How Icedout turned virtual try-on into 16,764 лв of attributed revenue

"Does AI try-on actually sell?" is the first question every merchant asks before installing a try-on app on their store — not "is it cool," because cool doesn't show up on a P&L. The real question is whether shoppers who try a product on actually buy it, at a scale that justifies the subscription.

Icedout, a Bulgarian men's jewelry brand at icedout.bg selling necklaces, bracelets, and sunglasses across 274 active products, has been running Tryvio long enough to answer that with real numbers instead of a demo reel. Here's what its live Tryvio analytics show, verified as of mid-July 2026.

The setup: one click, no theme code

Getting Tryvio live on Icedout's store didn't involve a developer or a theme rebuild. Installing a try-on app on Shopify works like installing any other app: approve the permissions it needs — mainly reading products and orders, so it can show the right image and later attribute a sale — then turn the app embed on inside the theme editor. That toggle is the entire integration. No template edits, no custom code, no agency invoice. The "Try it on" button shows up on product pages the same way any theme app embed does: automatically.

That matters for the numbers below — this is what a normal Shopify jewelry store gets from a normal install, not from a custom build.

A real, unedited Tryvio generation — the shopper photo stays untouched; only the jewelry is added, at true scale.

The funnel, in real numbers

Here's Icedout's all-time Tryvio funnel, pulled from its live analytics dashboard:

Read the add-to-cart row again: close to a quarter of every shopper who started a try-on session added a product to their cart. That's the number that turns "AI try-on is a nice feature" into "AI try-on is a sales channel." Of those carts, 293 became real, attributed orders worth 16,764.95 лв total — not a projection or an average order value multiplied by a click count, but 293 individual orders Tryvio can trace back to a specific try-on.

It isn't slowing down, either. Over just the last 30 days, Icedout's widget logged 8,518 clicks, 5,424 try-on sessions, and 4,038 generated looks — close to its entire all-time session count inside a single recent month, with no relaunch or paid push behind it. This is the steady-state pace of a widget shoppers keep coming back to.

Why it works: shoppers see the product on themselves

A chain photographed on a mannequin or a model doesn't tell a buyer what it will look like on their neck, at their skin tone, at their proportions. Product photography answers "what does this look like" — it can't answer "what will this look like on me," and that second question is the one that actually closes a jewelry sale.

Tryvio closes that gap with identity-preserving AI. The shopper uploads their own photo, and the result keeps them recognizably themselves — same face, same proportions, same skin tone rendered convincingly rather than lightened, darkened, or smoothed into something generic. Onto that photo, Tryvio places the actual product from Icedout's own catalog image — not an approximation of it — at the right scale and position. Real product, real person, rendered convincingly no matter who's wearing it: that combination is what a static product photo can never deliver, and it's a large part of why the add-to-cart rate above looks the way it does.

The email-capture bonus

Try-on sessions produce a second asset most merchants don't expect going in: emails. Icedout gives shoppers a number of free try-ons, then offers bonus try-ons in exchange for an email address — a gate shoppers accept because they're already engaged and mid-decision, not cold. That flow alone captured 105 emails, and 76% of those shoppers opted into marketing consent — meaning most of that list is usable for Icedout's own campaigns, not just sitting in a database.

Honest methodology

None of the numbers above are estimates, and none of them are generous. Tryvio counts a purchase as attributed only when the shopper tried on the exact product they bought — either in the same session, or within a 30-day window afterward. Buy a different product in the same order, even one sitting in the same cart as something you tried on? That line item isn't counted. Try a necklace on, come back six weeks later, and buy it? That's outside the window, so it doesn't count either.

That's a deliberately conservative standard. It means Tryvio's attributed revenue understates Icedout's real influence — a shopper who tries a piece on, thinks it over for two months, and buys it after the window closes never shows up in that 16,764.95 лв. What it does mean is that every number in this post traces back to a real order, not a modeled estimate.

Try it on your own store

Icedout's entire setup was a Shopify App Store install and a theme toggle — nothing custom. Install Tryvio to see the same widget on your own product pages, or read the install guide first if you want the full walkthrough.