AI Frame · Original research

The future of fashion ecommerce, understood through data.

Research, real merchant experiences and practical insights into how technology is changing the online fashion shopping experience.

Explore our research
21,095size recommendations shown by one international fashion retailer in nine weeks.AI Frame merchant data
10%fewer size-related returns on items with size advice.Zalando
See every figure and its source
Research areas

What we're investigating.

We explore the questions shaping the online fashion shopping experience, combining merchant data, technology and practical ecommerce insights.

59%of shoppers were unhappy with an item because it looked different on them.Google and Ipsos
01

Virtual Try-On

How shoppers use visual product previews when exploring clothing online, and what the industry has measured about items that look different on the shopper.

Read the industry report
10%fewer size-related returns on items with size advice.Zalando
02

Sizing & Fit

How shoppers make sizing decisions and interact with personalised sizing guidance.

See the fit research
7 in 10buyers ordered the recommended size or one size up, using height-based sizing on the product page.AI Frame case study
03

Product Page Experience

What makes fashion product pages more useful, informative and engaging.

See the streetwear case study
34%of store orders came from shoppers who had been shown a size recommendation first.AI Frame case study
04

Merchant Performance

What real store data can tell us about shopper behaviour and ecommerce performance.

Read the case study
Featured research

Why fit is the biggest gap in online fashion.

Now published · July 2026

Why most fashion returns are about fit, and how to cut them.

An industry report on the confidence gap in fashion ecommerce: what shoppers cannot see before they pay, and what it costs.

It draws on published figures from the National Retail Federation, McKinsey, Coresight and others, and compares the technology built to close the gap: virtual try-on, size recommendation and fit prediction.

Read the report
Inside the report
  • The problem, quantified
  • The technology landscape
  • The evidence so far
  • The honest limitations
Fit research

What retailers have measured about fit and returns.

Zalando, Levi's, Google and others have published their own numbers. We put them in one place, with what each one measured and where it comes from. None of them are AI Frame clients.

  • 10%fewer size-related returns on items with size advice.Zalando
  • Up to 40%fewer returns on jeans in virtual fitting room pilots.Zalando with Levi's
  • 19.3%of US online sales were expected to be returned in 2025.NRF and Happy Returns
  • 59%of shoppers were unhappy with an item because it looked different on them.Google and Ipsos

Each figure on the research page says what was measured and where it comes from.

Explore the fit research
Merchant case studies

Real merchants. Real experiences. Measurable findings.

Merchant Case Study · Size Recommendations

21,095 size recommendations in nine weeks.

How an international fashion retailer introduced personalised sizing across 37 products and reached 4,917 shoppers in nine weeks.

Read the case study
21,095size recommendations shown
4,917shoppers reached
37products with recommendations
34%of orders came from shoppers shown a size first

This is exposure, not cause: the 34% does not show that a recommendation led to the purchase. Measured September to November 2026.

~380recommendations matched to a purchase
7 in 10ordered the size we recommended or one size up
82buyers saved a Fit Profile

AI Frame analytics on one hoodie, over about four weeks.

Merchant Case Study · Size from height alone

A US streetwear brand, sized by height.

A roomy unisex hoodie made chest measurements useless, so the brand published a height range per size. Shoppers get a size without a tape measure.

See what they learned
Coming next

What we are working on next.

  • The Fashion Ecommerce Product Page Report

    Research in development
  • What makes a good online size chart?

    In preparation
  • The role of visualisation in online fashion shopping

    In preparation
  • What fashion merchants should measure on product pages

    In preparation
  • How personalised size recommendations work

    In preparation
  • Understanding shopper behaviour on fashion product pages

    In preparation
Methodology and transparency

Research should be useful. And credible.

Our research combines aggregated merchant data, product usage insights and practical ecommerce analysis. We aim to explain not only what we find, but how we find it.

  • Real data

    Findings are based on actual merchant activity and available performance data.

  • Transparent methodology

    We explain the sample, measurement period and limitations.

  • Responsible reporting

    Merchant identities and customer information are protected, and we distinguish correlation from causation.

Industry figuresPublished by the retailer or body that measured them, with the source named and a note on what was measured. None come from AI Frame clients.
AI Frame dataAggregated merchant analytics with the period stated. Exposure is never presented as cause: a recommendation shown before an order is not a recommendation that caused it.

Better fashion shopping starts with better information.

Explore how AI Frame helps fashion merchants bring Virtual Try-On, personalised Size Recommendations and interactive Size Charts into their Shopify product pages.