<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Tuck Virtual Try On on Medium]]></title>
        <description><![CDATA[Stories by Tuck Virtual Try On on Medium]]></description>
        <link>https://medium.com/@tuck.now?source=rss-9d03e79fdb1a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*faCAbxWn3wavAfmKt3BxeQ.png</url>
            <title>Stories by Tuck Virtual Try On on Medium</title>
            <link>https://medium.com/@tuck.now?source=rss-9d03e79fdb1a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 22:52:58 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@tuck.now/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Why Most Virtual Try-On Demos Look Better Than the Real Thing]]></title>
            <link>https://medium.com/@tuck.now/why-most-virtual-try-on-demos-look-better-than-the-real-thing-46f13ca7ce7a?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/46f13ca7ce7a</guid>
            <category><![CDATA[shopify]]></category>
            <category><![CDATA[ecommerce]]></category>
            <category><![CDATA[virtual-try-on]]></category>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Sat, 10 Oct 2026 06:57:30 GMT</pubDate>
            <atom:updated>2026-10-10T06:57:30.420Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ydb8UajWKwElalSP.png" /></figure><p>Imagine a virtual try-on demo that looks incredible.</p><p>A clean product shot, a well-lit photo of a person, and a result that held up to scrutiny — the garment sat correctly, the fabric read convincingly, nothing about it felt generated. You could see exactly why this technology is worth having on a product page.</p><p>Then imagine the same tool running across a live catalogue. The results may be less consistent than the selected example.</p><p>This gap can be disappointing, but it does not by itself mean a product is bad or a vendor is dishonest. Software demonstrations often use favourable examples, and choosing an example that works is reasonable.</p><p>The problem is what a selected example can and can’t tell you.</p><p>A great demo proves a result is possible. It doesn’t prove the result is repeatable.</p><p>Those are different claims, and the gap between them is where important deployment challenges arise.</p><h3>The problem with evaluating one perfect image</h3><p>Consider what you’re actually being shown. One garment, chosen by someone who knows which garments work. One shopper photo, usually well-lit and front-facing. One pose. One set of favourable conditions.</p><p>That’s a single point. What you need to know about is a distribution — what happens across every garment in your catalogue, photographed however it happened to be photographed, worn by every shopper who arrives, in whatever picture they happen to upload.</p><p>One impressive result from a plain black shirt is genuine evidence of capability. It doesn’t tell you what happens with 300 patterned dresses, 50 embroidered garments, shoppers standing with their arms crossed, and product photography shot by four different people over three years.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*J8H5gndxc1nswQva.png" /></figure><p><em>Illustration of how catalogue and shopper conditions vary; not measured product results.</em></p><p>Researchers working on this problem have reached a similar conclusion about evaluation. <a href="https://arxiv.org/abs/2505.19571">VTBench (Hu et al., 2025)</a> proposes a benchmark for real-world virtual try-on covering image quality, texture preservation, complex backgrounds, cross-category size adaptability and hand occlusion. The authors note that conventional metrics and indoor test sets do not adequately capture all real-world scenarios.</p><p>That’s the academic version of the same observation. Testing under favourable conditions tells you about favourable conditions.</p><p>So it’s worth understanding why real conditions are harder. There are four reasons, and they compound.</p><h3>Why the product photo changes the result</h3><p>Image-based try-on systems start from a garment image, and catalogues are not uniform.</p><p>Across a real store you’ll find garments shot flat, garments on mannequins, garments on models, garments at an angle, garments partly folded, garments with another item layered over them. Backgrounds vary from seamless white to a studio floor to something cluttered. Resolution varies. New and older products may have been photographed under different conditions.</p><p>A try-on result can only be as informative as the garment information available to produce it. A crisp, clearly presented product image gives a system a great deal to work with. A low-resolution photo of a folded garment with a sleeve hidden behind a fold gives it considerably less.</p><p>Product photography is one practical source of the demo-to-reality gap. A selected demo may use a strong source image; a real catalogue may contain less suitable ones.</p><h3>Why garment complexity matters</h3><p>A plain t-shirt and a heavily embroidered outfit are not the same problem.</p><p>Think about what has to survive the process intact: patterns that need to stay aligned, logos that need to remain legible, printed text, lace, embroidery, layered pieces where one garment sits over another, flowing fabric, structured sleeves, a saree drape, any silhouette distinctive enough that a shopper would notice it being slightly wrong.</p><p>This is an active area of research. Tran et al. (2025), presenting <a href="https://openaccess.thecvf.com/content/ICCV2025W/RetailVision/html/Tran_DualFit_A_Two-Stage_Virtual_Try-On_via_Warping_and_Synthesis_ICCVW_2025_paper.html">DualFit at the ICCV Workshops</a>, discuss how generative virtual try-on may lose fine garment details such as logos and printed text. Mobinizadeh and Lakizadeh (2025), writing in <a href="https://www.nature.com/articles/s41598-025-18107-6">Scientific Reports</a>, address challenges involving garment alignment, occlusion, textures, patterns and complex poses.</p><p>Worth being precise about what those papers mean. Each proposes methods to address these challenges — they’re research into solutions, not evidence that every tool fails. What they establish is that these specific things are hard, which matters when you’re deciding what to test.</p><p>Because the useful question about a garment isn’t does the output look good. It’s does the output still look like this product. This distinction matters when distinctive details are part of why a shopper chooses a garment.</p><h3>Why real shopper photos are unpredictable</h3><p>Shoppers may upload ordinary phone photos rather than studio photographs.</p><p>That means arbitrary poses, odd crops, variable resolution, bad lighting, busy backgrounds, other clothing already being worn, arms folded, a hand in front of the body, a photo taken at arm’s length in a hallway at eleven at night.</p><p>These conditions can make try-on harder, and a curated demo may not test them.</p><p>Which turns this into a product design question rather than purely a technical one. A system that assumes good inputs can struggle with less suitable shopper photos. A system that guides people toward usable photos — telling them what works before they choose, letting them retry easily, saying clearly when an image won’t produce a good result — converts a technical limitation into a manageable experience.</p><h3>Realism is not the same as fidelity</h3><p>This distinction deserves more attention than it usually gets, because visual plausibility can hide changes to the garment.</p><p>An image can be entirely photorealistic — correct lighting, plausible anatomy, coherent background, nothing that reads as generated — while subtly changing the actual product. A print shifts. Embroidery softens. A colour moves half a shade. A neckline sits slightly differently. A sleeve ends somewhere it doesn’t end in reality.</p><p>These are easy to miss when you’re admiring an image and more apparent when you receive the garment.</p><p>So there are two separate questions: does the image look real, and does it preserve the product?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*hNo6V2QFAVXZRlqi.png" /></figure><p><em>Illustration of the distinction between a believable image and accurate garment details.</em></p><p>When you evaluate, put the generated image next to the actual product photo and compare them deliberately. Not is this convincing, but is this the same garment.</p><h3>Why a production system needs guardrails</h3><p>Everything above explains why a single generation is hard. Running try-on across a live store is a different undertaking again.</p><p>Production involves deciding which products are eligible, choosing which source image to use when there are several, guiding the shopper’s input, generating the result, communicating limitations where they exist, handling cases that aren’t supported, allowing retries, and measuring what happened.</p><p>These capabilities vary considerably between providers, and most of them are invisible in a demo because a demo skips straight to generation.</p><p>That’s the real difference between a system that demos well and one that deploys well. Not the quality of the best output — the handling of everything around it.</p><h3>When not showing a try-on is the better decision</h3><p>Here’s the idea that tends to surprise merchants.</p><p>More coverage is not automatically better. Enabling try-on indiscriminately can expose shoppers to results from unsuitable products. A misleading result may weaken trust in the feature, including on products where it would be useful.</p><p>Withholding try-on for an unsuitable product can be more valuable than displaying a misleading result.</p><p>But this is a trade-off rather than a rule with an obvious answer. Restrict coverage too aggressively and the feature becomes rare enough that shoppers never encounter it, which is its own kind of failure. The right balance depends on your catalogue, and it’s a question worth asking any provider directly: how do you decide, and can I override it?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*XeOlbxv2O-4TyH9V.png" /></figure><p><em>Illustrative quality-control considerations for a production workflow; capabilities vary by provider.</em></p><h3>What consistency actually means</h3><p>If one great image isn’t the measure, what is?</p><p>Think in terms of representative cases rather than best cases. The things worth knowing about a deployment:</p><p>How much of your catalogue is eligible for try-on at all. What share of attempted generations produce a result you’d be comfortable showing. How often shoppers retry or hit a failure. Whether garment details survive on the products where details matter. What proportion of shoppers who start a try-on finish one. How long generation takes, and how that varies. How much support burden the feature creates.</p><p>Two of those deserve reading together rather than separately. High coverage looks good until you notice it was achieved by exposing unreliable products to shoppers. Low failure rates look good until you notice they were achieved by excluding most of the catalogue. Neither number means much alone.</p><p>None of this requires sophisticated measurement. It requires testing a representative spread of your own products rather than the one the vendor picked.</p><h3>Why we think the system around the render matters</h3><p>Tuck is an apparel-focused virtual try-on and size-guidance product, and the thinking above shapes how we’ve built it.</p><p>A few positions we hold. A try-on result should represent the real product usefully, not just look appealing. Not every product image is equally suitable as a starting point. It can be better to withhold try-on for a product than to show a shopper something misleading. A good shopper experience has to deal with imperfect inputs rather than assume they won’t arrive.</p><p>At Tuck, we focus on catalogues and choose products where virtual try-on is available, rather than assume it should be enabled for every item. Size suggestion is a separate capability, using shopper information against the merchant’s own size chart — the try-on image shows how something might look, and it doesn’t determine anyone’s size.</p><p>We’d be overstating things if we claimed immunity to any of the technical challenges described here. The difference worth discussing is what a system does about them.</p><p>What we’d encourage, whoever you’re evaluating: test across your real catalogue conditions rather than a selected example, and ask what happens when conditions aren’t ideal.</p><h3>The real test</h3><p>A demo can win attention. A dependable experience has to earn trust repeatedly, across every shopper who tries it and every product they try it on.</p><p>Those are different achievements. Evaluating both requires looking beyond a selected sales demonstration.</p><p>The real test of virtual try-on isn’t its best image. It’s what happens after the demo ends.</p><h3>See Tuck with your catalogue</h3><p>Want to discuss virtual try-on for your store and the products where it is available? Book a demo with the Tuck team.</p><p><a href="https://www.tucknow.com/book-demo">Book a Tuck demo</a></p><p><em>Originally published at </em><a href="https://www.tucknow.com/blog/virtual-try-on-demo-vs-reality"><em>https://www.tucknow.com</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=46f13ca7ce7a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What to Actually Look For When Choosing a Virtual Try-On App]]></title>
