When Julia Fowler was working as a fashion designer in Australia back in the early 2000s, she found herself frustrated by the lack of information available to help her understand and respond to the latest trends.
2000年代初，当时还在澳大利亚从事时装设计工作的茱利亚o法勒发现了一个恼人的问题：手头上的信息源太少了，没法帮她及时了解和响应最新的流行趋势。
“We had internal data on the performance of previous seasons’ products and access to inspirational trend sites,” she recalls, “but no way to understand opportunities we’d missed or concrete data on how we could improve our product assortment.”
“我们掌握着前一季产品业绩的内部数据，也可以访问一些能够给人启发的时尚网站，但是没法知道我们错过了哪些机会，也没有具体数据告诉我们怎样才能改进我们的产品搭配。”她回忆道。

With nowhere to turn, Fowler decided to take it upon herself to develop a solution to the problem. Her timing was just right: A methodology and series of technologies collectively called “big data” was beginning to swell in the technology industry.
由于不知道向谁求助，法勒干脆决定自己开发一套解决方案。她挑选的时机再恰当不过。当时。一系列被合称为“大数据”的方法和技术刚刚开始席卷整个科技行业。
Fowler has since swapped her title of designer for that of co-founder at Editd (pronounced “edited” and stylized in all caps), a company she launched five years ago with technical co-founder Geoff Watts, who now serves as the company’s CEO. Their mission: to help the world’s apparel retailers, brands, and suppliers deliver the right products at the right price and the right time.
没过多久，法勒的头衔就变成了Editd公司联合创始人。另一名负责技术的联合创始人吉夫o瓦茨目前担任这家公司的CEO。他们的目标是帮助全球服装零售商、品牌和供应商在正确的时间、以正确的价格交付正确的产品。
“Every time you see a product on discount, it’s because the wrong decisions were made,” Fowler says. “This leads to a lot of wastage in the industry. I wanted to fix that problem.”
法勒表示：“每次你看到一个产品打折，那都是由于错误的决策导致的。它导致这个行业出现了大量损耗，我希望解决这个问题。”
Editd says it now has the biggest apparel data warehouse in the world. It offers that data up to customers along with real-time analytics and an assortment of other tools, powered by 120 servers and hundreds of terabytes of data. The London-based company, which has 27 employees and $6 million in investment, counts Gap and Target among its customers. It’s also profitable, Watts says, though he declined to disclose the company’s revenues.
Editd公司号称拥有目前全世界最大的服装数据库。凭借120台服务器和几百兆兆字节的数据，该公司不仅向客户提供各类服装数据，还提供实时分析与各种其它工具。总部设在伦敦的Editd公司目前拥有27名员工和600万美元的资本，快时尚品牌Gap和塔吉特百货(Target)等大公司都是它的客户。瓦茨声称，Editd公司目前已经盈利，不过他拒绝透露该公司的具体收入。
53 billion data points
530亿个数据点
Part of Editd’s secret sauce is the way it aggregates fashion trend and sales information from a wide variety of sources around the globe—from retail sites, social media, designer runway reports, and blogs covering trends—and then makes it accessible in real time. The company’s dataset includes no fewer than 53 billion data points on the fashion industry dating back more than four years. It covers more than 1,000 retailers around the globe and boasts 15 million high-resolution images. Its Social Monitor feature, an aggregated dashboard of social activity by fashion influencers and experts, includes more than 800,000 people.
Editd的成功秘诀之一是，它汇总了来自全球各种来源的流行时尚数据和销售信息——从零售网站、社交媒体，到设计师的T台走秀报告，再到流行博客——然后设法实时获取这些数据。该公司的数据库包含了至少530亿个来自时尚行业的数据点，有些信息可以追溯到四年前。它还涵盖了全球1000多个零售商，同时拥有1500多万张高清图片。它的“社交监控”功能监控着全球80多万名有影响力的时尚潮人和专家的社交活动。
To keep its data readily accessible, Editd stores most of it in memory, not on disk. “That’s really important,” Watts explains. “We need to access all of that and query that in any possible way. It needs to be super-responsive.”
