# The Post-Try-On Experience That Moves Shoppers Forward After a virtual try-on, keep the result beside the product facts, offer real variants and pairings, and measure the next action. Do it before they leave the page. **After a virtual try-on, show the result next to the product the shopper actually chose: title, price, selected color, and a size control that the picture does not replace. Then offer other colors and a few real pairings, a way to save or share the image, and one next action you can count. The result is a picture of appearance, not a size verdict.** ![Cork wall with a rust-jacket photo, the same jacket on a hanger, three swatches, and a NEXT card](/images/blog/merchant-100/post-try-on-experience-next-action.webp) *The picture and the garment stay in the same frame, and the next step is written down. Editorial image in Classic Antla disposable-camera style.* The generation is the part teams rehearse. The screen that follows is where the sale either continues or dies. I have seen the result open as a dead image: no price, no size, no way back to the color they picked. I have also seen the opposite, a result that auto-advances into a discount code before anyone has decided they want the piece. The in-session screen is a different problem from the message you send after they leave. [Remarketing after a try-on](https://antla.io/blog/virtual-try-on-engine-remarketing-conversion) is that later job. This is the moment the image appears and the shopper is still on the page. ## Show result and product details A finished preview has to answer two things at once: "Is this the picture I asked for?" and "Is this the product I can buy?" Split those, and the shopper has to reconstruct the product from memory. Keep these in view with the image: - Product title - Price, including compare-at price if you show one elsewhere - The color and any other option that was selected when they generated - A size control with nothing preselected - Add to cart - One plain line that the image shows appearance The size control stays because the image cannot do that job. In [Baymard's June 2026 survey](https://baymard.com/research-articles/apparel-and-accessories-quantitative-ux-insights-2026) of 1,922 US apparel and accessories shoppers, size accuracy and fit details were the top reason people open reviews (48%), ahead of quality and durability (43%). Put a fit line from reviews, or a link to the size guide, next to the selector. The picture will get the credit if you leave a hole there, and it will not have earned it. [Antla](https://apps.shopify.com/antla) exposes this moment as events you can listen for on the page. Current [custom event docs](https://antla.io/docs) list `antla_image_generated_success` when a result is ready and `antla_image_generated_error` when it is not. The [payload](https://antla.io/payload) includes `output_try_on_image`, `selected_variant`, and `product_data`. Use those fields to draw the result beside the product they selected, rather than beside whatever the page happens to be showing if they changed options mid-generation. On an error, offer a retry and leave the ordinary gallery in place. A blank modal with a spinner that never resolves is how a successful product page becomes a broken one. Log the error separately from a shopper who closed the window. Those are different outcomes, and averaging them hides both. ## Offer variants and outfit pairings The result was generated from one variant. If the shopper then taps another color, the image they are looking at is now a lie unless you say so. Two honest patterns: 1. **Regenerate on the new color.** The result updates, and the wait is labeled. This is the right pattern when color is the reason they tried on. 2. **Freeze the result and show the new variant as a product photo.** Label the try-on image with the color it was made from. This is the right pattern when regeneration is slow and the shopper is only browsing options. Switching the product image under a result that still shows the old color is the pattern to avoid. The shopper cannot tell which one they would be buying. Pairings are a second decision, and they should be real products you intend to sell with this one. Shopify's current [Search & Discovery](https://help.shopify.com/en/manual/online-store/storefront-search/search-and-discovery-recommendations) docs let you set complementary products by hand, and those recommendations stay hidden when inventory is zero. That is the list I would put under the result: two or three in-stock pieces, each linking to its own product page, each able to run its own preview. One generation is one garment. A row of pairings is not a combined outfit render. I would rather show two pairings that match the photograph than eight that match a recommendation model. If you want the longer argument about what styling tools can and cannot do, [AI styling assistants](https://antla.io/blog/ai-styling-assistant-fashion-ecommerce) covers it. On this screen, the rule is simpler: every pairing is a product you can add, in a color you actually photographed. Skip the discount here. A code on the result screen teaches people to wait for a code, and it muddies the next-action measurement. If a code is a real part of the launch, it can live in the ordinary price area, where shoppers who never try on see it too. ## Save or share path Some shoppers