            <link>https://medium.com/@tuck.now/what-to-actually-look-for-when-choosing-a-virtual-try-on-app-374617d946c1?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/374617d946c1</guid>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Thu, 08 Oct 2026 12:58:49 GMT</pubDate>
            <atom:updated>2026-10-08T12:58:49.062Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*g60HF3Aw9U52IqEO6QouNw.png" /></figure><p>You’re evaluating try-on apps. You upload a clean product image and a clean photo of yourself. The result comes back and it’s good — the garment looks right, the proportions hold, and you can see what it would be like to own it.</p><p>That’s a useful signal: virtual try-on can help shoppers visualize a garment on their own photo. Whether it is worth deploying still depends on your products, customers and results.</p><p>But that demo answered exactly one question: <em>can this produce a good image?</em></p><p>Here are some it didn’t answer. What happens when a shopper uploads a dim photo taken at arm’s length in a hallway? What happens on a phone? What happens when she likes the result and then asks which size to order? What happens to her photo afterwards? And what happens to everything the interaction just told you about her interest?</p><p>The render is the most visible part of a virtual try-on system. It is not the only part to assess.</p><h3>Quality is the entry requirement, not the differentiator</h3><p>Image quality matters enormously. If shoppers don’t find the visualisation useful or believable, nothing else about the product is worth discussing.</p><p>But it’s a threshold. Once several options can produce a convincing demo, your decision turns on what happens before the render, around it, after it, and at volume.</p><p>Seven things to evaluate.</p><h3>1. Output quality — across more than one garment</h3><p>Ask better questions than <em>does it look realistic?</em></p><p>Does the garment still look like your actual product? Patterns are where this shows first — a stripe that drifts or a print that smears tells a shopper the image isn’t trustworthy. Are colors faithful? Do logos and details survive? Does the silhouette stay credible?</p><p>Then check the hard things: sleeves, drape, layers, dark fabrics where detail disappears, complex dresses and unusual poses.</p><p><strong>The practical test:</strong> pick 10 to 20 products spanning a simple t-shirt, a structured shirt, trousers, a dress, something heavily patterned, something very dark, something very light, a layered look, a complex silhouette, and a few shot in different photography styles. This is a rough working number rather than a scientific sample — the point is breadth.</p><p>Testing one ideal SKU tells you about that SKU.</p><h3>2. Shopper experience</h3><p>A system can be technically excellent and still be difficult to use. Your customer experiences the interface, not the technology underneath it.</p><p>Count the taps. From “try it on” to seeing herself takes some number of steps, and each one is a place she might leave. Does she upload her own photo, and does the app tell her what kind works before she picks an unsuitable one? Does the whole experience work on the devices your customers actually use? If she wants to see a second garment, does she start over?</p><p>Then check what happens once the image is on screen.</p><p>This can be a moment of meaningful product engagement. Check whether the experience offers a useful next step. Can she get size guidance, reach the cart, or try another colour?</p><p>A try-on that ends in a window she closes may leave the next part of her decision unresolved.</p><h3>3. Can it handle your catalogue?</h3><p>Demos may begin with one carefully chosen garment. Your store doesn’t.</p><p>Your catalogue has 50 products, or 500, or 5,000, photographed over years by different people with different equipment. Some products have eight images; some have two.</p><p>Does every product automatically get try-on, and should it? Can you control which ones do? Can unsuitable products be excluded? How is the product image selected? What happens when you add 200 products next season?</p><p>The setup required can affect how widely your team deploys the app. Ask what is needed as your catalogue grows.</p><p>With Tuck, check which products have VTON available and use the catalogue controls to manage where it appears. Build featured catalogues around those eligible products. Test representative garments rather than assuming eligibility guarantees output quality.</p><h3>4. Look and size are different questions</h3><p>A try-on image answers roughly <em>how might this look on me?</em></p><p>It does not answer <em>which size should I order?</em> Separate questions, separate inputs — one needs an image of the shopper; the other needs information about her against your size chart.</p><p>A provider might offer visualisation only, visualisation plus a static size chart, visualisation plus a size recommendation, or sizing as its own system.</p><p><strong>The question worth asking:</strong> when she finishes the try-on and wonders about size, what happens next? If the answer is “she opens the size guide and works it out herself,” the sizing question remains unresolved.</p><p>Tuck keeps these deliberately separate. Try-on visualises the garment on the shopper. Size suggestion uses height, weight, age and fit preference against your own size chart.</p><p>The image doesn’t determine the size, and our terms describe sizing outputs as indicative estimates — they help with guesswork rather than removing it. Treat a guaranteed-fit claim cautiously and ask for supporting validation.</p><h3>5. What does it do with the interaction?</h3><p>Think about what a try-on represents commercially. A shopper has chosen one garment out of your catalogue, taken an action requiring effort, and waited for a result. That can be a useful signal of product interest; it does not establish purchase intent.</p><p>Now ask what the app does with it.</p><p>Without measurement, it is harder to assess the commercial value of that interaction.</p><p>The useful questions are about connection:</p><p><strong>Analytics:</strong> which products are tried, and can the data be connected to subsequent actions? Verify the metrics available rather than assuming abandonment or conversion tracking is included.</p><p><strong>Returns:</strong> test whether visualisation and size guidance affect return reasons and rates. Do not assume a reduction.</p><p><strong>Engagement and marketing:</strong> ask whether consent-based follow-up is supported directly or requires a separate integration.</p><p>These are evaluation criteria, not a promise that every capability is built into Tuck.</p><p>Does the interaction produce analytics you’d actually use? Can it support a useful next step without pushing the shopper through an unnecessary marketing journey?</p><p>Assessing these outcomes requires measurement alongside output quality.</p><h3>6. What happens when it fails?</h3><p>A demo may focus on successful generations. Ask about the other ones.</p><p>Shoppers will upload photos that are too dark, cropped at the shoulder, or taken at an odd angle. Some garments won’t work well. Sometimes generation will be slow or fail outright.</p><p>Does the shopper get useful guidance, or a confusing result? Can she retry easily? Does the system catch unsuitable inputs before generating something poor? Is there a sensible fallback?</p><p>Judge a production system not only on how it succeeds but on how gracefully it fails.</p><p>7. Read the privacy policy, not the listing copy</p><p>Your customer is uploading a photograph of herself to your storefront. That makes photo handling a question to resolve before you install.</p><p>What’s collected, and why? Are the original photos kept, and for how long? Are they used to train the provider’s systems? Which infrastructure providers process them? Can a customer delete her data herself, or does she have to email someone? If she can save a profile to reuse, how does that differ from a one-off session?</p><p>Look for a policy that answers these questions directly.</p><p>At Tuck, photos shoppers upload are processed to produce the try-on. We do not currently use raw photos or personally identifiable biometric data to retrain or fine-tune our proprietary models.</p><p>Users can choose to delete their data using the <strong>Delete My Data</strong> feature. For users who have not used that feature, their data is automatically deleted after <strong>30 days of inactivity</strong>.</p><p>We do not sell personal information. Trusted infrastructure providers act as processors.</p><h3>And then: the economics if people actually use it</h3><p>The subscription price is the easy number. The useful question is what happens as usage grows.</p><p>Check the monthly fee, how many try-ons are included, what extra ones cost, whether usage can be capped, what happens when the allowance runs out, which features sit behind higher tiers, and whether analytics are included or charged separately.</p><p>A simple planning frame:</p><p><strong>Estimated monthly cost = base plan + additional usage + any required add-ons</strong></p><p>Tuck’s public Shopify pricing, checked on 7 October 2026, is in USD and billed every 30 days:</p><ul><li><strong>Starter:</strong> $14.99 for 200 try-ons.</li><li><strong>Growth:</strong> $49.99 for 800 try-ons.</li><li><strong>Scale:</strong> $129 for 3,000 try-ons.</li></ul><p>Extra try-ons cost $0.079, $0.066 and $0.045 respectively. The listing offers up to 50, 200 and 500 extra try-ons on Starter, Growth and Scale. Confirm the available allowance before a high-volume campaign.</p><p>Analytics are included on all three paid plans. Contact the team about enterprise requirements and separate commercial terms.</p><p><a href="https://apps.shopify.com/tuck-hyper-realistic-try-on">See the current Shopify listing.</a></p><p>Divide price by allowance and you get approximately $0.075, $0.0625 and $0.043 per included try-on.</p><p>One caveat applies to any vendor quoting per-unit figures, ours included: those rates hold only if you use the full allowance. Buy 3,000 and use 1,000, and you’ve paid about $0.13 each.</p><h3>A longer feature list isn’t the same as a connected one</h3><p>Apps in this category advertise email capture, marketing automation, catalogue creation, size charts, size recommendations, analytics, outfit building, customer history, sharing and post-try-on offers.</p><p>More features is not automatically better. Features you don’t use can add complexity, and an experience with too many steps can become difficult for shoppers.</p><p>Another risk is a try-on connected to nothing else: a render in a window, with no sizing answer, no path onward, and no way to tell afterwards whether it helped.</p><p>The question isn’t how many features an app has. It’s whether they connect to each other and to the decision your shopper is trying to make.</p><h3>The 15-minute VTO test</h3><p>Use fifteen minutes for an initial assessment:</p><p><strong>Minutes 1–3:</strong> test several garments, not one. Include a pattern, something dark, and something with structure.</p><p><strong>Minutes 4–5:</strong> use a deliberately imperfect shopper photo — poor light, an awkward crop, or an ordinary phone snap.</p><p><strong>Minutes 6–7:</strong> do it all again on a phone.</p><p><strong>Minutes 8–9:</strong> find out what happens with a product that isn’t suitable. Does it warn you, or produce a poor result?</p><p><strong>Minute 10:</strong> check what happens when a shopper wants a size.</p><p><strong>Minute 11:</strong> look at the analytics. Would you use what’s there?</p><p><strong>Minute 12:</strong> read the privacy policy on photo retention and deletion.</p><p><strong>Minute 13:</strong> check how many try-ons are included and what extra ones cost.</p><p><strong>Minute 14:</strong> check which features sit behind which plan.</p><p><strong>Minute 15:</strong> ask yourself: could I confidently put this in front of eligible shoppers on my store?</p><p>If the answer is no, use the trial to investigate the gaps before a wider rollout. Fifteen minutes is an initial screen, not a production validation.</p><h3>Buyer scorecard</h3><p>Score each area from 1 to 5 for every app you’re assessing:</p><ul><li><strong>Visual quality:</strong> does it preserve the product convincingly?</li><li><strong>Consistency:</strong> does quality hold across representative products?</li><li><strong>Shopper experience:</strong> is try-on easy on desktop and mobile?</li><li><strong>Catalogue:</strong> can I control what gets try-on?</li><li><strong>Sizing:</strong> what happens when the shopper asks which size?</li><li><strong>Connected value:</strong> what does the system do with the interaction?</li><li><strong>Analytics:</strong> can I measure usage and outcomes?</li><li><strong>Failure handling:</strong> what happens when generation fails?</li><li><strong>Privacy:</strong> do I understand photo and data handling?</li><li><strong>Economics:</strong> does cost work at expected volume?</li><li><strong>Operations:</strong> can my team realistically manage it?</li></ul><p>The total isn’t the point. Simply adding these scores together would overlook what matters most to your store.</p><p>A 5,000-product catalogue should give catalogue controls and operations considerable weight. A 40-product boutique may prioritise simplicity. A high-traffic store should examine economics closely.</p><p>The scorecard’s value is that it forces you to decide what you care about before a demo decides for you.