I started this Shopify App Store optimization pass to make two listings easier to find. It ended by making a product decision.
PopSmart Studio is now PopSmart AI Virtual Try On. PopSmart Clerk is no longer public while I pivot it.
That is a cleaner result than keeping two vague products live and calling both of them works in progress.
The first rename was only a patch
In June, I renamed both apps around the jobs merchants search for. Studio became “VirtualTryOn.” Clerk became “AIUpsell.” I was trying to spend the app-name field on discovery instead of a brand nobody knew.
The logic was right. The execution was incomplete.
“VirtualTryOn” removed the spaces from the actual search phrase. More importantly, the rest of the Studio listing still described a broad AI image toolbox. The name pointed toward one job while the screenshots and product story kept pointing in several directions.
ASO exposed that mismatch. A keyword cannot rescue a product page that has not decided what the product is.
Studio now leads with one job
The current name is PopSmart AI Virtual Try On. The spaces are back. The product now leads with the thing an apparel merchant can understand immediately: put a virtual fitting room on a Shopify product page so shoppers can see the look before they buy.
The listing, product page, and demo store now tell the same story.
Virtual try-on comes first. Fit confidence and fewer wrong-fit returns are the merchant problem. The Theme App Extension on the product page is the main product surface.
The AI model photos, product scenes, background tools, templates, and social publishing are still there. They now sit behind the primary job instead of competing with it. They make the product broader after the first sentence is already clear.
I also replaced the listing images around that order: shopper try-on first, then the fitting-room block, AI model photography, product-photo tools, templates, and publishing. The sequence now explains one workflow instead of presenting six unrelated features.
The full current product file is in the PopSmart AI Virtual Try On archive entry.
Clerk went private
Clerk did not earn the same conclusion.
The first version tried to be an AI store clerk that watched shopping behavior and produced contextual product messages. I tested names, listing copy, screenshots, pricing, ads, and a demo angle. The central problem stayed harder to state than it should have been.
So I took the Shopify App Store listing private.
This is not a tombstone yet. Clerk is paused while I pivot it toward a more concrete merchant job. But I am not going to keep asking new merchants to install the old version while the product itself is changing.
Private is the honest state between defending the current idea and pretending the next one is already ready.
What ASO actually did
I used to think of ASO as distribution work: choose the search term, tune the listing, improve discovery, measure installs.
It is also a product test.
If the keyword, title, first screenshot, category, and opening sentence cannot all describe the same job, the problem may not be optimization. The product may still be avoiding a decision.
Studio passed that test by becoming specific: Shopify virtual try-on first, visual production second.
Clerk did not. So Studio gets a sharper public listing, and Clerk gets privacy and time to change.
I do not have an ASO victory graph yet. The next evidence is still listing impressions, visits, installs, and actual merchant use. For now, the useful outcome is simpler: one product is easier to understand, and one product is no longer public before it is ready.