How AI Can Leverage a Body Data Foundation to Power Apparel Fit

One of the areas where Artificial Intelligence (AI) and Machine Learning (ML) can create highly accurate results in the apparel industry is in fit and sizing. By analyzing large datasets of body measurements and using advanced algorithms, AI can help apparel companies create more accurate and personalized sizing recommendations for their customers at a scale that would, under normal circumstances and with just human action, be unachievable. The first step in achieving accurate sizing through AI is to build a foundation of body data to power machine learning models predictions.

For example, Bold Metrics is an AI sizing solution provider that uses advanced algorithms and data analytics for apparel brands to accurately determine the sizing and fit of clothing and apparel for individual consumers based on their unique body measurements. Bold Metrics' AI solutions sees measurement accuracy to within +/-1% of a Master Tailor’s in-person measurements. To achieve this level of accuracy, Bold Metrics focuses on 5 key areas.

1. Extensive Training Data

For AI outputs, the quality and quantity of the foundational data upon which the ML algorithms are trained will impact the results' accuracy. AI algorithms require large amounts of high-quality training data to make accurate predictions. Bold Metrics has spent over eight years growing its AI Body Scan database and generated over 67 million AI body scans. Its proprietary ML algorithms are trained on over 2 million new body models a quarter and, with a simple survey of four to six questions, can accurately determine over 50 individual customer body measurements in seconds. 

2. Machine Learning Capabilities

Bold Metrics' algorithms use machine learning to improve their accuracy over time. As more data is collected and analyzed, the algorithms become more refined and accurate, making better predictions for individual customers. For example, Men’s Wearhouse online tuxedo rentals leverage Bold Metrics’ Virtual Sizer solution, resulting in an average return rate decrease of 47.4 percent for online orders of tuxedo coats, pants, and shirts due to customers getting best-size recommendations based on their individual body measurements.  

With over eight years of data training on millions of high-fidelity AI Body Scans, Bold Metrics’ algorithms can accurately determine results at speed and scale. Added to this is the layering in of additional data points, such as purchase and returns customer feedback, to update models and optimize predictions with real-time consumer data. 

3. Multivariate Analysis

Bold Metrics' algorithms use a range of simple data inputs to determine customer body measurements, including age, weight, and height. The algorithms take a multivariate approach, simultaneously considering all of these factors to make more accurate determinations while reducing the opportunity for bias or inaccuracy with a simple, straightforward survey. 

4. Granular Results on a per Garment/ Style-By-Style Basis

Fit is personal - how someone likes their clothes to fit varies from individual to individual. Bold Metrics can provide contextualized fit results because its AI takes into account not just individual customer body measurements but also layers in garment specifications and tech pack data on a per-style basis to provide specific best-size recommendations on a granular basis. By layering elements like fit preferences and geographical differences, Bold Metrics’ ML-based algorithms can recognize these patterns, trends, and data. The recommendations become smarter and more accurate over time. 

5. Close the Feedback Loop

Bold Metrics’ AI solutions surface a shopper’s unique body measurements to personalize the retail experience for every customer. When a shopper utilizes the Smart Size Chart, purchase, and returns data is tracked to allow for a continuous feedback loop, enabling a deeper understanding of how garments are fitting customers and how purchases and returns are impacted by sizing. A recent partnership with Loop, the exchange-first returns platform for Shopify brands, enabled a scalable way to provide contextualized fit returns data by garment style. This actionable insight allows for operational optimizations, elevating the online retail experience while reducing fit-related returns for brands.

Bold Metrics' AI-powered approach to fit and sizing provides customers with better-fitting clothing based on their body measurements and fit preferences. It generates valuable brand insights to optimize technical design, distribution, and operations.

Creating a personalized best-fit recommendation with accuracy is crucial for a company’s bottom line and improving customer satisfaction. A survey from Coresight Research found that 42.4% of consumers in the U.S. returned unwanted products from March 2020 to March 2021, most of which were clothing. The lack of certainty around sizing when it comes to apparel creates low conversion rates and higher returns when customers indulge in bracketing behavior when they purchase multiple items in the same style to try on at home, only to return those that don’t fit.

By building and analyzing vast amounts of customer data over time and constantly fine-tuning the feedback process, Bold Metrics' algorithms make highly accurate predictions for individual customers at a scale and level of personalization made possible only through AI. As our proprietary algorithms continue to learn from new data and incorporate customer feedback, the accuracy of Bold Metrics' sizing recommendations will continuously improve, resulting in a more efficient and rewarding shopping experience for shoppers that have been proven to lead to higher average order values. For brands, this means increased conversion that stays high and reduced returns to help improve sustainability efforts while giving customers a seamless, helpful, and accurate way to find their best fit while shopping online.