</p><h3>What not to use as your decision rule</h3><p><strong>The best demo image.</strong> It may be a selected sample rather than a representative result.</p><p><strong>The lowest starting price.</strong> Compare the allowance, extra usage charges and required features at your expected volume.</p><p><strong>The longest feature list.</strong> Check whether the features connect to something useful.</p><p><strong>The highest app store rating.</strong> Reviews can offer useful evidence about support and merchant experience. But a handful of positive reviews is not extensive validation, and ratings alone don’t establish rendering quality.</p><p><strong>The most impressive AI terminology.</strong> Technical vocabulary doesn’t tell you how well the experience works for your shoppers.</p><p><strong>A conversion number with no method attached.</strong> Ask what it was measured against and over what period. A comparison of engaged shoppers against all other traffic can be affected by selection bias, because those groups may differ in purchase intent. That applies to any number you’re shown, including ours.</p><h3>Where Tuck fits</h3><p>Tuck is built around a specific reading of the apparel problem: let shoppers see the garment on themselves, help them choose a size, manage where try-on is available across a real catalogue, and make the interaction produce information merchants can use.</p><p>In practice, that means shopper-photo apparel try-on, size suggestions handled separately, catalogue eligibility controls, analytics included on every paid plan, and lower included per-try-on rates on the higher-volume plans.</p><p>The thread connecting those is the point. The render is one part of the experience. The shopper still needs information to make a decision, and the merchant needs a way to assess whether the feature is useful.</p><p>If that’s how you see the problem, Tuck should be on your evaluation list.</p><h3>When Tuck may not match your requirement</h3><p>If you primarily need merchant-side AI fashion photography or catalogue imagery rather than shopper-facing try-on, evaluate a content-generation product.</p><p>These categories can be confused because both involve generated images, but one serves your production workflow and the other serves your customer’s decision.</p><p>If your category sits outside apparel, verify compatibility before assuming the approach transfers.</p><p>And if you need something we don’t publicly document, ask rather than assume.</p><h3>The demo is the beginning of the evaluation</h3><p>Go back to that first good render.</p><p>It earned the product a second look. It didn’t finish the assessment.</p><p>A decision you can live with answers more than one question. Can it generate useful results? Can shoppers use it easily? Can it handle your catalogue? Can it cope with imperfect inputs? Can it help with the sizing question that follows? Does it produce useful information? Do you understand the data handling? Can you afford it when usage grows?</p><p>Virtual try-on may be useful when testing shows that it improves the shopper experience and its cost is justified.</p><p>The best app is the one you’re comfortable putting in front of real shoppers, across a real catalogue, at realistic usage — and whose results you can assess afterwards.</p><h3>See how Tuck fits your store</h3><p>Discuss catalogue eligibility, shopper experience, size guidance and pricing with the Tuck team.</p><p><a href="https://www.tucknow.com/book-demo">Book a Tuck demo.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=374617d946c1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why “I’ll Think About It” Usually Means “I Couldn’t Picture It”]]></title>
            <link>https://medium.com/@tuck.now/why-ill-think-about-it-usually-means-i-couldn-t-picture-it-1749f79f9e66?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/1749f79f9e66</guid>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Mon, 05 Oct 2026 06:17:58 GMT</pubDate>
            <atom:updated>2026-10-05T06:17:58.881Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lyLoMcOlBZI3CoQUvu8ePA.png" /></figure><p>A shopper turns her phone round to show a friend a dress.</p><p>“What do you think?”</p><p>“It looks great.”</p><p>She looks at it again. “Yeah. I like it.” A pause. “I’ll think about it.”</p><p>Nothing has obviously gone wrong. She likes the colour. The price isn’t impossible. Her size is in stock. She hasn’t found something better on another site.</p><p>And yet she hasn’t decided.</p><p>Before going further, a caveat the headline doesn’t have room for: “I’ll think about it” means a lot of things. Payday is Friday. She’s comparing three options. She wants her sister’s opinion. The shipping is slow. She’s not sure about the colour. Sometimes she’s just being polite about something she doesn’t want.</p><p>This article is about one particular version of it, and it’s a version that fashion ecommerce is unusually prone to producing.</p><p>She likes the dress. She can’t quite picture herself in it.</p><h3>Liking something isn’t the same as wanting it</h3><p>There’s a real distance between those two states, and it’s easy to miss because they feel adjacent.</p><p>You can like a jacket you’d never own. A coat in a magazine, a shirt on someone in a café, a dress on a mannequin. Liking is an aesthetic judgement about an object, and it commits to nothing.</p><p>Wanting something for yourself is different. It requires placing the thing in your life — in your wardrobe, on your body, at events you actually attend, next to the clothes you already own.</p><p>The product page asks the first question very well. <em>Do you like this garment?</em> Good photography is extremely effective at producing yes.</p><p>The purchase decision eventually asks a second one. <em>Do you want this garment for yourself?</em></p><p>Between those two questions sits a piece of work, and the shopper usually has to do it alone.</p><h3>The imagination gap</h3><p>We call this the imagination gap — our own shorthand rather than any established term. It describes something simple:</p><p><strong>The distance between seeing a product and being able to picture yourself wearing it.</strong></p><p>Traditional ecommerce asks the shopper to cross that distance mentally. She sees a model wearing a garment, and has to convert that into an image of herself wearing it.</p><p>Sometimes that’s effortless. If the model is roughly her build and colouring, if the styling matches how she dresses, if the garment is a simple shape she’s worn before — the translation happens instantly and she barely notices doing it.</p><p>Sometimes it isn’t. And when it isn’t, she doesn’t usually diagnose the problem. She doesn’t think <em>I lack sufficient visual information to complete this evaluation.</em> She just feels unsure.</p><p>That feeling leaves the page as “I’ll think about it.”</p><h3>Why fashion makes this harder than most categories</h3><p>A coffee machine can stay an object. You evaluate it on specifications, reviews, how it looks on a counter. It never really has to become personal.</p><p>Clothing can’t stay an object for long, because at some point it has to become <em>a jacket on me</em>.</p><p>Which means the shopper is weighing several things at once, whether or not she’d articulate them: the silhouette on her particular frame, the colour against her skin, the length, the proportions, how she’d style it, what occasion it’s for, whether it reads as bold or understated on her, and the quiet underlying question of whether it feels like her at all.</p><p>No single feature resolves all of that, and claiming otherwise would be overselling. But it explains why visual self-reference matters more in apparel than in almost any other category.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q3AE2y5kk3I9lR6sKBAl5w.png" /></figure><h3>This isn’t an argument against model photography</h3><p>Worth being direct, because this argument gets made badly all the time.</p><p>Model photography does a job nothing else does. It communicates styling, silhouette, brand identity, context, aspiration, how the garment moves and falls, the creative world the brand is building. A strong campaign image sells a point of view, not just a dress. Brands invest in it because it works, and they should keep investing in it.</p><p>The limitation isn’t quality. It’s arithmetic. One image cannot show every future shopper.</p><p>The model photo answers <em>how does this garment look</em>. It was never designed to answer <em>how might this garment look on me</em>, and faulting it for that is like faulting a map for not being the place.</p><p>Those are complementary questions. Only one of them has historically been answerable online.</p><h3>From observation to self-evaluation</h3><p>A useful way to think about the journey — not a guaranteed funnel, just a shape:</p><p><strong>Observe the product → become interested → visualise it on yourself → evaluate it for yourself → decide</strong></p><p>The important move is in the middle. Everything before it is product evaluation. Everything after is self-evaluation, and that’s where purchases actually get decided.</p><p>Most product pages are built entirely for the first half.</p><h3>What a fitting room does</h3><p>A physical fitting room removes most of the imagination requirement. You don’t have to picture how something looks on you. You put it on and look.</p><p>Now the question is concrete. The answer might be yes. It might be no. It might be <em>try the black one</em> or <em>this needs a size up</em>.</p><p>Notice that a fitting room doesn’t exist to make people buy. Shops don’t measure them that way. It exists to turn an abstract question into something a person can actually assess.</p><p>That’s the right framing for virtual try-on too.</p><h3>What virtual try-on brings online</h3><p>Instead of only seeing <strong>model + garment</strong>, the shopper can see <strong>shopper + garment</strong>.</p><p>That doesn’t tell her whether to buy. It gives her something to evaluate instead of something to imagine.</p><p>And here’s the part worth sitting with: a good try-on can produce <em>yes, I like this on me</em>. It can equally produce <em>no, that’s not for me</em>.</p><p>Both are useful outcomes. One of them is a sale and one isn’t, but both resolve the question — which is what the shopper was stuck on.</p><h3>Where Tuck fits</h3><p>Tuck’s virtual try-on lets an apparel shopper see a garment visualised on her own photo. Its role in this story is narrow and specific: reduce the amount of imagination required to answer <em>how does this look on me</em>.</p><p>Not eliminate hesitation. Not guarantee confidence. Reduce one particular kind of visual uncertainty, which is one contributor among several.</p><p>Because even with that resolved, she may still be thinking: <em>which size do I order?</em></p><p>That’s a separate uncertainty needing separate information. Tuck’s Fit Intelligence compares shopper information against the merchant’s own size chart to suggest a size. It isn’t derived from the image, and no picture tells anyone whether to take the M.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bzwn-kdSNt3rrf1uP58tEw.png" /></figure><h3>A “no” can be a successful try-on</h3><p>Most marketing in this category celebrates one path: try on, add to cart, done.</p><p>Consider the other one.</p><p>A shopper likes a dress. She tries it virtually. She looks at the result and thinks <em>actually, no.</em> She closes the tab.</p><p>From a narrow conversion view, that session failed. Nothing was added to any cart.</p><p>From a decision-support view, it worked exactly as intended. She reached a conclusion she was previously unable to reach, and she reached it before placing an order rather than after it arrived.</p><p>We’re not going to claim that automatically reduces returns — that depends on far too many other things. But helping someone reject an unsuitable choice before ordering has obvious value, to her and to the merchant, and any honest account of what this technology does has to include it.</p><p>A try-on that only ever produces yes isn’t decision support. It’s persuasion wearing a decision-support costume.</p><h3>The space between a page view and a purchase</h3><p>Standard analytics sees discrete events. Page view. Click. Add to cart. Checkout. Purchase.</p><p>What it can’t see is why someone stayed undecided. There’s no event for <em>I’ll think about it</em>, and nobody should pretend to measure one.</p><p>But there’s a large space between a page view and a purchase where consideration happens, and that space is usually dark.</p><p>Interactive tools give merchants more observable behaviour inside it. Did she start a try-on? Finish it? Try several garments? Come back to the same product later? Save it? Add to cart afterwards?</p><p>None of that tells you what she was thinking. All of it tells you more than a page view does.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kSJiByZcu13ek9Pb8hY-Pg.png" /></figure><h3>What Shopify merchants should measure</h3><p>Eligible product-page traffic. Try-on start rate and completion rate. Garments tried per session. Repeat try-on behaviour. Add-to-cart after a try-on. Saves and wishlists where you capture them. Return visits. Category-level differences. Conversion among appropriately comparable cohorts rather than try-on users against all other traffic.</p><p>But the framing matters more than the list.</p><p>The question isn’t only <em>does try-on increase conversion</em>. It’s <em>does try-on help shoppers reach a decision</em> — where the decision is sometimes buy, sometimes try a different product, and sometimes not this one.</p><p>That’s a more demanding standard and a more honest one. It’s also the only standard under which a feature like this keeps earning its place once the novelty wears off.</p><h3>Back to the dress</h3><p>She liked it. She said she’d think about it.</p><p>Now imagine she can see it on herself.</p><p>That doesn’t necessarily turn her answer into yes. It might become yes. It might become no. It might become <em>try the black one</em>, or <em>maybe a size up</em>, or <em>save this for next month</em>.</p><p>All of those are better than what she had, because all of them are decisions.</p><p>The objective was never to manufacture certainty in the direction of the sale. It was to replace some of the imagination with something she could actually look at.</p><p>The opposite of “I’ll think about it” isn’t always “I’ll buy it.” Sometimes it’s just “now I know.”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1749f79f9e66" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why “I’ll Think About It” Usually Means “I Couldn’t Picture It”]]></title>
