Size Chart, Recommendation, or Size Comparison?

In fashion e-commerce, “what size should I choose?” sounds like a single question. In practice, it brings together three different needs: checking measurements, receiving guidance, and visually understanding the difference between alternatives. When a store calls everything a “size guide,” the interface hides what each resource actually does.

The clearest architecture separates the measurement chart, the size recommendation, and the visual size comparison. They can appear in the same journey, but they should not share a generic promise to “always get it right.”

The measurement chart answers objective questions: which measurements were recorded for each size? It turns a scattered grid into a visible, comparable reference. It is the simplest foundation of the journey, but it needs to be specific to the product or the correct pattern family.

A generic brand chart can fail when T-shirts, pants, dresses, and oversized garments follow different logic. The record needs to preserve the code, category, size range, and unit of measurement. When the pattern varies, the interface should explain which chart is being used.

When to use it

• • When the person wants to check measurements and compare their own reference.
• • When the brand needs to publish a consistent source for each product or pattern.
• • When recommendations or comparisons depend on a correctly recorded size range.

What it does not solve on its own

The chart does not interpret fit preferences, does not automatically know the person’s body, and does nottranslates every centimeter into a visual experience. It provides the data for a manual decision or for other modules in the journey.

The recommendation uses the data requested by the store and the registered sizing model to suggest a reference size. This recommended size serves as the journey’s natural anchor: it is the starting point, not a mathematical guarantee.

The flow needs to explain which information is requested, why it is necessary, and how the suggestion should be interpreted. If there is a preference for a closer or looser fit, it should appear as part of the experience—not as an invisible algorithmic detail.

When to use

• • When the store wants to guide the choice without requiring a manual review of the entire size range.
• • When product, fit model, and sizes are entered consistently.
• • When the brand can explain limitations and offer an alternative in case of uncertainty.

What it does not solve on its own

A recommendation does not necessarily show how two sizes change the fit. It also does not correct an inconsistent size range, an incorrect product reference, or an incorrectly identified fit model. Data quality remains part of the product.

The comparison tool answers a different question: “if I change sizes, what might happen to the fit?” At Provei.AI, the experience begins with the recommended size as the natural anchor and derives an alternative that is closer or looser according to the size range and configured flow.

This order matters. If the interface presents two sizes without explaining which is the recommendation and which is the alternative, the image creates more uncertainty. Labels need to accompany each result, the order must be stable, and the person must be able to return to the product without losing their choice.

When to use

• • When the uncertainty does not end with the recommendation and involves a fit preference.
• • When the store wants to compare the recommended anchor with an adjacent alternative.
• • When sufficient visual references and sizing data exist to build the experience.

What it does not solve on its own

The visualization is an AI-generated approximation. It does not measure the body from the image, does not replace the size chart, and does not guarantee that the perception of fit will be identical across all poses, fabrics, and bodies. The product details and recommendation remain visible.

The most coherent order is usually: identified product, available chart, necessary data, recommended size, and optional comparison. Not every store needs to show every step in the first interaction. The journey can reveal more depth as the question arises.

See the complete view of the Sizes family to understand how the three modules connect without taking over their owning pages.

Before the interface, the store needs to answer catalog and ownership questions. Who updates the size range? How does the product point to the correct chart? What happens when a variation has no data? What text appears when the recommendation cannot be calculated or when the comparison is unavailable?

If the foundation is not ready yet, start with the guide on how to create a size chart for e-commerce. Publishing a sophisticated interface on top of inconsistent data only makes the problem more difficult to diagnose.

The size component usually appears in a small space on the product page. Therefore, hierarchy and state matter more than the number of controls. The recommended size needs to be legible, the reason for the suggestion must be accessible, and the visual alternative cannot look like a second recommendation.

Keyboard, screen readers, and mobile devices need to be part of QA. Use real buttons, associate labels with fields, announce processing and errors, preserve focus after generation, and do not use only color to differentiate sizes. In the comparator, each image needs a text title and predictable order.

Also design the imperfect states: product without a chart, sold-out size, incomplete size range, unavailable recommendation, interrupted generation, and a result that needs to be redone. The safe path is to offer ainformation available and keep the next step moving — never fill the gap with an invented certainty.

Choose a product family with a stable size range and reviewed references. Before expanding, go through the journey on different devices, with available sizes and exception states. The brand should visually review the comparisons and confirm that the recommendation language reflects its policy.

Measure connected events: opening the size chart, starting and completing the recommendation, suggested size, requested comparison, selected alternative, return to the product page, and commercial progress. Do not turn click counts into proof of quality. The signal only makes sense when it preserves the product, context, and stage of the journey.

For fashion, the solution may also include Fashion Virtual Try-On and Provei Studio. They address visualization and content creation; they do not replace the size family. See the complete architecture in solutions for fashion and apparel.