The Latest Fashion, Trending on Google
By HIROKO TABUCHI

Normcore? So last year. String bikinis? Most definitely over.

Even interest in skinny jeans may be waning, if six billion fashion-related queries by Google users are any indication of this year’s most popular trends.

Instead, consumers are Googling tulle skirts, midi skirts, palazzo pants and jogger pants, according to the company, which plans to start issuing fashion trend reports based on user searches twice a year. The new trend aggregations are part of the company’s bid to become a bigger player in e-commerce and fashion beyond its product search engine or advertising platform.

In its inaugural report, Google distinguishes between “sustained growth” trends, like tulle skirts and jogger pants; flash-in-the-pan obsessions like emoji shirts and kale sweatshirts; and “seasonal growth” trends, or styles that have come back stronger every spring, like white jumpsuits. It makes similar distinctions among sustained declines (peplum dresses), seasonal ones (skinny jeans) and fads that are probably over and done (scarf vests).

Lisa Green, who heads Google’s fashion and luxury team, said the company had begun working with major retailers, including Calvin Klein, to help them incorporate real-time Google search data into fashion planning and forecasting. “Fast fashion” companies, for example, can take a trend identified by Google and run with it, Ms. Green said.

“We’re interested in being powerful digital consultants for our brands, not just somebody they can talk to about what ads they can buy online,” she said. “They can say, ‘Google has identified this as a trend, and we have six weeks to get this out on the racks.’ ”

Google’s foray into the fashion world is part of a scramble to define, inform and tap into how people search online for everything they can buy, be it clothes or jewelry, groceries or furniture.

The search giant has long experimented with e-commerce through services like Google Shopping, which lets shoppers compare prices among different vendors, and the recently introduced Shopping Express, which lets users make grocery purchases from local retail stores and receive them on the same day or the next one. But the company’s e-commerce business trails behind Amazon or Alibaba, the established go-to sites for a plethora of products, and in fashion, Yoox and Net-a-Porter are about to merge and flex their muscles as a luxury retailing powerhouse.

For Google, product search is increasingly important for its mainstay business of selling ads alongside search results, including fashion search results.

The company is hardly the first to deploy data to predict what might be a hot trend this season. IBM, for example, analyzes posts on blogs, social media and news sites to gauge “social sentiment” in a variety of brand categories, including fashion and retailing.

In one early experiment in 2013, IBM declared that “steampunk” — an industrial aesthetic inspired by 19th-century Britain — was set to “bubble up, and take hold, of the retail industry.” That prediction has played out to a certain extent, judging from the popularity of “industrial” or “salvaged” furniture, for example, or body trainers and corsets.

Spotify, the music-streaming service, also offers free analytics to artists to help them gauge the popularity of their music or estimate how much they can earn from new tunes.

“People tend to make trend predictions based on a very limited number of observations, and that’s very hit and miss,” said Trevor Davis, a consumer products expert who led the project at IBM. “The ability to detect trends very early on before they really become noticeable, and to follow them, is invaluable.”

Just how much more accurate or useful Google’s search data on fashion will be has yet to be seen, Mr. Davis said. One obvious weakness is that Google’s data encompasses all searches that appear related to apparel, regardless of whether the person searching actually bought something, or even intended to buy something. A search for “tulle skirt,” conceivably, could signal a shopper looking for an item on sale or a baffled fashion novice looking for a definition of it.

Asked by Google to assess its data analysis, Ellen Sideri, founder of ESP Trendlab, a forecasting agency, said that the data’s value lay in its focus on what real people were interested in from across the country.

“The hardest thing to explain is that a trend doesn’t come from one place,” Ms. Sideri said. “Every trend is multifaceted.”

For now, Google says it will share this database free to retailers and trend-followers in the hopes of winning partners and clout in the fashion arena. And Google promises not to match up its search data with customer data from retailers to target ads at individuals.

Ms. Green of Google said the sheer volume of its data — six billion data points — meant any patterns Google detected were a significant indicator of trend awareness and eventual purchasing behavior.

INstock, a data tool from the fashion trend forecaster WGSN, takes another approach to predicting trends. It analyzes fashion trends from pricing and inventory data on more than 40 million products daily across 12,000 brands in the United States and a handful of other markets. Styles picked up by a growing number of retailers signal a trend, and out-of-stock items likely demonstrate high demand, while heavy discounts raise a red flag.

Some current trends identified by INstock overlap with Google’s. Jumpsuits are in this season “as an elegant, yet edgy, alternative to the party dress,” with inventory offered by retailers growing by 12 percent between September and January compared with a year earlier, according to Loree Lash-Valencia, a vice president at WGSN. On the other hand, her data had not yet picked up substantial interest in tulle skirts.

While predicting trends remains difficult, accurate forecasting has become vitally important as fast fashion speeds up product cycles, Ms. Lash-Valencia said. Misreading trends can result in millions of dollars lost either from marking down inventory that does not sell, or not stocking enough of styles that do, she said.