            <link>https://medium.com/@tuck.now/why-ill-think-about-it-usually-means-i-couldn-t-picture-it-80edc0f5e5ea?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/80edc0f5e5ea</guid>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Sat, 03 Oct 2026 13:36:29 GMT</pubDate>
            <atom:updated>2026-10-03T13:36:29.375Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lyLoMcOlBZI3CoQUvu8ePA.png" /></figure><p>A shopper turns her phone round to show a friend a dress.</p><p>“What do you think?”</p><p>“It looks great.”</p><p>She looks at it again. “Yeah. I like it.” A pause. “I’ll think about it.”</p><p>Nothing has obviously gone wrong. She likes the colour. The price isn’t impossible. Her size is in stock. She hasn’t found something better on another site.</p><p>And yet she hasn’t decided.</p><p>Before going further, a caveat the headline doesn’t have room for: “I’ll think about it” means a lot of things. Payday is Friday. She’s comparing three options. She wants her sister’s opinion. The shipping is slow. She’s not sure about the colour. Sometimes she’s just being polite about something she doesn’t want.</p><p>This article is about one particular version of it, and it’s a version that fashion ecommerce is unusually prone to producing.</p><p>She likes the dress. She can’t quite picture herself in it.</p><h3>Liking something isn’t the same as wanting it</h3><p>There’s a real distance between those two states, and it’s easy to miss because they feel adjacent.</p><p>You can like a jacket you’d never own. A coat in a magazine, a shirt on someone in a café, a dress on a mannequin. Liking is an aesthetic judgement about an object, and it commits to nothing.</p><p>Wanting something for yourself is different. It requires placing the thing in your life — in your wardrobe, on your body, at events you actually attend, next to the clothes you already own.</p><p>The product page asks the first question very well. <em>Do you like this garment?</em> Good photography is extremely effective at producing yes.</p><p>The purchase decision eventually asks a second one. <em>Do you want this garment for yourself?</em></p><p>Between those two questions sits a piece of work, and the shopper usually has to do it alone.</p><h3>The imagination gap</h3><p>We call this the imagination gap — our own shorthand rather than any established term. It describes something simple:</p><p><strong>The distance between seeing a product and being able to picture yourself wearing it.</strong></p><p>Traditional ecommerce asks the shopper to cross that distance mentally. She sees a model wearing a garment, and has to convert that into an image of herself wearing it.</p><p>Sometimes that’s effortless. If the model is roughly her build and colouring, if the styling matches how she dresses, if the garment is a simple shape she’s worn before — the translation happens instantly and she barely notices doing it.</p><p>Sometimes it isn’t. And when it isn’t, she doesn’t usually diagnose the problem. She doesn’t think <em>I lack sufficient visual information to complete this evaluation.</em> She just feels unsure.</p><p>That feeling leaves the page as “I’ll think about it.”</p><h3>Why fashion makes this harder than most categories</h3><p>A coffee machine can stay an object. You evaluate it on specifications, reviews, how it looks on a counter. It never really has to become personal.</p><p>Clothing can’t stay an object for long, because at some point it has to become <em>a jacket on me</em>.</p><p>Which means the shopper is weighing several things at once, whether or not she’d articulate them: the silhouette on her particular frame, the colour against her skin, the length, the proportions, how she’d style it, what occasion it’s for, whether it reads as bold or understated on her, and the quiet underlying question of whether it feels like her at all.</p><p>No single feature resolves all of that, and claiming otherwise would be overselling. But it explains why visual self-reference matters more in apparel than in almost any other category.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q3AE2y5kk3I9lR6sKBAl5w.png" /></figure><h3>This isn’t an argument against model photography</h3><p>Worth being direct, because this argument gets made badly all the time.</p><p>Model photography does a job nothing else does. It communicates styling, silhouette, brand identity, context, aspiration, how the garment moves and falls, the creative world the brand is building. A strong campaign image sells a point of view, not just a dress. Brands invest in it because it works, and they should keep investing in it.</p><p>The limitation isn’t quality. It’s arithmetic. One image cannot show every future shopper.</p><p>The model photo answers <em>how does this garment look</em>. It was never designed to answer <em>how might this garment look on me</em>, and faulting it for that is like faulting a map for not being the place.</p><p>Those are complementary questions. Only one of them has historically been answerable online.</p><h3>From observation to self-evaluation</h3><p>A useful way to think about the journey — not a guaranteed funnel, just a shape:</p><p><strong>Observe the product → become interested → visualise it on yourself → evaluate it for yourself → decide</strong></p><p>The important move is in the middle. Everything before it is product evaluation. Everything after is self-evaluation, and that’s where purchases actually get decided.</p><p>Most product pages are built entirely for the first half.</p><h3>What a fitting room does</h3><p>A physical fitting room removes most of the imagination requirement. You don’t have to picture how something looks on you. You put it on and look.</p><p>Now the question is concrete. The answer might be yes. It might be no. It might be <em>try the black one</em> or <em>this needs a size up</em>.</p><p>Notice that a fitting room doesn’t exist to make people buy. Shops don’t measure them that way. It exists to turn an abstract question into something a person can actually assess.</p><p>That’s the right framing for virtual try-on too.</p><h3>What virtual try-on brings online</h3><p>Instead of only seeing <strong>model + garment</strong>, the shopper can see <strong>shopper + garment</strong>.</p><p>That doesn’t tell her whether to buy. It gives her something to evaluate instead of something to imagine.</p><p>And here’s the part worth sitting with: a good try-on can produce <em>yes, I like this on me</em>. It can equally produce <em>no, that’s not for me</em>.</p><p>Both are useful outcomes. One of them is a sale and one isn’t, but both resolve the question — which is what the shopper was stuck on.</p><h3>Where Tuck fits</h3><p>Tuck’s virtual try-on lets an apparel shopper see a garment visualised on her own photo. Its role in this story is narrow and specific: reduce the amount of imagination required to answer <em>how does this look on me</em>.</p><p>Not eliminate hesitation. Not guarantee confidence. Reduce one particular kind of visual uncertainty, which is one contributor among several.</p><p>Because even with that resolved, she may still be thinking: <em>which size do I order?</em></p><p>That’s a separate uncertainty needing separate information. Tuck’s Fit Intelligence compares shopper information against the merchant’s own size chart to suggest a size. It isn’t derived from the image, and no picture tells anyone whether to take the M.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bzwn-kdSNt3rrf1uP58tEw.png" /></figure><h3>A “no” can be a successful try-on</h3><p>Most marketing in this category celebrates one path: try on, add to cart, done.</p><p>Consider the other one.</p><p>A shopper likes a dress. She tries it virtually. She looks at the result and thinks <em>actually, no.</em> She closes the tab.</p><p>From a narrow conversion view, that session failed. Nothing was added to any cart.</p><p>From a decision-support view, it worked exactly as intended. She reached a conclusion she was previously unable to reach, and she reached it before placing an order rather than after it arrived.</p><p>We’re not going to claim that automatically reduces returns — that depends on far too many other things. But helping someone reject an unsuitable choice before ordering has obvious value, to her and to the merchant, and any honest account of what this technology does has to include it.</p><p>A try-on that only ever produces yes isn’t decision support. It’s persuasion wearing a decision-support costume.</p><h3>The space between a page view and a purchase</h3><p>Standard analytics sees discrete events. Page view. Click. Add to cart. Checkout. Purchase.</p><p>What it can’t see is why someone stayed undecided. There’s no event for <em>I’ll think about it</em>, and nobody should pretend to measure one.</p><p>But there’s a large space between a page view and a purchase where consideration happens, and that space is usually dark.</p><p>Interactive tools give merchants more observable behaviour inside it. Did she start a try-on? Finish it? Try several garments? Come back to the same product later? Save it? Add to cart afterwards?</p><p>None of that tells you what she was thinking. All of it tells you more than a page view does.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kSJiByZcu13ek9Pb8hY-Pg.png" /></figure><h3>What Shopify merchants should measure</h3><p>Eligible product-page traffic. Try-on start rate and completion rate. Garments tried per session. Repeat try-on behaviour. Add-to-cart after a try-on. Saves and wishlists where you capture them. Return visits. Category-level differences. Conversion among appropriately comparable cohorts rather than try-on users against all other traffic.</p><p>But the framing matters more than the list.</p><p>The question isn’t only <em>does try-on increase conversion</em>. It’s <em>does try-on help shoppers reach a decision</em> — where the decision is sometimes buy, sometimes try a different product, and sometimes not this one.</p><p>That’s a more demanding standard and a more honest one. It’s also the only standard under which a feature like this keeps earning its place once the novelty wears off.</p><h3>Back to the dress</h3><p>She liked it. She said she’d think about it.</p><p>Now imagine she can see it on herself.</p><p>That doesn’t necessarily turn her answer into yes. It might become yes. It might become no. It might become <em>try the black one</em>, or <em>maybe a size up</em>, or <em>save this for next month</em>.</p><p>All of those are better than what she had, because all of them are decisions.</p><p>The objective was never to manufacture certainty in the direction of the sale. It was to replace some of the imagination with something she could actually look at.</p><p>The opposite of “I’ll think about it” isn’t always “I’ll buy it.” Sometimes it’s just “now I know.”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=80edc0f5e5ea" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Endowment Effect and the Fitting Room]]></title>
            <link>https://medium.com/@tuck.now/the-endowment-effect-and-the-fitting-room-1accff170115?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/1accff170115</guid>
            <category><![CDATA[virtual-try-on]]></category>
            <category><![CDATA[shopify]]></category>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Sat, 03 Oct 2026 04:52:44 GMT</pubDate>
