[HBR] The CEO of Williams-Sonoma on Blending Instinct with Analysis – Laura Alber

In Blog
Scroll this
From the September 2014 Issue

 

 


Photography: Stephanie Diani


The Idea: Thanks to its catalog heritage, the retailer was well positioned to collect and use data from customers and other sources. Now it has reorganized to take its analytics to a whole new level.

As I walk each day into the Williams-Sonoma headquarters on San Francisco’s waterfront, I’m struck by the varied types of creativity on display. In one area designers are developing textiles, painting furniture, or spinning pottery. In another, brand and merchandise specialists are reviewing photography and revising catalog and website layouts according to projected sales. And in another, data analysts are at their computers, crunching numbers, building models, and analyzing reports.

If Williams-Sonoma has a “secret sauce,” it is these teams working together in remarkable alignment to develop and execute our strategic and tactical priorities. In my 19 years at the company and four years as CEO, I’ve found that the very best solutions arise from a willingness to blend art with science, ideas with data, and instinct with analysis.

The recent launch of Mark and Graham, our monogrammed gifts and accessories brand, is a perfect case study in how we apply this kind of thinking to build successful new business concepts. Since we introduced the Pottery Barn Kids brand, in 1999, personalization has been a key component of our business, and by 2010 we could see that it presented a clear opportunity to develop a stand-alone brand. We set up the Mark and Graham website, printed the catalog, and marketed to customers in our database who had previously bought personalized items from us. We also placed ads targeting online shoppers who browsed that category. Mark and Graham is now a scalable brand.

The process is the same for nearly every decision we make, from brand introductions to ad buys, website and catalog design, and, increasingly, supplier relationships. We start with a creative idea, test it, prove that it works, and then roll it out. This discipline has allowed us to build Williams-Sonoma into America’s preeminent home retailer, with seven brands and annual revenue approaching $5 billion. We account for 4% of all home furnishings sales in the United States, almost half of which come through the web, making us the 21st largest online retailer in the country.

Inspiration and Expansion

Williams-Sonoma started with a brilliant idea. In 1956 the company founder, Chuck Williams, decided that America’s home chefs, himself included, should have access to the high-quality French cookware he had discovered on a recent trip to Paris. He set up shop in Sonoma, California, with a woodcut of a pineapple (symbolizing hospitality) as his logo, and began to welcome customers. So many of them came from San Francisco that he soon decided to move his operations there. In the early 1970s he launched a catalog to serve customers from around the country. That was when the company first began to capture and use data: By observing where our catalog customers lived, we could make better decisions about where to locate our retail stores. It was an effective strategy that we—and our competitors—still use today. And I am delighted to report that at age 98, Chuck Williams still keeps office hours at our headquarters.

As the company grew its brick-and-mortar presence, it also built up its mail-order business, first acquiring an existing garden products catalog and then launching the Hold Everything catalog (a brand we later folded into the rest of our portfolio). The company went public in 1983 to finance further expansion, and in 1986 it bought Pottery Barn, a group of home goods stores located predominantly in Manhattan. We kept the name but immediately started changing the merchandise. The following year Pottery Barn’s first catalog was mailed across the United States.

I joined Williams-Sonoma as a Pottery Barn senior buyer in 1995, at a time when the company was just developing its “lifestyle” identity, selling its first sofas through a catalog and introducing new product categories. We projected this identity by pioneering the use of photography that featured products in home settings, not just the products alone, and redesigning our stores to inspire our customers in the same way.

Our foray into e-commerce began in earnest in 2000, with both the Williams-Sonoma and Pottery Barn brands launching fully transactional websites that year. Hard as it is to imagine now, e-tailing, especially in the furniture category, didn’t feel like a sure thing at the time. We wanted to be certain we could provide the same kind of shopping experience online that we did in our stores and catalogs—offering tips, gift ideas, customer service, and quick and convenient purchasing. So we took small steps, launching an online version of the Williams-Sonoma bridal registry in 1998 and then potterybarn.com, featuring virtual home tours, in 2000. As we began our web data collection efforts, we continued to use purchase information from the catalogs and stores to guide our business decisions. For example, when a catalog-only Williams-Sonoma product immediately sold out, we not only reordered it from our supplier but put a bigger picture of it in the next mailing and started stocking it in our stores. When we noticed that bed and bath items were some of Pottery Barn’s best sellers, we created a separate catalog for those two categories.

