The Evolution of our Magic Fit® technology The Evolution of our Magic Fit® technology

The Evolution of our Magic Fit® technology

Online shopping is always a risk, but especially when it comes to clothing. Which is why we’re transforming the experience of shopping for clothes at Citizen Wolf to make it better for the environment and better for you. 

Sizes are hardly standard and guessing the right one is always a gamble; sometimes it fits but mostly it doesn’t. 

We realised that the perfect shopping experience allows you to pick the styles, colours and fabrics you love without worrying about the vagaries of vanity sizing, which is exactly why we created Magic Fit® over 18 months ago.


So what's changed?

Not much on the surface: you still choose your favourite style and our job is to make sure it fits perfectly. Under the bonnet it's a different story.

Magic Fit® is now the love child of statistics, natural language processing (NLP), our experience and machine learning. 

Each of these layers plays a vital role in creating your perfect fit and as the saying goes, "a perfect Tee in time saves nine (Tees going to landfill)". After more than 10,000 Tees made using the algorithm, we're now able to predict measurements which were previously unavailable through the statistical model alone.

Custom fit to your body with magic mathematics.


Numbers don't lie.

The foundational layer of Magic Fit® is (not so) simple statistics. To create the first version of the algorithm we analysed the anthropometric data of over 8,000 individuals across 12 variables and superimposed this with more than 2,000 manual fittings.

Things such as waist diameter, arm length and bicep (plus much more) could now be calculated using only three biometric data points that everyone knows: height, weight, and date of birth. Women’s bodies have more variance and so we needed a fourth data point via bra size.

This statistical model would be right most of the time. But most of the time isn’t good enough, so we evolved.


You know yourself best.

A statistical model alone could never account for all the individual variations of the human body resulting from genetics, diet, and/or fitness.

Some people were on the swim team, others carry their weight in their tummy or maybe their thighs, and some have disabilities which affect body shape (sometimes significantly). Nature and nurture need to be taken into account, so we added Natural Language Processing (NLP).

You know yourself best so we take your advice on board in the form of notes on every product page and throughout the checkout flow. We want to make the perfect Tee for everyone and now we can thanks to this layer of Magic Fit®.


Our experience still matters.

They say there’s no substitute for human experience and we're inclined to agree. Thousands of manual fittings have allowed us to create the perfect Tee for all sorts of bodies and see what works with our own eyes (and what doesn’t).

The first two layers calculate your body measurements and this layer adds the right tolerances throughout the Tee to make it absolutely perfect.

We are constantly doing IRL fittings with people of different heights, weights, and body shapes to build up our datasets which are used to backfill any gaps in the statistical model.


Never stop learning.

Not every Tee we make is perfect the first time around, which is why we offer free alterations. We actually love it when we get it wrong because it's a chance to learn and get better!

Our Production Team looks at your original Tee to diagnose exactly what went wrong and how to fix it. The corrections are then analysed automatically by the algorithm to constantly improve itself.

The more Tees we get wrong, the more accurate Magic Fit® becomes.

The data sets from the fittings done in the previous layer also allow us to catch mistakes before they even happen, allowing us to train Magic Fit® proactively on all kinds of different bodies.


But don't take our word for it...

Even though you now know exactly how Magic Fit® works, it still feels like magic, which is the whole point. So please give it a try and if you have any recommendations, comments or criticisms, we're all ears. 



Written by Rahul Mooray