            <atom:updated>2026-10-03T04:52:44.533Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*r0aA5TT9yCTQnRThhg779w.png" /></figure><p>Two versions of the same shopper, ten minutes apart.</p><p>In the first, she’s standing at a rail looking at a navy jacket. She likes it. She checks the price, looks at the sleeve, puts it back, and walks on. The jacket stayed a jacket — something she was looking at.</p><p>In the second, she takes it into a fitting room. Puts it on. Turns sideways. Rolls the sleeve back once, then again. Thinks about the wedding in June and whether it would work over the grey jumper she already owns. Takes a photo in the mirror.</p><p>Then she takes it off, and has to decide whether it goes back on the rail.</p><p>That second decision feels different from the first one. Not dramatically. But differently.</p><p>Nothing about the jacket changed. She doesn’t own it, no money moved, it’s the same garment it was five minutes ago.</p><p>What changed is what she was evaluating. In the first version, she was looking at a product. In the second, she was looking at herself in it.</p><p>That shift is small, and online retail mostly removed the thing that causes it.</p><h3>What the fitting room actually does</h3><p>It’s easy to describe a fitting room as an information booth. You go in unsure whether something suits you, you come out knowing.</p><p>True, but it misses the more interesting part.</p><p>A fitting room isn’t somewhere you observe a garment. It’s somewhere you <em>use</em> one. You wear it, control it, adjust it, move in it, and start placing it in your actual life — the wedding, the office, the jumper you already own.</p><p>That’s a different kind of encounter from studying the same jacket on a hanger, however carefully you study it.</p><p>And it changes the sentence in the shopper’s head. Before: <em>that looks good.</em> After: <em>I like how that looks on me.</em></p><p>Those two sentences are doing very different work. Only the second one contains her.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*d9_G5YuGwFqYmFqEKZcprQ.png" /></figure><h3>Two ideas that help explain it</h3><p>Economists have a name for the way ownership changes how we value things. Once something is ours, we tend to price it differently — usually higher than we’d have paid to get it in the first place. That’s the <strong>endowment effect</strong>, and most people recognise it instantly once it’s described. It’s why selling a car you’ve driven for six years is such an irritating experience.</p><p>The fitting room isn’t quite that, though. She doesn’t own the jacket. She had it for four minutes and handed it back.</p><p>The closer idea is <strong>psychological ownership</strong> — the feeling that something has become personally connected to you before you formally own it.</p><p>That fits better. She doesn’t own the jacket. But she has seen herself in it, adjusted it, and imagined using it. It moved from something she was looking at to something she was weighing up as part of her own life.</p><p>Two terms, one simple observation underneath: trying something on changes your relationship with it.</p><h3>Why ecommerce loses this</h3><p>Standard fashion ecommerce is very good at showing products. Model images, product shots, detail crops, video, descriptions, reviews. All of it does real work and none of it is the problem.</p><p>But it’s all external. The shopper looks <em>at</em> the garment throughout.</p><p>There’s no moment where she interacts with it, controls it, or sees it against herself. The nearest thing available is imagination — studying a model who’s taller, differently proportioned and professionally styled, and translating that into herself.</p><p>Some shoppers do that easily. Plenty don’t, and those are the ones who leave still unsure.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gushzNYsSaGSp2GF9wb7GA.png" /></figure><h3>What virtual try-on brings back</h3><p>A physical fitting room turns <em>how does this jacket look</em> into <em>how does this jacket look on me</em>.</p><p>Virtual try-on is an attempt to bring part of that transition online. Instead of only seeing the garment on a model, the shopper sees it visualised on herself. The reference point moves from <strong>model + product</strong> toward <strong>shopper + product</strong>.</p><p>We’re not claiming that produces psychological ownership, and certainly not an endowment effect. Those are useful ways to think about what a fitting room does, not switches anyone can flip.</p><p>The behavioural observation is enough on its own: something changes when a shopper can evaluate a garment on herself rather than only look at it.</p><h3>What it can’t bring back</h3><p>A photo is not a fitting room, and any argument pretending otherwise deserves suspicion.</p><p>A physical fitting room gives you touch, fabric weight, movement, physical fit feedback and direct manipulation. You learn things in there that no screen conveys — whether the wool itches, whether the shoulders pull when you reach forward.</p><p>Virtual try-on gives you personal visualisation, visual self-reference, digital interaction, and the ability to explore a garment on yourself without having it in your hands.</p><p>The claim isn’t that virtual try-on equals a fitting room. It’s narrower and more defensible: <strong>virtual try-on can restore one important behaviour ecommerce lost — evaluating the product in relation to yourself.</strong></p><p>That’s the part worth building for.</p><h3>Where Tuck fits</h3><p>Tuck’s virtual try-on lets an apparel shopper see a garment visualised on her own photo. That’s the whole job: moving the product page from model-plus-product toward shopper-plus-product.</p><p>Sizing is a separate question and we treat it that way. Tuck’s Fit Intelligence compares shopper information against the merchant’s own size chart to suggest a size. It isn’t derived from the try-on image, it has nothing to do with any of the psychology above, and an image doesn’t tell anyone whether to take the M.</p><h3>This isn’t a conversion trick</h3><p>Ideas like these get treated as a toolkit of tactics, and that’s the wrong reading.</p><p>The question isn’t how to exploit psychological ownership so shoppers feel reluctant to leave. It’s how to give online shoppers some of the product interaction physical retail provides by default. She should stay completely free to say yes, no, not sure, save it, or try something else. Decision support, not decision manipulation — and an experience that only works by narrowing someone’s options isn’t support.</p><h3>What a Shopify merchant should measure</h3><p>You can’t measure psychological ownership from Shopify analytics, and shouldn’t try. What you can measure is behaviour.</p><p>Try-on engagement rate. Completion rate. Garments tried per session. Repeat try-on interactions. Add-to-cart after a try-on. Save and wishlist behaviour. Return visits. Conversion among appropriately comparable cohorts rather than users against everyone. Differences between product categories.</p><p>These tell you whether shoppers behave differently after interacting with a try-on. They don’t prove anything about what shoppers are feeling.</p><p>The practical version of this article’s argument is a change in the question merchants ask. Not <em>how many product images should I show</em> — that’s an observation question. But <em>how do I help this shopper evaluate the garment for herself</em>.</p><h3>Back to the fitting room</h3><p>She puts the jacket on. Looks. Takes it off. Decides whether it goes back.</p><p>The interesting thing isn’t that psychology can make her keep it. It can’t, and chasing that is the wrong ambition.</p><p>The interesting thing is that after four minutes, she’s evaluating something different from what she was evaluating outside. Not a jacket. Herself in a jacket.</p><p>Online retail removed most of that and spent twenty years compensating with better photography.</p><p>The fitting room was never only where shoppers checked the clothes. It was where the clothes stopped being abstract.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1accff170115" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Your Try-On Tool Is Also a Lead Channel. Here’s the Maths.]]></title>
            <link>https://medium.com/@tuck.now/your-try-on-tool-is-also-a-lead-channel-heres-the-maths-1b2ae2d0cb8d?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/1b2ae2d0cb8d</guid>
            <category><![CDATA[fashiontech]]></category>
            <category><![CDATA[shopify]]></category>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Sat, 26 Sep 2026 05:49:52 GMT</pubDate>
            <atom:updated>2026-09-26T05:49:52.870Z</atom:updated>
            <content:encoded><![CDATA[<p>A shopper opens a dress on your product page. She uploads a photo, runs a try-on, looks at the result. Maybe she runs it again in a second colour. Then she closes the tab.</p><p>Your analytics records the session as no purchase.</p><p>Commercially, though, something happened. This person selected a specific garment, spent real time with it, and evaluated it against herself. That is not the behaviour of someone idly browsing a category page.</p><p>So the question worth asking is whether that intent has to evaporate when the tab closes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*m8rLRN1AeG1w0qltDJRl2A.png" /></figure><h3>A try-on is a different kind of product-page event</h3><p>Most product-page metrics measure presence. Someone loaded the page, scrolled some portion of it, stayed for a while. Useful, but weak signal — people land on product pages by accident, from search, from a mis-clicked ad.</p><p>A try-on is participation. The shopper has picked one garment out of your catalogue, taken an action that requires effort, and waited for a result.</p><p>Worth being careful about the claim here: not every try-on user is a high-intent buyer. Some are curious about the technology. Some are trying things they’d never buy. But as a behavioural signal, a completed try-on sits meaningfully above a passive page view, and most merchants treat both the same way in their reporting.</p><h3>From engagement signal to lead opportunity</h3><p>Here’s the distinction the headline needs, because it’s easy to overstate.</p><p>A try-on is not a lead. A shopper becomes a lead when there’s an identity and permission event — choosing to save a look, asking to be sent the result, continuing an account or fit profile, adding to a wishlist, opting into a notification about that item.</p><p>The try-on creates the opportunity. A permissioned next step is what turns it into a lead.</p><p>Which means the honest version of this article’s title is: your try-on tool <em>can be</em> a lead channel, if you deliberately build a continuation path around it. Left alone, it’s a feature that either converts in-session or doesn’t.</p><h3>Don’t gate the value</h3><p>One design principle matters more than the rest: give the shopper the value first.</p><p>The tempting pattern is an email form in front of the try-on. Enter your address to see yourself in this dress. It will capture addresses, and some will be fake, and you’ll have introduced friction at precisely the moment a shopper was leaning in.</p><p>The better sequence:</p><p><strong>Product page → try-on → result → value experienced → optional next step</strong></p><p>The shopper came for utility. Give them the utility. Then ask whether they’d like to keep it — save this look, save this item, notify me if my size comes back, see this later.</p><p>Those are illustrative phrasings rather than a feature list, and how you implement them depends on your stack. The principle is what matters: value before capture. A capture mechanism that degrades the experience it’s attached to will eventually reduce the number of interactions worth capturing.</p><p>This is also why the comparison with a generic email pop-up is about context rather than performance. A pop-up fires because someone arrived at your website. A try-on follow-up appears after they chose a garment, engaged with it, and saw themselves in it. I’m not claiming a higher conversion rate — that would need evidence neither of us has. I’m claiming the intent context is fundamentally different, and that difference is worth designing around.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*m4xKsi8a9tFPorppqRV7Jw.jpeg" /></figure><h3>Here’s the maths</h3><p>Four steps, each of which you can populate from your own analytics.</p><p><strong>Step 1 — Try-on engagement</strong>Eligible PDP visitors × try-on engagement rate = try-on users</p><p><strong>Step 2 — Permissioned leads</strong>Try-on users × permissioned capture rate = new leads</p><p><strong>Step 3 — Lead-assisted purchases</strong>Permissioned leads × subsequent conversion rate = purchases from the captured cohort</p><p><strong>Step 4 — Revenue</strong>Purchases × average order value = revenue associated with the captured cohort</p><p>That fourth line needs precision. It is <em>associated</em> revenue, not incremental revenue. Some of those shoppers would have come back and bought anyway. Treating the whole figure as caused by the lead capture will overstate your result, sometimes considerably.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*asgcl1fW9ZZsZRcWKxksiw.jpeg" /></figure><p>100,000 eligible PDP visitors × 8% use virtual try-on → <strong>8,000 try-on users</strong> × 20% choose to save or continue → <strong>1,600 permissioned leads</strong> × 10% later purchase → <strong>160 purchases</strong> × $120 AOV → <strong>$19,200 revenue associated with the captured cohort</strong></p><p><strong>THESE ARE ILLUSTRATIVE ASSUMPTIONS, NOT TUCK PERFORMANCE BENCHMARKS.</strong></p><p>The 8%, 20%, 10% and $120 are placeholders chosen to demonstrate the equation. They are not industry averages and they are not results from Tuck deployments. Change any one of them and the outcome changes substantially — which is the point. Use your own traffic, engagement, capture rate, follow-up conversion, AOV and margin.</p><h3>Contribution, not just revenue</h3><p>Revenue isn’t profit. To get closer to what this is worth:</p><p>Revenue from captured cohort × gross margin − cost of the try-on experience − follow-up / CRM cost − any incentive cost = estimated contribution</p><p>Estimated, again, and not automatically incremental.