Pottery Barn Kids is another case in point. In 1998 I was pregnant with my first daughter and couldn’t find quality furniture and accessories for her nursery. I knew lots of other soon-to-be or new moms (including many colleagues) whose experience was similar, and I saw a business opportunity. But even then instinct wasn’t enough. We decided to test the idea by creating a collection and marketing it in a small, targeted catalog mailing. In some cases we simply scaled down popular adult products, such as our sleigh beds; in others—toys, for instance—we asked our designers to think back to their own childhoods and get creative. Sales were so strong that we didn’t have sufficient inventory to do the second mailing we’d planned. But we had proved that the Pottery Barn Kids concept would work.

The next question was whether to open stores. The pitch my team and I made to our then CEO, Howard Lester, and other top leaders again involved both creativity (we set up a mock store in our parking garage) and data (we mapped out the neighborhoods where people had made the most Pottery Barn Kids catalog purchases and found the best-performing shopping centers or streets in each). As a result, the company opened eight stores in 2000 and added many more, plus a website, the next year. PBteen soon followed, and our West Elm and Williams-Sonoma Home brands were born in the same way in 2003 and 2004.



Capitalizing on Data

In the years since, Williams-Sonoma has taken its use of data to an entirely new level—one that I know is unique in the retail sector. Today analytics infuses—and enhances—all areas of our business, in decisions both big and small. To cite examples in marketing, each brand’s website is designed to have personalized content based on what we know about the customer visiting it: If the previous purchases of registered users or the browsing history of nonregistered users suggests that they’re interested in sales, we’ll make sure current promotions and discounts are featured at the top of the home page. If you always buy boys’ items from PBteen, we’ll send you a catalog with a boy’s bedroom featured on the cover. If we know that you just bought a custom upholstered sofa from Pottery Barn, we’ll send you care tips. If we see that you’ve been looking at West Elm beds online, you might find a reminder ad on other websites you visit in the future. Our overarching goal is always to enhance the customer experience. Relevance is our guiding principle.

 

Personalized Shopping


When customers come to the website of any Williams-Sonoma brand, they are greeted with pages tailored to their preferences, as indicated by past purchases or browsing history.

Banners try to entice customers with deals based on their previous buying behavior. Does she shop sales? Does he wait for free shipping?

Photos of products in room settings were an early differentiator. Brands also use plain product shots and track which approach leads to more sales.

External websites such as Pinterest can be an important driver of traffic. Data analysis shows that people clicking through from elsewhere are more likely to buy if they’re presented with multiple products, not just one.

 

 

Recommendations are increasingly algorithm-driven. Someone browsing sofa pillows might also need a new cocktail table. A buyer of bedding for a girl’s room might want to check out girls’ backpacks.


We’re beginning to use data in our supply chain in much the same way, tracking not only key service metrics but also design and production processes to ensure that best practices are spread throughout the company.

Everyone at Williams-Sonoma now recognizes that our business is both an art (creating products that people will love) and a science (presenting those products to our customers when they want or need them). Creative curation is what distinguishes us from generalist e-tailers; data analytics is what puts us ahead of other home furnishings retailers. To get to this point, however, we had to refine our thinking in a few key areas: structure, people, tools, and culture.

 

Williams-Sonoma Facts & Financials


Founded: 1956

Headquarters: San Francisco

Store count: 585

Full-time employees: 20,300

 

 

Source: Williams-Sonoma

Organizing for Analytics

Our strategy is simple: We help customers create the homes of their dreams by providing them with a best-in-class, multichannel retail experience. But in the early 2000s our organizational structure was preventing us from achieving the second half of that goal. Our businesses were still divided into stores, catalogs, and websites, which meant that people from the same brands were competing with one another for sales. So we restructured, assigning a single president for each brand and all its channels. Since becoming CEO, I’ve worked to bring our information technology and e-commerce teams closer to the brands. We’ve centralized and integrated our marketing, analytics, and IT functions so that we can more easily share and execute on ideas and information that benefit the entire group. We also created two new senior vice president roles: One SVP heads a combined e-commerce engineering and marketing group, reporting to both our CIO and our CMO, and the other oversees multichannel direct marketing across our brand portfolio.