为了随时读取这些数据，Editd公司把大部分数据储存在内存而不是硬盘里，对此瓦茨解释道：“这是非常重要的。我们需要以任何可能的方式读取和查询所有数据，它必须具有超强的响应力。”
It also needs to be easy for a layperson to grasp. “People shouldn’t have to be data scientists to understand the insights,” Watts adds.
另外，它必须足够简单易懂，让外行也能知道数据的意义。瓦茨表示：“用户不必非得是一名数据学家才能理解这些数据的含义。”
With Editd’s service, apparel professionals in merchandising, buying, trading, and strategy can set up and tailor their own dashboards and monitor whatever they choose from virtually any device. The service spans menswear, womenswear, children’s apparel, accessories, and beauty. Because the output can be customized, a denim merchandiser at a premium retailer, for instance, would see a very different set of data than a women’s knitwear buyer at a mass-market chain store.
借助于Editd提供的服务，从事新品规划、采购、贸易和战略规划等工作的服装业从业者几乎可以在任何设备上设置他们自己的“社交监控器”。Editd的服务涵盖男装、女装、童装、配饰和美容等多个领域。由于输出端的信息是可以定制的，所以一家高端服装店负责牛仔服的业务员所看到的数据，与一家平价服装连锁店的女款针织衫采购员所看到的数据是截然不同的。
Editd issues daily and weekly retail reports to highlight new and discounted products in chosen market categories. Its analytics tools are intended to help industry professionals track the competition and refine their own product planning. A visual merchandising archive helps shape promotion strategies for upcoming seasons.
Editd公司每天和每周分别都会发布反映特定市场类别的新品和打折商品情况的零售报告。它的分析工具则致力于帮助业内人士追踪竞争情况，改进自己的产品规划。Editd还有一个虚拟的销售规划档案工具，可以帮你制定下一季的促销战略。
One of the biggest benefits of using Editd is that industry professionals no longer need to “comp shop,” short for competitive shopping, to research the competition. At one of Editd’s more data-driven customers, the entire buying and merchandising team used to stop work for one week every six to spend the time visiting competitors’ websites for information —how many types of skinny jeans are on offer, for example, and how they were priced.
使用Editd的最大好处之一，就是业内人士们不必再去“竞争性购物”（即调查竞争对手）了。比如Editd公司就有一个非常重视数据的客户，该公司的整支采购和销售团队每过六个星期就要专门抽出一周时间，到竞争对手的网站上搜集信息，比如他们有多少款紧身牛仔裤，每款定价多少钱等等。
“They’d put together the reports in Excel, then the booklets were bound and distributed around the company,” Fowler says. “That was their playbook for the next six weeks.”
法勒表示：“他们要把这些数据汇总到Excel表格里，然后做成小册子在公司里散发。这就是他们接下来六个星期里的‘销售兵法’。”
Not only was the process time-consuming, but it was “fraught with danger,” Watts says. “So many errors creep into things.” In some cases, items might get double-counted. In others, different data collection methodologies might be used.
瓦茨表示，这种方法不仅非常耗时，而且“充满了危险，很多错误都可能发生。”在一些情况下，有些项目可能被重复计算，还有些时候，一些不同的数据收集方法可能被混用。
In a boundary-blurring business like fashion, categorizing products across retailers is another challenge. Pants, capris, or shorts—or something else entirely? “The way we analyze the kinds of products and the categories of products is very important,” Watts says. “We use computer vision and natural language processing to understand, for example, ‘This is a floral dress’ or ‘This is a cardigan.’ Unifying that and making it one consistent, clean data set is an incredibly important part of what we do.”
在时尚业这样一个边界比较模糊的产业里，光是给产品分类就是一个不小的挑战。比如裤子就有长裤、七分裤、短裤等许多种类。瓦茨表示：“我们分析产品种类的方法也非常重要。我们使用了计算机视觉和自然语言处理程序给服装分类，比如‘这是一件印花连衣裙’或‘这是一件羊毛开衫’等等。对于我们的工作来说，统一分类标准，生成一个干净、一致的数据库是一个极为重要的部分。”
Today, an Editd user can simply run a query on cardigans, for example, and receive results in under a second, Fowler says. More than 50 million SKUs are tracked by the system, she adds.
