Start a virtual try-on pilot with a coherent, purchasable assortment that answers one customer question. Choose products shoppers genuinely need to compare, keep enough stock behind those choices, and include the range of styles you would need to support after rollout. A shortlist selected only because it looks impressive in a demonstration cannot establish whether the experience helps sell your collection.
For a fashion founder, merchandising director or ecommerce leader, the first catalog decision is therefore commercial: which part of the range has a discovery problem worth solving, and what would a better buying journey look like? This guide offers a selection framework for evaluating iKawn Mirror. It does not prescribe a universal SKU count or claim that a particular garment category will perform best.
Why assortment scope belongs in the budget conversation
There is pressure to invest in AI, but interest in technology is not proof of demand for your assortment. DHL’s 2026 E-Commerce Trends Report announcement, published June 2, 2026, reports that 73% of surveyed businesses anticipate using generative AI more over the next five years. The research covered 29,000 shoppers and 5,800 businesses across 29 countries. This is a broad ecommerce investment signal, not a fashion virtual try-on conversion benchmark.
Meanwhile, the BoF–McKinsey State of Fashion 2026, published November 17, 2025, projects low single-digit global fashion growth in 2026 and describes consumers seeking value. That context makes a focused budget request easier to scrutinize: identify the buying friction, show which products it affects, and test whether better visual discovery changes a useful next action.
Consider a store that has a wide collection but little display space, or an ecommerce range whose similar product photographs make styles hard to distinguish. Those are candidate problems to investigate. If the real obstacle is an unavailable size, an unattractive price or a delivery promise customers reject, adding a try-on experience does not resolve that obstacle. Review search behavior, product-page exits, staff observations and stock availability before deciding which category deserves the pilot budget.
Define the buying decision before counting products
Write the pilot question in language a merchandiser can use: “Can shoppers compare the occasionwear silhouettes we carry and choose which to examine physically?” or “Can customers discover credible alternatives to the one style they arrived to buy?” These are proposed test questions, not reported outcomes.
Choose one channel and one next action. For a store, that might be asking an associate for the selected garment. For ecommerce, it might be visiting the correct product page and choosing an available variant. Keep the selection relevant to that action; combining unrelated garments, channels and campaigns makes the result harder to interpret.
Also agree what “product” means in the scope document. A style, a color variant and a size SKU are different units. A quote covering styles can describe a very different workload from one covering every sellable variant. Ask the provider and catalog owner to record those units explicitly, including which changes require review.
Use five gates to choose the first assortment
The following is an editorial decision framework, not a research-derived scoring model. Apply each gate to a proposed category and then to the individual styles. A commercially important failure should not disappear inside an average score.
- There is a specific discovery need. Identify where shoppers hesitate between looks, overlook alternatives or need assisted comparison. Use your own evidence to distinguish a visualization problem from fit, price or fulfillment friction.
- The assortment offers meaningful choices. Include variations in silhouette, color, detail or occasion that reflect the actual range. A selection of near-identical bestsellers may produce activity without testing whether shoppers discover something useful.
- The products are available to buy. Check the sizes and variants customers are likely to request, the relevant store or delivery region, and expected availability during the test. Decide who removes a sold-out option and how quickly. Confirm the proposed availability workflow; do not assume a real-time stock integration is included.
- The collection can pass a brand review. Ask merchandising to evaluate representative items in the intended experience before customers see them. Keep unsuitable items out of the customer pilot and record the exclusions. Product representation is an acceptance gate here, not an image-production tutorial.
- The pilot can produce a usable decision. Choose an assortment with enough eligible shopper exposure to evaluate the agreed action and a named owner who can maintain it. If exposure is too low, narrow the question or extend the observation period before making a rollout claim.
Assign a merchandising owner to the shortlist, an operations or ecommerce owner to availability and handoff, and an analytics owner to the readout. Before commissioning extra work, ask for the initial catalog cost, the process and cost of replacements, any usage allowance, and the ongoing review burden. A narrow launch should also reveal what maintaining a larger range would require.