Of course, the right organizational structure is nothing without the people supporting it, and Williams-Sonoma is intent on hiring, retaining, and developing the best designers, merchants, marketers, technologists, and data scientists we can. I am proud of the fact that the average tenure of our top 14 executives is almost 14 years. We’ve also worked to attract exceptional advisers. A few years ago we recruited the head of global marketing at Google to join our board, and she’s been instrumental in helping us set and achieve increasingly ambitious goals. Our San Francisco location, close to Silicon Valley, gives us access not only to an extraordinary amount of talent but also to terrific industry knowledge. For example, we were early partners of Pinterest, and we frequently work with tech companies on beta tests.

To put data to work, you also need tools, and we’ve made those investments. Our databases document transactions made by our customers across our brands and channels. This information spans 30 years, and more comes in every day to give us insights and create a more relevant, personalized experience for the 57 million people who have bought our products. It informs every customer-facing aspect of the business, from targeted catalogs, ads, and e-mails to the extremely high-touch service that our retail sales associates and customer care centers provide. We have product tracking systems that show the status of every item in our pipeline—what’s on back order, in direct ship, being returned—and allow us to build detailed vendor scorecards, so we can celebrate and spread exceptional performance while also identifying weak spots. None of this would be possible if our whole team didn’t embrace an open, collaborative, cross-disciplinary, “create, test, prove, roll” culture. So I try to emphasize those values every day—with my words and my actions. I encourage colleagues at every level to challenge the status quo and offer their opinions.

One of our marketing executives likes to tell a story about my initial reluctance to replace merchant-driven product recommendations on our e-commerce websites with algorithmically driven ones. We have a world-class merchant team that understands our products and the related items our customers might want to consider. Although I was sure that expertise would always be more effective than a machine in driving cross-selling opportunities, Pottery Barn Kids had had some success with marketing that used predictive analytics, so I agreed to test it in other areas. As it turned out, the brand’s product assortment had grown so rapidly that our merchants had trouble keeping up with shopping recommendations. Algorithms can more easily scale and better use large amounts of customer data. They also perform much better, not only driving incremental revenue, but giving our merchants time to focus on other critical activities that differentiate our user experience.

We Are All Owners

As our organization has embraced analytics, we’ve found that major decisions have become less stressful. A good example is the launch of West Elm. Initially, some were concerned that it might cannibalize our existing brands: Would mailing the West Elm catalog to previous Pottery Barn customers hurt Pottery Barn sales? We quickly found, from analyzing the data, that people who received both catalogs actually bought more from both brands than those who received only one brand’s catalog. That allowed us to expand West Elm mailings quickly, knowing that the new brand’s offerings were meeting different customer needs.

The blending of art and science has also brought us together as an organization. We like to say that “we are all owners.” And we are. Every team member—regardless of level, function, brand, or channel—works for the good of the entire enterprise. We’re a creative group that uses data to make smarter decisions, to strengthen our relationships with customers and suppliers, and to grow our business in the United States and around the world.

A version of this article appeared in the September 2014 issue of Harvard Business Review.

Laura Alber is the CEO of Williams-Sonoma.
Tags: / Category: Blog

1 Comment

  1. 윌리엄 소노마는 미국의 카달로그 주문으로 큰 인테리어 업체인데, 카달로그로 쌓아온 고객데이터를 Decision making, locating Store, 고객 구매유도 등 다양한 분야에 IT 기술을 적용하고 있다고 한다. 특히 인상깊은 사례는 전문가에 의한 추천이 물론 더 퀄리티가 좋겠지만, 급격히 증가하는 고객을 하나하나 다 전문가가 추천하는데 시간을 버리는 것 보다. 알고리즘을 활용한 자동 추천을 도입함으로서 전문가들은 더욱 중요한데 주의를 기울일 수 있게 했다는 부분이었음. 되도 않되는 똥고집으로 무작정 매출만 올리라는 짓을 하는 일반 기업들에 비하여 과학적이고 합리적인 분석을 바탕으로 사업을 펼쳐나가는 부분이 참고할 만한 점이었음

Submit a comment

이메일 주소는 공개되지 않습니다. 필수 필드는 *로 표시됩니다