法勒表示，Editd的用户现在只需要输入“羊毛开衫”几个字进行查询，不到一秒钟便可以获取结果。她还补充道，Editd的系统可以追踪到5000多万个SKU（注：SKU即‘库存最小单位’。对于服装业来说，某一款服装的某一个颜色的某一个尺码，即是一个SKU。）
One Editd customer, the British online retailer Asos, credits the company’s services for the 33% jump in sales it saw in the last quarter of 2013. The company gave 200 of its employees access to the Editd system with a particular focus on improving the pricing of its goods.
Editd的用户之一英国在线零售商Asos声称，使用了Editd的服务后，其2013年第四季度的销售额跃升了33%。这家公司尤其注重产品定价环节的改善，已经给予200多名员工进入Editd系统的权限。
“What this technology and the changes to the industry are unlocking is the ability for customers to have exactly what they want and not necessarily what’s been decided for them,” Watts says. “It lets consumers be more fluid with their tastes and it lets the market be more efficient and more green.”
瓦茨表示：“这项技术以及它给行业带来的变革，使客户能够获得他们真正想要的东西，而不是由别人决定给他们什么东西。它使客户可以更加动态地掌控他们的时尚格调，也使市场更加高效、绿色。”
A million products, 11 million SKUs
100万个产品，1100万个SKU
Editd isn’t the only fashion-focused company dipping its toes in the big-data waters. Vying for a share of the market is the British trend forecaster WGSN, which just last year launched its own first big-data offering, Instock.
Editd并不是唯一一家试水大数据的时尚公司。英国时尚预测机构WGSN也想在这个市场上分一杯羹。WGSN去年刚刚推出了它的首个大数据服务Instock。
WGSN claims its dataset has more than a million products and 11 million SKUs each day from more than 10,000 global online brands and retailers. Instock, essentially a retail analytics service, is intended to complement its widely used trend-forecasting service by adhering to the same product-categorization taxonomy.
WGSN称，它的数据库每天都从全球10000多个在线品牌和零售商那里搜集100多万个产品和1100多万个SKU数据。Instock本质上是一项零售分析服务，它恪守着同一种产品分类方法，旨在补充该公司被广泛使用的时尚趋势预测服务。
“We link the taxonomy from the trend side in terms of how we categorize a specific shirt or dress or kimono and how we track it coming through and being presented in WGSN Instock,” explains Helen Slaven, global managing director for Instock. It’s a single, end-to-end taxonomy, in other words. By unifying the many ways in which different companies might interpret the same product line, industry professionals can make more effective decisions, she says.
该公司负责Instock业务的全球常务董事海伦o斯拉文表示：“我们对一件T恤、一条裙子或一件和服进行分类，并且将这种分类与它在WGSN Instock上的分类展示结合起来。”换句话说，它是一种统一的、端对端的分类方法。斯拉文指出，鉴于不同的公司对同一条产品线的命名可能存在差异，通过统一不同的命名口径，业内人士可以据此做出更有效的决策。
More than 6,000 customers use WGSN’s trend service today. The newer Instock service counts almost 50 global clients in nine countries. This season, WGSN plans to complement its existing data on womenswear, footwear, and accessories with information on kids’ apparel and menswear. A new service called Analysis+ will offer custom cuts of the data and the option of additional analysis.
目前已经有6000多个客户在使用WGSN的趋势服务。最新推出的Instock服务也已经在9个国家拥有了50名全球客户。除了女装、鞋类和配饰之外，WGSN还计划在本季继续补充童装和男装数据。另外，该公司还计划推出一项名叫Analysis+的服务，用于向用户提供定制数据和附加分析功能。
“It’s a really exciting time for big data and retail,” Slaven says. “By providing a lot more actionable insight, it’s completely changing the way retailers think about their process.”
斯拉文表示：“对于大数据和零售业来说，现在真是个非常令人兴奋的时代。通过提供大量更加有可操作性的见解，大数据正在彻底改变零售商对业务流程的看法。”
Watts, of Editd, agrees. “We help retailers have the right product at the right price and the right time,” he says. “That’s the kingmaking thing in retail. When you get that right, it unlocks a fortune.”
Editd公司的瓦茨也认同这一点。“我们帮助零售商在正确的时间，以正确的价格，提供正确的产品。这在零售业可以说是惊天动地的事情。如果你做对了，它会为你带来一大笔财富。”