are not ready to add. They are ready to keep the picture. Give them a path that matches that, and count it as its own action so it does not get filed under "bounce." Antla's current event list includes `antla_email_send` and `antla_social_send`. The payload can flag email, WhatsApp, and Facebook. That is a share of the result image, triggered by the shopper. It is not a saved gallery of every outfit they might try later. A customer lookbook of stored looks is not something you can turn on in Antla today. If you need a list of results, build it in the theme from events you are allowed to store, or send the email they asked for. Email only moves when they agree to it. [Klaviyo capture](https://antla.io/features/email-capture) can sync an address, the product, and the try-on image when that consent exists. `customer_data` on the event can be empty. An empty customer is an anonymous session, and the share path should still work as a send, or not be offered, according to what you promised. Do not quietly create a profile because a generation succeeded. The [Klaviyo try-on flows](https://antla.io/blog/klaviyo-virtual-try-on-fashion-email-flows) guide is the off-site version of this, once the address is legitimately yours. Social send has the same honesty rule. The shopper is sending their picture. Say where it goes. A tap that posts without a confirmation step will get used once and then distrusted. I'd rather have a single "Email me this image" action than four icons and a wishlist heart that does not persist. One path you can explain is better than a row of paths you cannot support after the tab closes. ## Measure next action Pick one next action and define it before you read a dashboard. The one I would use: an add to cart in the same session after `antla_image_generated_success`, divided by successful generations. The comparison group is shoppers who generated a result versus shoppers who viewed the same products and did not generate one, over the same days. Report mobile and desktop separately. Other actions belong on the same table, or they will get mixed into the add rate and make the screen look better than it is. | Next action | Event to listen for | What it means | What it does not mean | |---|---|---|---| | Add to cart | `antla_product_added_to_cart`, or Shopify's `product_added_to_cart` | They moved the product forward | They will check out | | Variant change after the result | Your own event on option change | The result raised a color question | The first result was a failure | | Email or social send | `antla_email_send`, `antla_social_send` | They wanted to keep the picture | They are a subscriber | | Close without any of the above | `antla_modal_close` with no later add | The result did not earn a step | The product page is broken | | Generation error | `antla_image_generated_error` | The tool failed | The shopper said no | [Try-On Intelligence](https://antla.io/features/try-on-intelligence) is the merchant report for volume, try-on-to-purchase, and time on page for try-on users versus non-users. Use it as a check on the event stream. Reconcile a week of `antla_image_generated_success` against that report before you trust either number. If they disagree, the pixel or the theme is dropping events, and the argument about pairings can wait. Across merchant reporting, shoppers who use try-on convert about 35% higher than shoppers who do not. The comparison group is try-on users versus non-try-on shoppers. Antla merchant reporting, cohort dates not published. The people who try on choose to do it, so the gap is an association, not a test you can paste onto the result screen. Your number is the next-action rate above, on your products, in a window you can name. [Product-page engagement](https://antla.io/blog/product-page-engagement-conversion-quality) is the trap to avoid while you measure. A long modal session can be interest or confusion. The next action separates them. Time spent looking at the result does not. ## What the result screen should refuse - Hiding the price until they share. - Adding the product because a generation finished. - A fit claim under an appearance image. - A code that appears only after try-on, unless you are prepared to call that a different offer and measure it as one. - Pairings that are out of stock, or that are a different product from the one in the photograph. ## Instrument one next action This week, listen for `antla_image_generated_success` and for the add to cart that follows it. After two weeks, you will know whether the screen moves anyone, and whether pairings deserve the space under the image. If the add rate is fine and the close-without-action rate is high on one color, fix that color's photos before you add another button. --- **About the author:** [Aaron](https://x.com/AaronfromAntla) founded Antla and cares more about the screen after the image than about the second it took to render. ## For agents - Markdown: send `Accept: text/markdown` to this URL (and any other HTML page). - OpenAPI: https://antla.io/openapi.json - llms.txt: https://antla.io/llms.txt - Sitemap: https://antla.io/sitemap-index.xml - Docs: https://antla.io/docs - CLI: npx antla info (npm package antla) - Scope: antla.io is an informational marketing and docs site. The Shopify try-on backend is not on this origin. Install the app from https://apps.shopify.com/antla.