Around 7.5 trillion photos were taken in 2017, according to multimedia storage platform Forever, with the majority of them snapped on our phones. If you needed a sign that the present (and future) is visual, this is it. The ubiquitousness of cameras has made images and videos one of the main ways we communicate with others and taking photos – one of the primary uses for our devices. This, paired up with advancements in emerging technologies such as AI and machine learning is changing the way we search for and dicsover information and products, namely by moving from text- to image-based search.
For fashion brands and retailers, adding image recognition and visual search capabilities to their online stores or apps bring enormous potential for better product recommendations, improved customer experience and upselling opportunities, which positively impact their bottom lines.
“The benefits for retailers are many,” says Matthias Dantone, CEO & co-founder at Fashwell, a technology company specialising in deep learning image recognition. “Through our customers, we have seen how Fashwell’s recommendation engines increase user engagement up to 250% and increase chances for conversion by 35%. In a recent study, we found that not only do 30% of text-based searches fail to show the desired products, but also that Visual Search is 8x faster than traditional text-based searches.”
Founded in 2014, the Swiss startup provides image recognition tools for fashion, home & living ecommerce businesses. Fashwell’s technology can be used to recognise products in any type of image and make them shoppable.
The most popular use case for Fashwell’s technology is visual search, which allows users to upload an image to a brand or retailer’s app to shop the products in the image. The startup claims its image recognition technology to be 10% above industry standard. It currently powers visual search functionalities for retailers including Zalando, Nelly.com and Bonprix.
“We saw a natural adoption to visual shopping with our customer Zalando. As soon as shoppers learned of the feature, we noticed how Zalando’s Visual Search queries and overall engagement were growing steadily from month to month. That’s because images are so intuitive for humans – customers prefer visual shopping,” says Dantone.
It’s not just an observation by Fashwell. According to tech company BloomReach, people who used visual search when visiting an ecommerce site viewed 48% more products, were 75% more likely to make a return visit and placed orders worth 9% more than those who did not use the technology.
Depending on their goals, retailers can also use Fashwell’s technology “behind the scenes,” for similar product recommendations (to find and display visually similar products), outfit recommendations (to analyse a model image, find remaining products she or he is wearing and add them to the page for customers to shop the full look) or product tagging.
Shop Spring, a leading fashion marketplace based out of New York, for example, approached Fashwell to add an element of personalisation to their app by sorting products according to the Spring style profiles.
When visiting the Spring app, shoppers can choose between different outfits. In the backend, each outfit and each product of that outfit are labelled with a corresponding aesthetic profile. This ranges from “classic” to “bohemian” or from “street” to “heritage”. Thanks to this selection process, Spring can offer personalised recommendations to each of their shoppers.
Fashwell’s classifier was tasked with identifying the style profile of each product based on its aesthetic attributes. Just as with the physical attributes, Fashwell’s engineers trained the classifier on images with the relevant aesthetic attribute label. For instance, flowing peasant skirts usually equal a “bohemian” disposition, just as plaid shirts and workman boots are considered “heritage” wear. “The end result is an unrivaled product: Fashwell’s attribute classifier caters to the personal style of any Spring customer and it gives Spring staff a great tool for managing their stock internally,” says Dantone. “After testing and looking at the product recommendation engagement, we also noticed a significant difference in customer behavior. Compared to standard product recommendation systems, our automatic recommendation engines more than doubled overall engagement.”
As the prolifiration of these functionalities are expected to increase, do brands still need to be educated around the topic? “According to Gartner’s 2018 CIO Agenda Survey, only 4% of CIOs have added AI functionalities to their business plan, with 46% having plans to do so in the near future. We see a similar pattern: there’s a definite interest in our technology, however it’s still quite a new concept to many,” explains Dantone. “We therefore take a unique approach, working with brands one on one to build a full business case that fits with their workflow.”
As technology continues to impact every step of the industry – from design, to production, to sales – early adopter brands are more likely to stand out, learn faster and reap the rewards that come with being open to change instead of fashionably late.