Inconsistent Sizing in Fashion: Analysing 1 Million Online Returns
We analysed over 1 million online returns and uncovered a striking reality: inconsistent sizing in fashion is the number one driver of refunds in fashion retail. This issue spans across brands, product categories, and even individual garments of the same style. For retailers, the cost is immense – lost revenue, reduced customer trust, and an ongoing battle to improve conversion rates.
In this deep dive, we explore why inconsistent sizing in fashion remains such a major problem, how traditional sizing methods are failing customers, and how AI-driven fit finders are changing the game.
The Scale of the Sizing Inconsistency Problem
Recent research shows that 66% of UK consumers cited poor fit as the reason for returning online purchases (Statista, 2020).. Our analysis of 1 million returns reveals that a significant portion of these returns stem from inconsistent sizing in fashion between brands, and even within the same brand across different categories and collections. For example:
- Nike: A ‘Medium’ for men covers a chest size of 38-41 inches.(96.52 x 104.14 cm)
- Zara: A ‘Medium’ is designed for chest sizes of 39-40.5 inches.(99.06 x 102.87 cm)
- Calvin Klein: A ‘Medium’ fits a 37-38 inch chest.(93.98 x 96.52 cm)
- Gant: A ‘Medium’ ranges between 38-40 inches.(96.52 x 101.6 cm)
- H&M: A ‘Medium’ is intended for 37-39 inches. (93.98 x 99.06 cm)
The variation is clear: a shopper with a 38-inch chest could be a ‘Small’ in one brand and a ‘Medium’ in another. This inconsistent sizing in fashion forces customers to either guess their size or order multiple sizes, leading to higher return rates.
Inconsistencies Within the Same Brand
H&M has publicly acknowledged the need to improve consistency in its sizing. In September 2023, Chief Executive Helena Helmersson addressed customer concerns, stating: “There’s always improvement to make.” She emphasized that the retailer is actively working to ensure that “whatever customers buy, they want to keep,” including enhancements to sizing guides. However, even with these efforts, the same size garment may still vary depending on fit (H&M Product Page). This means that two ‘Medium’ T-shirts from H&M could have different measurements if one is a slim fit and the other is a relaxed fit.
This issue extends beyond H&M:
- Zara’s sizes fluctuate based on different collections, making it difficult for loyal customers to shop with confidence.
- Nike’s sportswear and casualwear fit differently, with some ranges running smaller due to performance-focused designs.
- Calvin Klein’s denim and formalwear use different sizing structures, meaning a shopper might need different sizes across their product categories.
Adding another layer of complexity, colour variations within the same product can impact fit. Some retailers use different fabric suppliers for different colours, leading to slight differences in stretch and structure. This is a commonly reported issue, with customers noticing that a ‘Black’ version of a T-shirt might fit tighter than the ‘White’ version. A BBC investigation revealed that H&M customers experienced drastic differences in fit between trousers of the same style but in different colours, despite having identical size labels (BBC).
One shopper, Lesley Hodgson, discovered that her black trousers fit perfectly, while the beige version in the same size was far too tight, highlighting inconsistencies in manufacturing and fabric treatment. The report also noted that many customers feel they are being unfairly charged for returns caused by these inconsistencies, especially under H&M’s new returns policy.
A report from The Jerusalem Post further supports this issue, explaining that darker fabrics often undergo additional processing, making them slightly less flexible than their lighter counterparts (Jerusalem Post).
The Shortcomings of Traditional Fit Finders and the Shift to AI
Some fit finder solutions attempt to solve this issue by asking customers which size they wear in a known brand and then suggesting an equivalent size in a new brand. However, this approach relies on outdated statistical methods that fail to address real-world sizing inconsistencies. These tools assume that past sizing choices are accurate predictors of future purchases, ignoring the significant variations between brands, product categories, and even colour-specific fit differences.
This static, rule-based system is fundamentally flawed because:
- Sizing varies significantly between brands, as demonstrated above.
- Even within a brand, different categories and styles fit differently.
- Customer perception plays a major role. If a shopper is used to buying from a brand that runs large, they may automatically size down when shopping elsewhere, further complicating recommendations.
Instead of relying on outdated statistical models, retailers need predictive AI-driven solutions that dynamically analyse real-world purchase and return data, garment-specific dimensions, and individual customer body profiles. Predictive AI continuously learns from new inputs, improving accuracy over time, unlike traditional fit finders that rely on static brand-to-brand comparisons.
Retailers that stick to these older methods risk alienating customers with incorrect recommendations, driving up return rates, and damaging brand trust. The solution is a data-driven, machine learning approach that understands fit at the deepest leveldown to fabric stretch, pattern grading, and customer shape preferences.
Case Study: How Playful Promises Reduced Returns with AI-Powered Fit Prediction
Playful Promises, a UK-based lingerie retailer, implemented AI-powered size prediction after struggling with high return rates due to inconsistent sizing. The retailer had relied on standard size charts and customer-reported size preferences but found that return rates remained high, particularly among new customers.
By integrating Prime AI’s fit recommendation system, Playful Promises achieved:
- A 27% reduction in returns, particularly in fitted lingerie items such as bras and bodysuits.
- An 18% increase in conversion rates, as customers felt more confident selecting the right size, reducing abandoned carts.
- A significant decrease in customer service inquiries related to sizing, reducing the workload on support teams.
The AI-driven system leveraged real purchase data, biometric analysis, and even accounted for variations between different colours of the same product. Unlike traditional size charts, the Prime AI system continuously improved its accuracy with each new purchase and return, refining its predictions to deliver an optimal fit recommendation for each individual shopper.
The Features and Benefits of AI Fit Prediction
AI-powered fit finders offer far more than just basic size recommendations. They provide a truly personalised shopping experience that addresses individual customer needs. The key features of predictive AI-driven sizing solutions include:
- Granular product dimension analysis – AI measures garments at a deep level, ensuring accuracy beyond standard size charts.
- Customer biometrics and past purchase data – Predictive AI learns from previous returns and successful purchases to refine recommendations.
- Real-time machine learning improvements – The AI system continuously adapts to new data, making predictions increasingly precise.
- Fabric and colour variation awareness – Unlike traditional tools, AI understands that different fabrics and colours affect fit and adjusts recommendations accordingly.
- Seamless integration with e-commerce platforms – AI fit finders can integrate with Shopify, Magento, and custom-built platforms to deliver an effortless experience for customers.
Retailers using predictive AI-powered fit finders see a significant reduction in returns, higher customer satisfaction, and improved sustainability by minimizing unnecessary shipments and returns processing.
Take the Next Step: Reduce Returns and Boost Sales with Prime AI
Sizing inconsistencies are a major cause of lost revenue and customer dissatisfaction in fashion retail. By implementing AI-driven fit prediction, brands can significantly reduce return rates, improve customer satisfaction, and enhance sustainability.
Want to see how Prime AI’s Fit Finder can work for your brand? Don’t let outdated sizing models cost you sales. Book a free demo today and see how Prime AI can help your brand deliver a seamless shopping experience, reduce returns, and build long-term customer trust.
Speak to us today
More from the blog
