Alexandra Van Houtte spent her early career chasing down runway images. As a fashion assistant at the Paris offices of Numéro, Grazia, Glamour, and assisting on international editions of Vogue, she was tasked with researching looks from the four major Fashion Weeks that take place around the globe all year long. At each event, designers present hundreds of runway shows totaling many thousands of looks.
If her editor wanted a photo of a "red dress," Van Houtte, who's originally from the U.K., would spend 10 hours sifting through images of shows to supply viable options. "I’d have to go through every single show, screenshot-ing it and putting it into a folder," says Van Houtte, 27. "It took up so much energy. There are so many collections every year, so many styles, and nothing to sort them out… It was really tiring."
By her own admission, Van Houtte is "impatient." And she was frustrated. She saw efficiency in so many other industries but hers: Hollywood has IMDB. Academics have JSTOR. Lawyers have LexisNexis. Where was the transformative research tool for fashion?
"[The fashion industry] is so forward-thinking on so many things but when it came to sourcing out collections, it took so much time," she says.
After one too many nights poring over thousands of images of ruffles, Van Houtte decided to end the tyranny of runway research. She quit her job at Grazia in May 2015 to create TAGWALK, the world’s first fashion search engine. Launched in January 2016, the site allows users to search Fashion Week collections by brand, season, fabric, style, city, and trend. Unlike shopping search engines, which present consumers with options broadly related to keyword searches, Van Houtte's startup is designed to help the fashion industry track down looks with as much specificity as possible, from runway shows in Paris, Milan, London, and New York that occur twice a year. It's like Google, but in Anna Wintour-speak.
Van Houtte envisioned a neutral resource where people could find every Fashion Week designer, big and small, established and up and coming. As the website explains, "TAGWALK is based on three words: fashion, simplicity, and rapidity." The idea was to build a database that housed every single look from each show as soon as it was presented, with multiple keyword "grades" for each image. For example, "denim" is one keyword. If a look is a full Canadian tuxedo, it would receive a 10/10 for denim, whereas a jacket with denim lining might get a 3/10.
The first step was finding a web developer. And from there... the initial six months were rough. "I did it all on my own," says Van Houtte, who personally contacted every single brand involved in Paris and Milan Fashion Weeks, requesting images to feed her growing database. It was, much like her assistant days, painstaking and tedious. She edited and tagged every single image by herself.
"That’s 10,000 pictures that I referenced with my own hands," she says. Eventually, she hired a team of fashion experts to help her sort through the mountains of visuals.
Today, the process works more smoothly. TAGWALK has a Paris-based staff of seven, including five ex-stylists who, with help from an algorithm, tag each image. TAGWALK now saves time by purchasing the images directly from photo agencies, but the manual labor is still required to achieve Van Houtte’s vision. While a computer can detect colors or specific items like a hoodie, it can’t detect niche trends, like balloon sleeves.
"The human eye is really important," says Van Houtte. Sometimes an ensemble can look beige, but it can also be interpreted as another color, like blush. These are the details that can't be fully executed by a software program (yet). That means the TAGWALK team personally judges each look and decides on specific keywords, such as "circus" or "powder pink." Brands can also provide their own keywords (or subtract those they deem incorrect). It’s a collaborative processes between Van Houtte's team and the fashion houses to help make the catalog as precise and user-friendly as possible.
Anyone can search the site, and if you type in, say, "Victorian," you'll turn up more than 230 runway images of Chanel, Gucci, Dior, and other big names synonymous with expensive. Related tag suggestions will include "Royal," "Princess," and "Tudor." Much like the internet as a whole, TAGWALK will lead you down a black hole that's at once confounding and fascinating.
"When I launched it, I was a bit naive," Van Houtte admits. "I thought it would really only help assistants... But the more people sign up, I realize it’s the whole industry using it."
Searching the site is free of charge, but brands can pay $50-$250 a month to have their images included in the database—an appealing option for smaller, independent labels that don’t have the budget to stage traditional runway shows, which can cost from $30,000 to $200,000. Today, the site has 7,500 subscribers, many of them buyers, merchandisers, editors, stylists, trend influencers, and fashion school students. Last August, TAGWALK had 3,000 unique visitors. By this past January, that number had tripled to 9,000 unique visitors—plus 400,000 impressions. "It’s a whole spectrum of people I wasn’t expecting," says Van Houtte. "It’s a happy surprise."
TAGWALK also sells general user trend data to companies showcased on the site. "We can see, per brand, which looks are the most looked at in which country and how a brand's collection has been tagged in comparison to other brands overall," Van Houtte says. This data gives companies an idea which items have consumers buzzing and could be safe bets to stock in retail stores. It also provides an overview of how the industry is reacting to a collection.
Sophie Roche Conti, VP of Catherine Miran, a PR firm that works with labels such as Isabel Marant and Carven, welcomed the ease of TAGWALK. "It’s incredibly obvious that it should have existed all along. It’s just that no one took the time to do it," she says. "We have all of this data [from shows and presentations], all of this creativity, and somehow we couldn’t organize it."
Shortly after it launched, Conti saw TAGWALK quickly spreading throughout the fashion community, both in the States and abroad. "It took over Paris really quickly," she says. "It’s become the tool that everyone uses."
Before TAGWALK, Conti would spend huge chunks of time putting together trend reports for editors and buyers. Now Van Houtte's database does all the heavy lifting. "Having that information helps me better pitch editors," she says. "It’s helped condense information in a way that’s just so useful.
Conti can’t imagine someone without a fashion background being able to pull off such an endeavor. The TAGWALK staff are fashion experts who know their Fendis from their Fords and are especially sensitive to emerging trends. On a recent day, the site's trends page highlighted everything from "White" and "Sequins" to "Extra Long Sleeves" and "Doll."
"It’s specific, but it’s user friendly," says Conti. "Even for the die-hardest of stylists, it takes a season for many to see what the trends are going to be."
Sophie Fontanel is a fashion journalist who previously wrote for French Elle and now writes for the weekly French news magazine Le Nouvel Observateur. When Van Houtte first told her about TAGWALK, Fontanel had a sense that it would not only "change the industry," but also create a new way of deciphering fashion data. "It's as if finally the avalanche of show images were given space to breath and be seen," Fontanel says via email.
Moving forward, TAGWALK plans to expand its service beyond Fashion Week and start tagging other noteworthy fashion trends, like streetwear.
"We're focusing on what younger generations are looking at and what interests them," says Van Houtte. "It's important to brands." She's well aware how powerful social media stars have become. When they wear a particular garment and post about it, that sometimes generates more press and industry chatter than a traditional Fashion Week show. As top fashion executives are increasingly paying attention to social media personalities, TAGWALK intends to do the same and include their outfits in the database.
Soon, the database will also start collecting searchable images of models, tracking which models walked which shows, who walked the most, and who are the new faces to know.
"I want to take clever steps," says Van Houtte. She knows that somewhere out there, an assistant is poring over Instagram photos of streetwear stars and runway models, wishing there were an easier way to find the needle in the haystack.
"Even if you’re really good at your job, you definitely can’t remember 9,000 looks—it’s physically impossible," says Van Houtte. Which is why she’s glad she created something that does all the hard work, within seconds. As she reflects, "I basically created what I would’ve loved to have."