Why Inaccurate Size Recommendation Tools Are Making the Fit Problem Worse

I rarely get this angry, but after observing the new size recommendation tools that are emerging on the market, I’ve reached my breaking point. Some of these tools are already backed by investors and being used by major brands. But I have a serious issue with the direction they’re taking.

Dear developers, AI researchers, and anyone without patternmaking experience, do you realize the disaster you’re causing by providing inaccurate size recommendation tools?

Do you honestly think the fit problem can be solved by someone who has never made a single pattern for clothing? I’m not here to throw shade at technology; I’m here to highlight a major flaw that could ultimately harm customers and fashion brands alike.

Let’s break it down. I recently looked at six different size recommendation tools. Every single one of them asked for basic measurements like bust, waist, and hips (some even just asked for weight and height), only to recommend the same size for every product I opened.

How is this possible?

What is the logic behind recommending the same size for an entire product category or, even worse, the same size across all categories?

This is where the problem lies, and here’s why:

1. Size Inconsistency: Do you understand that size inconsistency exists among brands and even within the same brand? One brand’s size medium might not fit the same as another’s.
2. Elasticity and Design: Do you realize how critical elasticity is when defining the size of a garment? Stretch fabrics behave differently than non-stretch fabrics, and fitted designs fit differently than loose ones. These tools are missing that nuance, leading to wrong recommendations.
3. Misleading Customers: By giving the same size recommendation for every product, you are misleading customers in their search for the right fit. This isn’t just a disservice to customers, it’s costing fashion brands sales and increasing return rates.

For every tool I tested, it was the same result. No matter the design, elasticity, or whether the products came from different brands, I was always recommended the same size. Whether it was an elastic dress, a fitted blazer, or a loose-fit jacket – it didn’t matter.

This is exactly the kind of error that fuels the growing frustration with online shopping. Customers receive their orders, try them on, and send them back because the fit is wrong.

I understand the urgency in creating a size recommendation solution to one of the fashion industry’s biggest problems: fit. The problem isn’t just a customer inconvenience; it’s a real issue that’s affecting the bottom line of fashion brands.

But if you think AI or machine learning alone can solve this, think again. Technology is powerful, but it can’t replace the craftsmanship of a seasoned patternmaker. Craftsmanship should be embedded in your code. You can’t just throw a bunch of data into an algorithm and expect it to know how to design for the intricacies of human body shapes and fabric behavior.

If you don’t understand how to incorporate that craftsmanship into your solution, please don’t release your code. It’s like me starting to use Copilot and calling myself a developer who can write real code. The end result is more harm than good.

The road to solving the fit problem is long and requires more than just a few simple metrics and data points. It requires a deeper understanding of clothing design, fabric properties, body types, and sizing consistency across different brands.

AI can help, but it needs to be grounded in the real-world expertise that only comes from patternmaking and fashion design. Until that happens, these tools are not the solution we need – they’re just another band-aid on an ever-growing wound.

So, to the developers and AI experts out there: before you go live with another “AI-powered” tool, consider this: if you don’t know how to incorporate the necessary craftsmanship into your code, maybe it’s time to partner with someone who does. The future of fashion fit depends on it.