“In our industry, there’s been a push to go faster and faster, and that’s one reason everyone is having problems: No one has time to plan,” she said. “Instinct’s no longer enough. Data can’t replace every merchant, but there’s such accurate data available now — you really need to use it.”

Ms. Green said the Google search data helped to unearth fashion-forward hot spots that had escaped the industry’s attention. “The industry might subscribe to certain perspectives, but our search data allows us to see what people really want to buy,” she said. “And our data shows that it’s not just that every trend starts in New York or L.A. and everyone else catches on.”

A geographical breakdown of Google’s data shows, for example, a flash of search activity for white jumpsuits in May 2013, and for palazzo pants in August of that year, in Jackson, Miss. Both styles then spread nationally. And “tulle skirt” surged in popularity as a search term in October 2013 around Salt Lake City, Utah, before that trend also spiked elsewhere.

While Google said it had no idea why that was the case, a little digging turned up possible clues.

Jason Bolin, a stylist in Jackson, held a series of local fashion shows for the 2013 spring and summer seasons that featured both jumpsuits and palazzo pants — and “everyone here went crazy for them afterwards,” according to Monique Pruitt, who has run Fashion Week Mississippi, a series of fashion events that feature local and Southern designers, since 2012.

The jumpsuit trend further took off in late 2014 when Solange Knowles, the sister of the singer Beyoncé and a fashion icon in the South, was spotted heading to her New Orleans wedding in a striking Stéphane Rolland pantsuit. They were also frequently seen on Kim Kardashian. (Ms. Kardashian also most likely spawned another rising trend identified by Google, the waist trainer.)

Mr. Bolin, who sells his styles online, said he was unsurprised by the Google data and lauded his region’s sense of style.

“Jackson’s a very boutique-y place,” he said. “We even have a fashion week,” he said. “It might be because there’s not a lot else to do. But we’re on trend with the world, believe it or not.”

The tulle skirt trend, for its part, most definitely came from a craze for fairy-style skirts that has swept weddings and baby showers across Utah in the past few years, said Sherene McClellan and Deanna Sorenson, who run their Tulle Skirt Shop on the Etsy marketplace from the tiny town of Mount Pleasant, Utah.

The sisters initially set up shop in September 2013 selling corsets and other period costumes, but soon found that demand for their handmade tulle skirts far exceeded that for other products. Last April, they decided to focus on the skirts, and have sent out about 700 of them since then, they said.

“Boy, we do weddings and babies,” Ms. McClellan said. “I think the trend comes from our girls going for sweet and romantic, rather than sultry and sexy,” she said.

“We are such a tiny rural area, and the fact that we might be a trendsetter for anything cracks me up.”

Google插手時尚 流行我說了算

隨興穿搭風？已經是去年的事了。綁帶比基尼？恐怕快要退流行了。如果Google的六十億筆服飾相關搜尋資料真能預測今年的流行，那麼連緊身牛仔褲可能都開始退燒了。

紐約時報報導，根據Google即將每半年發表一次的服飾流行趨勢統計，消費者現在蒐尋的大熱門是紗裙、寬管褲和收腳褲。Google希望藉由這份統計，跨出搜尋引擎和廣告平台市場，在電子商務和服飾領域扮演更大角色。

這份報告還細分為「持續性成長」趨勢，例如紗裙和收腳褲，以及「季節性成長」趨勢，例如每逢春天都會熱賣一陣的白色連身衣。退燒的部分也區分為持續性衰退（如荷葉邊洋裝）和季節性衰退（如緊身牛仔褲）。

Google服飾與精品部門主管麗莎．葛林說，他們已開始和凱文克萊等知名品牌合作，使用Google的資訊協助服飾業者預測和計畫要生產哪些服飾。她說，標榜「快速時尚」的品牌也可以參考Google公布的資訊，跟著趨勢走，未來這些服飾業者可能會說，「Google已經指出這個趨勢了，我們有六個星期的時間趕緊讓衣服上架」。

Google一直想跨進電子商務領域，也推出了Google Shopping等比價服務，但亞馬遜、阿里巴巴等網購大咖已成氣候，要急起直追並不容易，Google只好換一種方式切入。

同樣在做數據分析趨勢預測的IBM專家戴維斯說，Google的服飾搜尋數據究竟能不能精準預測時尚走向，有待觀察，畢竟這些資料只反映出網民搜尋的熱潮，但搜尋後不一定會出手購買。

對服飾業特別是「快速時尚」業者而言，預測時尚走向是困難但必要的工作，預測錯誤，不是產量不夠嚴重缺貨，就是存貨堆積賣不出去，代表數百萬美元的損失。

原文參照：
http://www.nytimes.com/2015/04/27/technology/the-latest-fashion-trending-on-google.html

2015-04-28．聯合報．A5．話題．編譯組