A public example of a deliberately bounded assortment
Zalando’s April 24, 2023 virtual fitting room announcement described a pilot with a selected range of 22 jeans items. Customers created an avatar and saw size-related heatmaps. It is a historical example of a retailer defining a category and a bounded set of products for a specific learning objective.
That number is not a recommended iKawn catalog size. Zalando’s avatar and fit-oriented workflow is different from evaluating Mirror for visual discovery and preselection. The useful inference is to match the assortment to the question being tested; the announcement does not prove that 22 items, jeans or that workflow is right for your business.
Balance commercial relevance with honest coverage
Use three roles when assembling the shortlist. Core items should represent normal demand and the styles customers already consider. Discovery items should be plausible alternatives that deserve more consideration. Boundary items should expose the difficult but commercially important cases your eventual range would contain.
Boundary items belong in an internal acceptance review first. If they fail, exclude them from the customer experience and say what the pilot no longer covers. A successful test of simple garments cannot validate an entire collection containing materially different silhouettes, layers or decorative details.
For example, an occasionwear team might propose a capsule containing familiar styles, overlooked alternatives and several detail-rich pieces for internal review. This is an illustrative planning scenario, not an iKawn deployment or a promise of category support. The exact items should be agreed after a demonstration using representative products.
A full-catalog launch offers wider coverage but adds more availability checks, approval work and opportunities for the assortment to change during measurement. A tiny selection is easier to operate but may leave too little meaningful choice. Start with the smallest set that represents the intended buying decision, then document the categories and variants it excludes.
Keep the assortment stable enough to learn
Before launch, save the approved product list, variant mapping, selling prices, availability rules and date of approval. Log additions, removals, promotions and stockouts during the pilot. If the featured products or their prices change materially, separate those periods in the readout. Otherwise, a stronger promotion could be mistaken for an effect of try-on.
Track the journey from eligible exposure to starting a session, completing a useful comparison and taking the agreed next action. Report these measures by category or assortment role where sample sizes permit. For a store, distinguish a garment request from a completed sale. Online, distinguish a product-page visit from an order. An engaging session alone does not establish commercial value.
When the question is incremental sales, use a planned comparison with similar eligible shoppers or comparable store periods, and account for availability, promotions and assisted service. Comparing enthusiastic try-on users directly with all non-users can mislead because the groups may already differ in purchase intent. If the sample cannot support a reliable sales conclusion, report operational feasibility and observed behavior, then decide what further test is justified.
Set the expansion rule before reviewing results: the next category must have a viable buying journey, acceptable representation, adequate stock coverage and an owner. Use the Mirror ROI calculator to explore assumptions, but replace assumptions with pilot evidence before committing to a wider rollout.
Resolve the objections before approving the shortlist
“Can we use only bestsellers?” They can provide a useful reference, but a bestseller-only pilot does not answer whether Mirror helps customers discover the wider range. Include the alternatives that matter to your commercial question.
“Should clearance stock get priority?” Only if it still meets the buying-journey gates. Sparse size availability and changing markdowns can make clearance a poor first test. Treat clearance as its own question if the purpose is to evaluate that specific journey.
“Do we need the whole catalog to get a fair result?” No universal count makes a result fair. Represent the decision you intend to support, define the exclusions and resist applying the result to untested categories. More products do not compensate for insufficient shopper exposure.
“Can visual try-on settle size and comfort questions?” Do not scope a discovery pilot around that promise. Physical fit, fabric feel and final sizing judgment remain separate evaluation needs. Make the intended role clear in both the shopper journey and the internal business case.
Bring a commercial shortlist to an iKawn Mirror demo
iKawn Mirror supports evaluation of live and photo-based virtual try-on for fashion discovery across ecommerce, stores and activations. Start the conversation with your priority channel, the choice shoppers struggle with, representative styles, availability constraints and the action you want them to take next. Confirm product suitability, catalog maintenance, integrations and measurement scope for the proposed deployment.
The useful outcome of the demonstration is an agreed first assortment and a clear list of exclusions—not an assumed uplift. If you are planning an assisted store pilot, pair the merchandising decision with the store staff training and handoff plan.