</p><p>If you’re deliberately treating try-on as an acquisition surface, cost per lead is the natural metric:</p><p>Total attributable cost ÷ permissioned leads generated = cost per lead</p><p>But be careful how you allocate. Virtual try-on is a multi-outcome investment. The same interaction may be producing same-session conversion, better-informed purchases with different post-purchase behaviour, and lead capture. Charging the entire program cost against lead generation will make your CPL look terrible and tell you nothing useful. Charging none of it will flatter the channel. Pick a defensible split and stay consistent.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*M-mUfejVKZaIcStCAPcpRQ.jpeg" /></figure><h3>What is a try-on lead actually worth?</h3><p>A list of email addresses is worth nothing by itself. The commercial question is what happens next — follow-up engagement, return visits, revisits to that product, add-to-cart, purchase, AOV, margin, and return behaviour on those orders.</p><p>There’s also context. A newsletter signup tells you someone wants email from your brand. A permissioned save after a try-on may additionally tell you which garment they were considering. That supports more relevant follow-up: a saved-look reminder, a back-in-stock notification for their size, a price-drop alert, a wishlist nudge, related pieces from the same collection.</p><p>Those are merchant journeys you’d build in your own CRM and stack — describing what’s possible around the interaction, not a claim about what any try-on provider does for you automatically.</p><h3>Privacy is part of the product, not a disclaimer</h3><p>There’s a wrong version of this strategy, and it’s worth naming so nobody drifts into it.</p><p>The wrong version treats the try-on as a data-harvesting event: capture as much as possible because the shopper engaged, and market to them on the strength of it.</p><p>Two things need to stay separate. Processing a try-on is one permission. Receiving marketing communications is a different permission, and consenting to the first is not consent to the second. Shoppers understand this intuitively, and brands that blur it tend to find out the hard way.</p><p>The image itself should never become marketing data. Tuck’s public FAQ states that identifiable shopper photos are not stored, and that only the numerical information required for the Fit Profile is retained, encrypted at rest. The continuity mechanism is a fit profile the shopper chooses to keep, which is separate from any marketing relationship.</p><p>The context worth carrying into follow-up is ordinary commerce context — the product, the collection, the saved item, a stated preference — where the shopper has permitted it. Not their photo, and not biometric information.</p><p>A lead channel that costs you trust isn’t an acquisition strategy. It’s a one-off extraction with a long tail of unsubscribes.</p><h3>Why this changes how you value try-on</h3><p>The usual evaluation runs: did the shopper use try-on, and did they buy in that session?</p><p>That’s a reasonable question and an incomplete one, because the interaction may produce value along several paths: immediate conversion, post-purchase outcomes on better-informed orders, permissioned lead capture, and later conversion from that captured cohort.</p><p>Measuring only same-session conversion <em>can</em> understate the commercial role of the interaction. Not always — if you’ve built no continuation path, there’s nothing to understate. But if you have, and you’re not measuring it, the feature is doing work that doesn’t appear anywhere in your reporting.</p><p>Worth saying plainly: establishing what’s genuinely incremental requires holdout groups, controlled campaigns or comparable cohorts, plus a defined attribution window. Without that, you have an associated-revenue figure, which is useful directionally and shouldn’t be presented internally as anything more.</p><h3>What the dashboard should track</h3><p><strong>Try-on engagement</strong> — eligible PDP sessions, try-on starts, completions, engagement rate.</p><p><strong>Lead capture</strong> — prompts shown, permissioned captures, capture rate, cost per captured lead.</p><p><strong>Follow-up</strong> — CRM engagement, return visits, product revisits, add-to-cart, purchases.</p><p><strong>Commercial</strong> — revenue from the captured cohort, gross margin, return behaviour, cost per try-on, campaign cost.</p><p>Don’t stop at lead count. Lead count is the easiest number to grow and the least informative. What matters is what the cohort contributes.</p><h3>Back to the shopper who left</h3><p>She tried the dress. She didn’t buy it.</p><p>Depending on what you built around that moment, three things may have happened anyway. She may have left with more confidence about the garment than she arrived with. You may have learned that genuine product-level consideration existed on that SKU. And if she chose to save the look, you may have earned permission to continue the conversation.</p><p>The rendered image doesn’t have to be the end of the funnel.</p><p>But it only becomes the start of a relationship when the value comes first, the permission is explicit, and somebody is actually measuring what happens next.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1b2ae2d0cb8d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Reverse Logistics and Return Spoilage: The Cost Nobody Puts in the Model]]></title>
            <link>https://medium.com/@tuck.now/reverse-logistics-and-return-spoilage-the-cost-nobody-puts-in-the-model-0ba20b0ca917?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/0ba20b0ca917</guid>
            <category><![CDATA[shopify]]></category>
            <category><![CDATA[shopify-store]]></category>
            <category><![CDATA[fashiontech]]></category>
            <category><![CDATA[fashion]]></category>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Thu, 24 Sep 2026 09:16:18 GMT</pubDate>
            <atom:updated>2026-09-24T09:16:18.689Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pBoTkv3ZBL6zrX0q7JdT6A.png" /></figure><p>An item sells for $80. The customer sends it back. You refund $80.</p><p>On that arithmetic, the cost of ecommerce returns looks simple. It’s the mental model most merchants carry, and it’s the one most reporting quietly reinforces.</p><p>But when the refund clears, the garment isn’t back on the shelf. It’s somewhere between a customer’s hallway and your next available sale, and the gap between those two points is where the rest of the cost lives.</p><h3>A return is a process, not an event</h3><p>Processes vary between merchants, 3PLs and platforms, but a returned apparel item generally moves through some version of the same sequence: the customer initiates the return, the item travels back, it’s received, inspected and graded, the refund is reconciled, it’s repackaged where possible, and then it’s restocked, marked down, redirected, liquidated or written off.</p><p>Every step takes time and most of them cost money — return shipping, carrier handling, warehouse receiving, inspection labour, sorting, repackaging materials, inventory updates, restocking.</p><p>The useful distinction is this: <em>returned</em> is not the same as <em>back in sellable inventory</em>. Between those two states sit days or weeks, and a decision about what the item is now worth.</p><p>Some of that value can erode along the way. I’ll call this <strong>return spoilage</strong> — an editorial framework for this article rather than a formal accounting term. It describes the recoverable merchandise value lost while an item moves through the return process: missed full-price selling time, markdown exposure, packaging that no longer presents as new, a product that can’t be sold as new, and time spent unavailable to the next customer who wanted it.</p><p>Not every return spoils. Plenty come back quickly, unworn and intact, and go straight into stock at full price. But two returns of the same SKU can produce very different recovery economics depending on processing time, condition, seasonality and which resale path they end up on. Return rate counts both the same.</p><h3>The scale behind the return rate</h3><p><strong>National Retail Federation, 2025 Retail Returns Landscape:</strong> approximately 19.3% of online sales projected to be returned in 2025, across retail generally rather than apparel specifically.</p><p><strong>National Retail Federation, 2025:</strong> approximately $849.9 billion in projected total US retail returns — all retail categories, not apparel alone.</p><p><strong>McKinsey apparel-retail research, 2021:</strong> processing costs between different return channels differed by roughly $5–$6 on average.</p><p><strong>McKinsey apparel-retail research, 2021:</strong> some faster in-store return paths reduced processing time by roughly 12–16 days compared with slower alternatives.</p><p><strong>McKinsey retailer survey, 2021:</strong> approximately 70% of surveyed returns were attributed to fit or style.</p><p>Three things are worth drawing out of that.</p><p>Processing has genuine cost, and the route a return takes changes it. Time has genuine economic consequence — a difference of two weeks in processing is a difference in which pricing window the item returns to. And historically, uncertainty about fit and style has been a substantial driver of apparel returns, at least among the retailers McKinsey surveyed.</p><p>A merchant carries a seasonal dress in limited size depth. A customer orders it and sends it back. The parcel spends several days in transit, then waits for inspection, then waits to be restocked. By the time it’s sellable again, the promotional period has started. The merchant may well resell it — but under different economics than the original full-price sale, and only after the size was unavailable to other shoppers during the weeks it was in the system. Nothing was damaged and nobody made a mistake. The value simply eroded in transit.</p><h3>The more interesting question: where did the return begin?</h3><p>Everything above concerns what happens once a return exists. Reverse logistics is the discipline of recovering as much value as possible from that point forward, and it’s worth doing well.</p><p>But it starts after a decision has already been made.</p><p><strong>Reverse logistics asks how much value can be recovered once a return exists. The more interesting question is whether some of that value loss needed to begin at all.</strong></p><p>That question points backwards, to the product page, and to the moment a shopper decided to order something they weren’t quite sure about.</p><h3>Where Tuck intervenes</h3><p>Tuck does not manage reverse logistics. It operates before the purchase decision is finalised, at the point where a shopper is still deciding.</p><p>The sequence looks like this:</p><p><strong>Product page → shopper uncertainty → Tuck → better-informed purchase decision → order → possible return</strong></p><p>Tuck works at the uncertainty stage. That scope needs stating honestly, because plenty of returns have nothing to do with it. Product defects, quality disappointment, late or damaged delivery, straightforward change of mind, gifting, fraud and deliberate bracketing all sit outside anything a product page can influence. No pre-purchase technology addresses those, and claiming otherwise would be dishonest.</p><p>What Tuck addresses is the subset of uncertainty that exists while the shopper is still on the page — and it treats that as two separate problems, because it is.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*S40hFw2QYp94XWPbF9x32g.png" /></figure><h3>Virtual try-on addresses visual uncertainty</h3><p>The question is <em>how does this look on me?</em></p><p>Without personalisation, a shopper does this translation themselves. They look at a garment on a model — a particular height, build and colouring, styled and lit for the brand — and mentally convert that into a guess about how it might look on their own body. Some shoppers are good at this. Many aren’t, and the ones who aren’t are guessing at the point of purchase.</p><p>Tuck’s virtual try-on gives them another input. The shopper provides their own photo, and the garment is visualised on them. Colour against their skin, how the silhouette reads on their proportions, whether the piece looks like something they’d actually wear.</p><p>This is visualisation, and the boundary matters: the image doesn’t determine physical fit, and it doesn’t tell the shopper which size to order. It resolves the visual question, which is a real question, and not the only one.</p><h3>Fit Intelligence addresses sizing uncertainty</h3><p>The question is <em>which size should I choose?</em></p><p>This is a different problem requiring different inputs. Tuck’s Fit Intelligence takes shopper information and compares it against the merchant’s own size chart to produce a size suggestion. Not a generic sizing assumption — the brand’s actual sizing, because a size 10 means different things at different labels and shoppers know it.</p><p>Two capabilities, two questions, kept deliberately distinct:</p><p><strong>Virtual try-on → visual confidence.Fit Intelligence → sizing confidence.</strong></p><p>They matter together because a shopper can hold both uncertainties at once, or either one alone. Someone can love how a jacket looks on them and still have no idea whether to take the M. Someone else can know their size in a brand and still be unsure the cut suits them. Resolving one and leaving the other means the shopper is still guessing about something when they click buy.