirt is for?
Or does it say:
Elevate your wardrobe with this stylish and unique tee…
and then spend three paragraphs saying nothing?
Descriptions should support the product, not simply fill the box.
Test #12: Trust
New shops have another challenge.
Nobody knows them yet.
There may be few products.
No reviews.
No sales history.
No recognizable brand.
That doesn’t mean a new shop can’t sell.
It means you shouldn’t compare your first week with a store that has been accumulating customers and reviews for eight years.
Building out a coherent storefront can help.
So can having multiple strong designs instead of one isolated product.
The more complete the shop looks, the more confidence a buyer may have that there’s a real brand behind the listing.
Test #13: Timing
Seasonal merch has an expiration date.
At least temporarily.
A Halloween shirt can be excellent and still struggle if you publish it after the main shopping window.
The same applies to Christmas, graduation, Valentine’s Day, Mother’s Day, Father’s Day, back-to-school merchandise, and countless other occasions.
If a seasonal product underperforms, ask when it went live.
Did shoppers have enough time to find it?
Order it?
Receive it?
Sometimes the product isn’t wrong.
It’s late.
Test #14: External Promotion
Marketplace traffic isn’t the only source of visitors.
You can also test whether outside promotion changes anything.
Pinterest.
Instagram.
Facebook.
TikTok.
A blog.
Your own website.
Email, if you eventually build a list.
The goal isn’t to spam the product everywhere.
It’s to create additional opportunities for the right people to discover it.
For Flowshirt, I can also use Public Nuisance Apparel as part of the content itself.
We research a product.
Design it.
Publish it.
Share it.
Then document what happens.
That means the blog and store can help feed each other without turning every article into an advertisement.
Change One Thing at a Time
This may be the most important testing rule.
Suppose a product isn’t selling, so you:
change the title;
replace all the tags;
lower the price;
change the artwork;
switch the product;
rewrite the description;
and replace the mockup.
Then traffic improves.
Great.
What worked?
You have no idea.
When possible, change one meaningful thing and give it enough time to produce useful data.
That way, you’re actually testing.
Otherwise you’re just shaking the vending machine.
Keep a Simple Testing Log
This doesn’t need to become a giant spreadsheet.
For each product, I’d keep track of things such as:
date published;
views or impressions;
clicks if available;
favorites;
sales;
main keywords;
price;
major changes;
and the date of each change.
That gives you something to compare over time.
After enough products, you may start noticing patterns that would otherwise be easy to miss.
Maybe certain designs consistently get more clicks.
Maybe a keyword family performs well.
Maybe one product type gets attention and another doesn’t.
Maybe most of your traffic comes from social media rather than marketplace search.
That information can shape what you create next.
Don't Test a Product That Hasn't Had a Chance
This is the part I’m reminding myself of right now.
A product with seven views hasn’t necessarily failed.
Seven people isn’t a market test.
If some of those views are yours, it’s even less meaningful.
That doesn’t mean sitting around forever waiting for magic.
It means recognizing when you simply don’t have enough data yet.
At that stage, I’d rather keep building the shop and creating more entry points than panic and redesign the first product.
Once the listing has received meaningful exposure, then its behavior becomes much more interesting.
Use the Funnel to Decide What to Test
I think the simplest way to diagnose merch is to think of it as a funnel:
No impressions or views?
Look at demand, keywords, competition, indexing, and visibility.
Impressions but no clicks?
Look at the title, thumbnail, design concept, and how it compares with surrounding products.
Clicks but no sales?
Look at the design, product choice, price, listing quality, trust, and purchase experience.
Some sales?
Now look for patterns and ask what you can repeat.
That framework keeps me from treating every disappointing result as a design failure.
Don't Let One Product Decide the Whole Business
One shirt can flop.
Another can surprise you.
One niche may be dead.
Another may bring in traffic.
One visual style may get ignored.
Another may immediately earn clicks.
That’s why testing multiple researched products is more useful than betting everything on one design.
The point isn’t to prove that every product works.
It’s to learn which ideas deserve more of your time.
When Nothing Works
Eventually, there may be a point where you’ve tested visibility, keywords, presentation, price, timing, and multiple products and still see very little response.
That’s useful information too.
Maybe the niche isn’t strong enough.
Maybe the marketplace isn’t a good fit.
Maybe the designs need a different direction.
Maybe the shop needs outside traffic.
Testing isn’t about forcing every idea to succeed.
Sometimes testing tells you to stop.
That can save you a lot more time than endlessly trying to rescue something the market clearly doesn’t want.
The Goal Is Diagnosis, Not Panic
When merch isn’t selling, resist the urge to rebuild everything.
Figure out what stage is failing first.
Then test the thing most likely to affect that stage.
Visibility problem?
Work on discoverability.
Click problem?
Look at presentation.
Conversion problem?
Look at the product, design, price, and buying experience.
No demand?
Move on.
Every product doesn’t need to become a bestseller.
But every product can give you information.
And if you use that information well, the next one should be a smarter test than the last.
Research it. Design it. List it. Test it. Move your merch.
