When a design does not sell, most people assume the design was bad. Usually it was not. It was invisible, and those are completely different problems with completely different fixes.
The useful thing about running a lot of listings is that you can separate them. Three numbers do it: how many people saw the listing, how many showed interest, and how many bought. Each combination points at one cause.
Nobody saw it
This is the common case by a wide margin, and it is worth being blunt about how common. In our own catalogue, the median listing has a small handful of lifetime views. Not a small handful per week. Lifetime.
A marketplace is a search engine with a checkout attached. If your listing does not match what people type, it does not appear, and if it does not appear, nothing else about it matters. The design could be excellent. Nobody is looking at it.
The mistake that causes this is naming a design instead of describing it. "Midnight Serenade" is a nice title and nobody on earth searches for it. A shopper types something closer to "funny cat shirt for coffee lovers", and your title has to contain the words they used or it will not be matched.
The fix is keywords, not artwork. Rewriting the title and tags of an invisible listing is the highest-leverage thing you can do, and redrawing it is close to the lowest.
One qualifier: a brand-new listing has no views because it is new, not because it is bad. Marketplaces take time to index and test a listing. Judging one after a week tells you nothing.
They saw it and did not click
If a listing gets impressions but few clicks, the problem is almost always the lead image, because in a search grid the lead image is the entire advert. Everything else, the title, the price, the description, is behind a click you did not get.
Common causes: the design is too detailed to read at thumbnail size, the shirt colour does not separate from the marketplace's white background, or the image is a flat artwork square instead of a photo of a product. That last one matters more than it should. Shoppers are buying a shirt, and a picture of a shirt outperforms a picture of a drawing almost every time.
The test is free: shrink your lead image to about 200 pixels wide and look at it. If you cannot tell what it says, neither can anyone scrolling.
They clicked and did not buy
This is the rarest of the three and the most encouraging, because it means the concept works. Something specific is stopping the purchase.
Usually it is price against the immediate neighbours, or a listing that promised something the page did not deliver. If your title says "sweatshirt" and the product is a t-shirt, you get clicks and no sales and a slowly worsening quality score. We audited our own catalogue for exactly this and found around seventy live titles advertising a garment we do not stock, because the copywriting had reached for a keyword rather than a fact. That is a self-inflicted wound and it converts terribly.
It can also be trust: no reviews, thin description, or shipping estimates that sit badly next to a competitor's.
The trap: reading a number too early
All three diagnoses need enough data to mean anything. Ten views and no sales is not a signal, it is noise. A listing needs a couple of weeks and a real number of impressions before its numbers say anything, and seasonal designs are worse still: a Valentine's design read in September looks like a failure and is simply out of season.
The instinct to change everything at once is also worth resisting. If you rewrite the title, swap the image and drop the price together, and it improves, you have learned nothing about which change did it, and you have permanently given up margin that may have been fine.
The order to work in
- No views: rewrite the title and tags in the words shoppers actually type. Do not touch the artwork.
- Views, no clicks: change the lead image. Test it at thumbnail size first.
- Clicks, no sales: check the price against the listings you actually appear beside, and check the listing is not promising something the page does not deliver.
- All three healthy: nothing is wrong. Make more like it.
The reason this is worth doing in order is that each fix is cheap and specific, while "the design failed" is expensive and vague. Most designs that get quietly abandoned were never actually tested.
We run this loop on the listings we manage: measure, work out which of the three it is, change one thing, and read it again a fortnight later. If you would rather your designs went through that without you running it, create a free account and we will handle the listing side.