</p><p>The argument here isn’t that Tuck prevents returns. It’s narrower and more defensible: if some portion of avoidable returns originates in unresolved visual or sizing uncertainty, then resolving more of that uncertainty before checkout reduces the number of doubtful purchase decisions entering the downstream system at all.</p><h3>What our numbers actually mean</h3><p>Tuck publishes several figures, and they’re worth interpreting rather than displaying.</p><p>Tuck reports upto <strong>24% conversion lift</strong> where its experience is deployed. That’s the pre-purchase side of the ledger: a signal that reducing uncertainty affects whether shoppers complete the purchase. It’s Tuck’s result in Tuck’s deployments, not a number any merchant should plan around.</p><p>Tuck’s published production results report upto <strong>25% reduction in returns</strong>. This is the figure most relevant to everything above, and the reason is worth spelling out. A return avoided upstream isn’t only a refund avoided. It’s an order that never enters return shipping, receiving, inspection, processing, restocking or markdown exposure. The refund line is the visible part; the operational cost and the value erosion sit behind it. That’s the bridge between a product-page feature and a logistics cost centre.</p><p>The range matters as much as the number. It varies by merchant, and it will vary for you.</p><p>And <strong>$0.043 average cost per try-on</strong>, from Tuck’s homepage. The useful way to think about this isn’t as a price. It’s that resolving uncertainty before purchase has a cost, and letting an uncertain purchase become a return has a different cost — shipping, handling, processing, inventory downtime and whatever value erodes on the way back. Those two costs are comparable, but only using your own return handling costs, margins and volumes. We’re not going to do that arithmetic for you with invented numbers.</p><h3>Prevention economics and recovery economics</h3><p>There are two ways to improve return economics, and they aren’t substitutes.</p><p><strong>Recovery economics</strong> applies once a return exists: route it efficiently, inspect it quickly, restock faster, preserve resale value, choose the right disposition. This is reverse-logistics work and it pays off regardless of what happens upstream.</p><p><strong>Prevention economics</strong> applies before the order: reduce visual uncertainty, reduce sizing uncertainty, help the shopper make a better-informed choice.</p><p>Every fashion retailer needs the first. Returns will happen regardless of how good the product page gets, and handling them badly is expensive.</p><p>But a business that works only on recovery is optimising what happens after the problem has already entered the system. The parcel is already moving. The best available outcome is a smaller loss.</p><h3>What should a fashion brand measure?</h3><p>Connecting these datasets gives a far better picture than conversion or return rate alone.</p><p><strong>Pre-purchase:</strong> virtual try-on usage, size recommendation usage, conversion among exposed cohorts, add-to-cart behaviour, bracketing patterns where measurable.</p><p><strong>Return outcome:</strong> overall return rate, size-related return rate, style or expectation-related return rate, return reason by SKU, return rate by size and category.</p><p><strong>Reverse logistics:</strong> processing cost per return, days from return initiation to receipt, days from receipt to restock or disposition, share restocked at full price, share requiring markdown, recovered merchandise value.</p><p><strong>Commercial:</strong> net sales, gross margin after returns, total return handling costs, total program cost.</p><p>Most merchants can’t measure all of this immediately, and don’t need to. The two timing metrics and the split between full-price restock and markdown are usually the highest-value place to start, because they’re what separates a cheap return from an expensive one.</p><h3>The cost begins before the parcel comes back</h3><p>The cost of ecommerce returns is not order value multiplied by return rate.</p><p>A serious model asks what was refunded, what it cost to process, how long the inventory was unavailable, and what value was recovered when it finally sold.</p><p>Return rate tells you how often products come back. Return economics tells you what happened to their value when they did. And the earliest place to influence either is the moment before someone decides.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0ba20b0ca917" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Fitting Room Was Never About the Garment]]></title>
            <link>https://medium.com/@tuck.now/the-fitting-room-was-never-about-the-garment-a8d6cf02deb5?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/a8d6cf02deb5</guid>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Sat, 19 Sep 2026 12:10:49 GMT</pubDate>
            <atom:updated>2026-09-19T12:10:49.762Z</atom:updated>
            <content:encoded><![CDATA[<p>A product page shows you clothes. A fitting room shows you yourself. Online apparel has spent twenty years working on the wrong half of that.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dBE2JBsEmigBFNTwC5_XnA.png" /></figure><p>Something specific happens in the two seconds after you pull the curtain closed and turn toward the mirror.</p><p>You stop looking at the garment.</p><p>Up to that point, you’ve been evaluating an object. Nice colour. Interesting cut. Reasonable price for what it is. Standard consumer assessment, the same posture you’d bring to a kettle.</p><p>Then you put it on, and the assessment changes shape entirely. You’re no longer asking whether this is a good jacket. You’re asking whether it’s a good jacket <em>on you</em> — whether the shoulders sit right on your frame, whether the colour does anything for your skin, whether you look like yourself in it or like someone attempting a different personality.</p><p>The garment has stopped being a product. It’s become a possibility.</p><p>That transition is the fitting room’s actual function, and it’s easy to miss because the room itself is so unremarkable. A curtain, a hook, a mirror, bad lighting. But the space is doing something a display rack cannot: it’s moving the shopper from evaluating an item to evaluating themselves wearing it.</p><p>Online, that transition mostly doesn’t happen.</p><h3>What the product page carries, and what it can’t</h3><p>This is where it would be easy to blame product photography, and that would be both unfair and wrong.</p><p>Photography does its job well. It shows the garment — the cut, the drape, the colour, the styling, the mood a brand is reaching for. A good shot communicates more about a piece in one frame than a paragraph of description manages. Brands invest in it because it works.</p><p>But photography answers <em>what is this</em>. It was never designed to answer <em>what is this on me</em>, and asking it to do so misunderstands what it’s for. A model wears the garment beautifully, and the shopper is left performing a mental translation from that body to their own, with whatever imagination and self-knowledge they happen to have available at 11pm.</p><p>Everything else on the page has the same character. The description covers fabric and construction. The size chart offers measurements in centimetres. Reviews supply strangers’ opinions about whether it runs small. A size selector asks for a commitment.</p><p>Each component is doing real and useful work. Collectively they describe the garment thoroughly and the shopper barely at all.</p><p>So the personal evaluation — the part the fitting room existed for — gets deferred. It happens after the box arrives, in a bedroom, with a return label as the safety net. The shopper still tries things on. They just do it on the other side of the purchase.</p><h3>The thinking that shaped Tuck</h3><p>The obvious way into this problem, for a company working with apparel imagery, is to ask how you generate a convincing picture of a garment on a person.</p><p>That’s a legitimate technical question. It also isn’t quite the shopper’s question, and starting there quietly changes what you optimise for. You end up chasing how the output looks rather than what it resolves.</p><p>We started from the fitting room instead. Not as a metaphor, but as a specification: what did that room let a shopper work out, and how much of it can happen before checkout rather than after?</p><p>Some of it can’t, and there’s no use pretending otherwise. Fabric weight, the way something moves when you walk, whether a collar irritates your neck — none of that survives the translation to a screen.</p><p>But two things a fitting room settles are addressable. They just aren’t the same problem, and treating them as one is where a lot of thinking in this space goes sideways.</p><p>The first is <em>how does this look on me</em>. That’s the mirror question, and it’s what Tuck’s virtual try-on addresses. The shopper provides their own photo, and the garment is visualised on them rather than on a fit model. Colour against their skin, how the silhouette reads on their proportions, whether the piece looks like something they’d recognise as theirs. The output builds visual confidence. It’s a visualisation, and it isn’t measuring anything.</p><p>The second is <em>which size should I choose</em>, and it needs entirely different inputs. Information about the shopper — height, weight, age, how they prefer clothes to sit — checked against the merchant’s own size chart, because sizing is brand-specific and generic assumptions don’t survive contact with how any particular brand cuts.</p><p>That’s Tuck’s Fit Intelligence, and we built it as a separate capability rather than folding it into the image. An image doesn’t tell you your size. A size suggestion doesn’t come from looking at a picture. They draw on different information, they can be wrong in different ways, and blending them would make it easy to imply one proves something it doesn’t.</p><p>Neither offers a guarantee. A suggestion is a suggestion, and bodies vary more than any chart accounts for. What changes is that the shopper gets two informed answers while they’re deciding, rather than a photograph and a link to a measurement table.</p><h3>What this is actually about</h3><p>The prevailing question in this category is how to make generated apparel imagery more convincing. That’s getting easier every year, and it will keep getting easier.</p><p>The more useful question is what the shopper is trying to work out, and whether the technology is helping them work it out.</p><p>Those sound similar and lead somewhere different. One produces better pictures. The other produces better decisions, which occasionally requires admitting that a picture isn’t the right instrument for part of the job.</p><p>A fitting room was never really about the garment. It was about the person in front of the mirror, briefly able to see what a thing looked like as part of their own life rather than as stock on a rail.</p><p>That’s the experience worth rebuilding. Not the room — the moment inside it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a8d6cf02deb5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why We Built Tuck Around the Fitting Room]]></title>
            <link>https://medium.com/@tuck.now/why-we-built-tuck-around-the-fitting-room-73b48b0266f6?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/73b48b0266f6</guid>
            <category><![CDATA[vton]]></category>
            <category><![CDATA[ecomerce]]></category>
            <category><![CDATA[virtual-try-on]]></category>
            <category><![CDATA[shopify-plus]]></category>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Fri, 18 Sep 2026 06:27:51 GMT</pubDate>
