June 7, 2019 Dave Desmarais

Case Study: Human vs Machine

Human versus machine, sounds like something out of a sci-fi movie.

In the world of technology, many rounds of testing are necessary to prove the viability of a new product. This is especially so when the product is a replacement for a manual process that has been in place for hundreds of years. When hearing about our garment measuring technology, fashion industry experts express excitement because it has the potential to replace a very laborious, time consuming task in the manual measuring of apparel. Of course, after the initial praise come the questions, and rightfully so.

One of our first defining A/B tests occurred in 2018 with a major US based fashion retailer. As the name of this post suggest, the test pitted 3 technical designers against Tailored’s Capture software to examine the viability of garment measuring technology; specifically accuracy, consistency and time savings.

What we did

Ten garments were used in the trial. Each person measured the garments in the same 5 areas. The human results were then compared with the results taken from Tailored, which measured each garment 3 times for comparison, and here is what we learned:

Accuracy & Consistency

We’ve talked at length about “accuracy” in garment measuring (see our post on accuracy here) and that due to the physical properties of apparel, there is no exact answer to the question “what is the measurement of this area” on a garment? That being said, accuracy is definitely important. Though an interesting point to note is those doing the manual measuring know the “answer” before they measure because they refer to the garment’s spec sheet.

For this test we wanted raw results and informed the team not to look at the spec sheets before or during the trial. Specifically what we were looking for was the number of times one technical designer recorded the same measurement as another, and we found this only occurred 54 percent of the time. This compared to Tailored at 65 percent of the time (For this trial we considered an exact match any measurement within 1/8 of an inch).

Speed

An unfortunate fact for us humans is most manual tasks cannot be done in parallel, we must perform one task at a time. This is not the case with machines/computers, and such Tailored was shown to improve efficiency during manual measuring tasks by 80 percent. This leads to an overall efficiency improvement of 20 percent or more, depending on the company and amount of manual measuring being done.

Conclusion

This test, along with others, continue to prove that Tailored will be a viable technology for many facets of the garment industry. Through tests like the one performed here and continuous advisement by our team of fashion industry experts, we know what is and isn’t necessary for Tailored to achieve this. That being said, we believe there will likely always be times when manual measuring is necessary.

As of this writing, Tailored is one week in to its newest software release; a web-based version of our garment measuring software. It’s primary use is for e-commerce in helping online shoppers find their best style/size. If you are a brand or retailer and are interested in how Tailored could help your business, please reach out.