            <atom:updated>2026-09-18T06:27:51.557Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NOk4OZNsK6LbigW9qpHkyg.png" /></figure><h3>What the Fitting Room Was Actually For</h3><p><em>In a store, trying something on costs nothing. Online, you have to buy it first. That inversion explains more about apparel ecommerce than most people realise.</em></p><p>The fitting room is one of the most quietly effective pieces of retail design ever built, and almost nobody thinks about why.</p><p>It isn’t a display space. It isn’t merchandising. Its function is permission: you can take three garments off a rail, put them on, look at yourself, and walk out having bought none of them. Trying is free, and you only commit once you’ve stopped being uncertain.</p><p>When apparel moved online, that sequence flipped.</p><p>To try something on now, you buy it. It arrives, you put it on in your bedroom, and if it isn’t right you undo the transaction — repack it, print a label, wait for a refund.</p><p>The trying didn’t disappear. It moved to after the purchase. And the cost of it moved onto the merchant.</p><p>I think this is the most underexamined fact in online apparel. Returns are usually discussed as a logistics problem, or a margin problem, or a sustainability problem. They’re all of those. But structurally, a return is a fitting room that happened in the wrong place, paid for by the wrong party.</p><h3>What the product page put in its place</h3><p>The fitting room’s job didn’t vanish. It got distributed.</p><p>Photography shows the garment on a model. A description covers fabric and cut. A size chart supplies measurements. Reviews offer crowd-sourced opinions on whether it runs small. A size selector demands an answer. A returns policy absorbs whatever the rest failed to settle.</p><p>None of these components is bad. Photography sells clothes, and it should. Size charts are useful if you own a tape measure.</p><p>The problem is what they ask of the shopper. Each piece holds a fragment of the answer, produced without this particular body in mind, and the shopper is expected to assemble a decision out of them. The model is 5&#39;11&quot; in a size S. The measurements are in centimetres. One review says size down and the next one disagrees.</p><p>A fitting room never asked anyone to do that. It just let them try.</p><h3>Starting from the decision, not the image</h3><p>When we started building Tuck, there was an obvious framing available: how do you generate a convincing image of a garment on a person?</p><p>That’s a real technical question. It’s also not the shopper’s question.</p><p>The shopper wants to know whether to buy this, and in what size. Image quality matters insofar as it serves that. Start from image generation and you optimise for how the output looks. Start from the decision and you end up asking a different question — what does someone actually need resolved before they’re willing to commit?</p><p>So we worked backwards from the fitting room. Not as a metaphor, as a specification. What did it resolve, and which of those things can move back to before checkout?</p><p>Some can’t. You can’t judge fabric weight through a screen, and pretending otherwise would be dishonest. But two things a fitting room settles are addressable digitally — and they’re addressable by different means, which turns out to matter a great deal.</p><h3>Two things, kept separate</h3><p>The first is seeing yourself in the garment. Not a model — you, with your colouring and proportions.</p><p>That’s what Tuck’s virtual try-on does. The shopper provides a photo, and the garment is visualised on them. It answers <em>how does this look on me</em>.</p><p>It’s a visualisation. It builds visual confidence. It doesn’t measure anything and it isn’t a verdict on whether a size will fit, which is a distinction worth holding onto rather than blurring.</p><p>The second thing is sizing, and a fitting room settles that physically rather than visually.</p><p>Online, it needs its own inputs: information about the shopper — height, weight, age, how they like clothes to sit — checked against the merchant’s own size chart, because sizing is brand-specific and generic assumptions don’t survive contact with how brands actually cut.</p><p>That’s Tuck’s Fit Intelligence, and we built it as a separate capability rather than folding it into the image. They use different information. They can be wrong in different ways. Merging them would make it easy to imply a picture proves something it doesn’t.</p><p>Neither guarantees anything. A suggestion is a suggestion, and bodies are more varied than any chart. But two informed answers at the point of decision is a different proposition from an image and a link to a measurement table.</p><h3>What this means for apparel commerce</h3><p>The prevailing question in this category is how to generate better images of clothes on people. That’s getting easier and will continue to.</p><p>The more interesting question is how much of the fitting room can come back.</p><p>Not the room. Not everything that happened inside it. But the central thing — being able to try before committing — was never inherently physical. It became physical because of how online retail was assembled, not because of anything fundamental about buying clothes.</p><p>A product page is currently a presentation surface. There’s no strong reason it couldn’t be a decision surface.</p><p>That’s the work. Tuck exists because when the fitting room disappeared, the problem it was solving stayed exactly where it was.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=73b48b0266f6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Fitting Room Answered Two Questions. Now Ecommerce Can Too.]]></title>
            <link>https://medium.com/@tuck.now/the-fitting-room-answered-two-questions-now-ecommerce-can-too-a2f4f14b6916?source=rss-9d03e79fdb1a------2</link>
            <guid isPermaLink="false">https://medium.com/p/a2f4f14b6916</guid>
            <category><![CDATA[fashiontech]]></category>
            <category><![CDATA[ecomerce]]></category>
            <category><![CDATA[virtual-try-on]]></category>
            <category><![CDATA[shopify]]></category>
            <dc:creator><![CDATA[Tuck Virtual Try On]]></dc:creator>
            <pubDate>Thu, 17 Sep 2026 05:33:54 GMT</pubDate>
            <atom:updated>2026-09-17T05:33:54.919Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Shoppers arrive at a product page with two separate uncertainties. Virtual try-on resolves the first one well. The second one needs its own answer.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pK1IxMO1dE5WPDVYBg9gHA.png" /></figure><p>A physical store solves something that ecommerce has never fully replaced, and it does it so quietly that most people never notice.</p><p>You take three things into a fitting room. You come out having answered two separate questions: whether this looks like you, and whether this is the right size. Both answers arrive within a couple of minutes, in the same room, and nobody experiences them as distinct. You just know whether to buy it.</p><p>Online, those two questions came apart and landed in different places.</p><p>The appearance question is handled by photography — a garment on a model, lit well, styled, shot on someone who is almost certainly not shaped like the shopper. The sizing question is handled by a table of measurements behind a link, often produced by a different team, sometimes inherited from a supplier, occasionally not updated when the pattern changed.</p><p>The shopper is handed both and quietly asked to reconcile them.</p><p>Most can’t. Some handle it by ordering two sizes, which solves the shopper’s problem and creates the brand’s.</p><h3>Two uncertainties, not one</h3><p>It’s worth being precise about what a shopper is actually uncertain about, because the two things get collapsed constantly.</p><p>The first uncertainty is visual. <em>How will this look on me?</em> Colour against skin tone, whether the proportion works, whether the silhouette reads the way the shopper wants to be read. This is subjective and immediate, and it’s resolved by seeing.</p><p>The second uncertainty is about sizing. <em>Which size should I order?</em> This is a question about a specific body against a specific brand’s sizing — not sizing in general, but this brand’s, which may run differently from the one they bought last month.</p><p>These feel adjacent because they arrive seconds apart. But they’re answered by different information, and the answer to one doesn’t produce the answer to the other.</p><h3>What virtual try-on actually solves</h3><p>Virtual try-on addresses the first uncertainty directly. A shopper provides their own photo, and sees the garment visualized on themselves rather than on a fit model.</p><p>That’s meaningful. It shortens the distance between a product image and the shopper imagining ownership. It answers <em>how this look on me</em> in a way a look-book can’t.</p><p>What it doesn’t do is tell the shopper what size to order.</p><p>This is the part the industry gets loose about, and it’s worth stating plainly: a generated image showing a garment on a shopper does not establish which size that shopper should buy. It’s a visualization. It builds visual confidence. It’s not a measurement, and it shouldn’t be presented as one.</p><p>That isn’t a shortcoming of try-on. It’s a description of the job try-on is doing. The problem only appears when the industry treats one solved question as though it settled both.</p><h3>Sizing is a different kind of problem</h3><p>Answering <em>which size I should order</em> requires different inputs entirely.</p><p>It needs information about the shopper — height, weight, age, how they prefer clothes to sit on them. And it needs the merchant’s own sizing information, because a size chart is brand-specific and no general assumption substitutes for it.</p><p>Put those together and you can produce a size suggestion. That’s guidance built from numbers, and it belongs to a different discipline than image generation.</p><p>Which is why collapsing the two into a single feature is a mistake. Not a marketing mistake — an architectural one. They draw on different inputs, they can fail in different ways, and a shopper needs both answers before they’re comfortable clicking buy.</p><h3>The button is only the surface</h3><p>There’s a second thing worth understanding about this category, and it applies to whether or not sizing is in the picture.</p><p>A try-on button is the visible layer of an experience. It’s the part a merchant installs, themes, and positions on a product page. Underneath it sits everything that actually determines the outcome: how the garment is represented, how the shopper is represented, how the output is generated, how that data is handled, and what other decision-support capability sits alongside it.</p><p>Two stores can have identical buttons and materially different experiences underneath.</p><p>This isn’t an argument that interface layers are worthless. Integration work is real work, and a badly implemented widget can sink a good system. The point is narrower: the presence of a try-on button tells you almost nothing about what the shopper’s experience will be, and evaluating the category on that basis leads merchants to compare the wrong things.</p><p>The more useful questions are about what’s underneath. Who controls the generation. What happens with difficult garments rather than easy ones. Whether the same technology can support anything beyond a single product-page widget. And whether the experience addresses the sizing question at all, or leaves it where it’s always been.</p><h3>Where Tuck sits</h3><p>Tuck is built around both halves of that decision, as two distinct capabilities rather than one blended feature.</p><p><strong>Virtual try-on.</strong> The shopper provides a photo. Tuck uses that image together with the garment image to visualize the shopper wearing the product. It’s apparel-focused, which matters, because apparel has structure and behavior that generic visual approaches aren’t built around — how a garment falls, where the shoulder sits, how a hem lands. The purpose here is visual confidence.</p><p><strong>Fit Intelligence, or size suggestion.</strong> This is separate. It takes shopper information — height, weight, age, fit preference — and compares it against the merchant’s own size chart to suggest a size. The purpose here is sizing confidence.</p><p>The distinction is deliberate and we keep it visible. The try-on image doesn’t determine the size. The size suggestion doesn’t come from looking at the picture. They’re two answers to two questions, delivered in the same place.</p><p>Neither is a guarantee. A suggestion is a suggestion, and bodies are more varied than any chart can capture. What changes is that a shopper isn’t left running two disconnected errands — study an image here, open a measurement table there, and reconcile the two alone.</p><h3>Why the combination is the point</h3><p>Consider how the hesitation actually unfolds.</p><p>A shopper likes a garment, but can’t picture it on themselves. Try-on addresses that, and they move forward.</p><p>Then they reach the size selector, and a second uncertainty arrives that the image did nothing to touch. If nothing helps them there, the same hesitation that existed before try-on is still sitting between them and the purchase. It just moved four inches down the page.</p><p>Solving one uncertainty well while ignoring the other means the shopper gets farther along and stops anyway. Seeing it builds one kind of confidence. Knowing what to order builds another. Purchase confidence tends to need both.</p><p>That’s the case for treating these as related but separate capabilities, rather than one feature with a single name.</p><h3>The shift worth paying attention to</h3><p>For merchants, the practical version of this is less about technology and more about where it sits in the journey. Try-on and size guidance are useful at the specific moment a shopper is deciding — on the product page, in the size-selection step, at the point the decision is actually made. Capability that lives somewhere else in the flow doesn’t do much.</p><p>But the broader shift is in what the industry is asking of itself.</p><p>The question has been: <em>can we show shoppers an AI try-on?</em> That’s a capability question, and the answer is increasingly yes for everybody. Which means it stops being interesting.</p><p>The better question is: <em>can we reduce the uncertainty involved in buying apparel online?</em> That’s a decision question, and it’s harder, because uncertainty comes from more than one source and a single impressive feature doesn’t address all of it.</p><p>Tuck is built around that framing — apparel visual try-on and sizing intelligence, treated as related but distinct parts of the same decision.</p><p>A fitting room was never a mirror. It was a place where two questions got answered at once. Rebuilding only the mirror was always going to leave something behind.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a2f4f14b6916" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>