Gap inc annual report 2017

GAP Big Data Case Analysis

A Y EL E T I S R A ELI

J I LL A V ER Y

Predicting Consumer Tastes with Big Data at Gap

In January 2017, Art Peck, chief executive officer and HBS MBA ‘79, was struggling to turn around Gap Inc. following two years of declining sales in an environment where many brick and mortar retailers were under pressure. Peck took over as CEO in February 2015, after serving as president of growth, innovation, and digital, when he envisioned and implemented Gap’s digital strategy using an analytical approach (see vitae in Exhibit 1). Gap’s troubles were not new to Peck; the company had been struggling to regain its footing since 2000.

One way he hoped to improve operations was to eliminate the positions of creative director for each of the firm’s fashion brands and to replace them with a more collective creative ecosystem fueled by the input of big data. Creative directors were the visionaries of a fashion brand, serving as guardians of its image and providing its taste inspiration and wellspring of ideas. These designers, such as Karl Lagerfeld for Chanel and Christopher Bailey for Burberry, established a design direction for each line, created a small number of inspiration pieces, and oversaw and approved the designs of other products in the line. Their personal vision established and reinforced the look, feel, tone, and spirit of the brand.

However, Peck was critical about the amount of power this concentrated in one individual. Many creative directors with top notch design experience had come and gone during his tenure without

making a significant mark to boost sales. Labeling creative directors “false messiahs”,1 Peck reflected, “We have cycled through so many, and each has been proclaimed as the next savior.”2 Instead of betting the future on the next savior, he replaced creative directors with a decentralized, collective process that no longer required the approval of a creative director. Rather than relying on a single person’s artistic vision, Peck pushed the company to use the mining of big data obtained from Google Analytics and the company’s own sales and customer databases as the backbone to inform the next season’s assortment. Ideas could thus arise anywhere, even from Gap’s external vendors, and would no longer have to be vetted by a creative director serving as maestro of the collection. Once a trend was spotted, it could be immediately and simultaneously incorporated into all three of the company’s brands, hitting stores within three months. “There is now science and art, and they can come together,” in this new process, proclaimed Peck.3 With the elimination of his creative directors, he was upsetting the delicate balance between creativity and commercialization, between designers and merchants, that existed at most fashion brands and that had supported Gap Inc.’s fashion cycles for decades.

Peck was also considering expanding online distribution by selling Gap’s brands on Amazon, an online retailer. His previous role at Gap taught him the importance of e-commerce and digital and he

Professor Ayelet Israeli and Senior Lecturer Jill Avery prepared this case. This case was developed from published sources. Funding for the development of this case was provided by Harvard Business School and not by the company. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management.

.

517-115 Predicting Consumer Tastes with Big Data at Gap

expressed his opinion that Gap could be at a disadvantage if it didn’t consider the Amazon opportunity. Selling on Amazon could provide an additional datastream about customer purchasing

behavior to inform Gap’s decision making.4

Company Overview

Gap Inc. was founded in 1969 by Donald and Doris Fisher and their son, Robert was chairman of the board in 2017. Gap was one of the creators of specialty retailing, retailers that focused on a particular product category rather than carrying a wide assortment and produced their own private label branded goods. It remained the largest example of the genre, with 135,000 employees and 3,659 company owned and franchised retail locations in 50 countries, accounting for 36.7 million square feet of selling space,

which generated global sales of $15.5 billion.5 (Also see Exhibit 7).

Gap Inc. managed five brands: Gap, Banana Republic, Old Navy, Athleta, and Intermix, and had historically been the authority on American casual style. The Gap brand offered female and male consumers casual, classic, clean, comfortable basics: jeans, khakis, button down shirts, pocket tees -- at accessible prices. Some called it democratic fashion, “ordinary, unpretentious, understated, almost lowbrow,” while others labeled it iconic: “They elevated incredible basics to not just an iconic status in terms of clothing, but also a spirit – you felt like there was such a strong attitude, so much energy.”6 In 1996, Gap was at the height of its cool; actress Sharon Stone wore a Gap turtleneck on the red carpet of the Academy Awards.

In 1983, Gap Inc. acquired Banana Republic, moving into a higher price/quality tier. Luxurious materials were combined with detailed craftsmanship to support more expensive price points and attract a higher income consumer. In 1994, Gap Inc. created a new brand, Old Navy, to compete with discount department stores and mass merchandisers, such as Sears and Target, ushering in a period during which it became chic for consumers of all income brackets to shop for a bargain. Offering “wardrobe must-haves” at “prices you can’t believe,” embedded in a fun shopping experience, Old Navy was an immediate success with families, becoming the first retailer to reach $1 billion in annual sales within four years of its launch.7 Two acquisitions followed, Athleta (2008) a women’s fitness apparel brand, capitalized on the shift in women’s fashion from a jean-based foundation to activewear apparel. Intermix (2012) a multi-brand retailer of luxury and contemporary women’s apparel, offered consumers the “most sought-after styles” from a carefully curated selection of “coveted designers.”

In 1983, Millard “Mickey” Drexler became chief executive officer. During his tenure, sales grew from $480 million to $14 billion in 2000 and Gap’s market cap swelled to $42 billion. Drexler, described as “a visionary executive [that] helped transform Gap from a grab-bag of styles into a trend-setting machine that made simple clothes look great, even elegant,”8 was dubbed “the merchant prince” for his trendspotting, design instinct, and merchandising prowess. However, after being one of the first to predict the rise of business casual in the 1990’s, Drexler lost his magic touch, as he attempted to inject more fashion into Gap to attract younger shoppers who were migrating to edgier competitors. After eight consecutive quarters of declining sales, Drexler left Gap in 2002. Explained fashion writers,

Clothing companies…depend upon the vision and taste of just one person...Everything at Gap depends upon Drexler’s eye; it isn’t like making turbine engines. If he’s off the mark…if he approves a line of clothes in colors that aren’t just right, sales collapse and so does Gap’s stock price. That is why Gap can never really be like

Coca-Cola – there is no Gap formula hidden in some vault; there’s only Mickey Drexler.9

2

Predicting Consumer Tastes with Big Data at Gap 517-115

Two CEOs followed but were unable to restore Gap’s success in what the New York Times called “a remarkable comedown for a chain that once seemed to dictate how America dressed”10 (see Exhibit 2 for sales and net profit since Gap’s IPO in 1976 through 2016).

Every season, Gap produced hundreds of unique products, each offered in a variety of colors and sizes. While the online website typically offered the entire product assortment, each brick-and-mortar

store, with an average footprint of 10,000 square feet,11 was somewhat limited due to space constraints and offered a carefully curated subset of the product line. Gap’s assortment in each of its primary categories (women, men, children, and baby) consisted of two types of products: basics with styles that endured across seasons and more fashion-forward, designed items that captured the spirit of a particular season. Creative directors influenced the full product line, but their touch was most heavily felt on the latter group, where more fashion innovation was desired.

Digital and Big Data at Gap Inc.

As president of growth, innovation, and digital, Peck invested heavily in digital capabilities to address consumers’ shift to omnichannel shopping, focusing on dissolving the wall between the physical and digital channels. He observed, “Our customers are omni today and that is a fundamental reality. Many of our customers begin their journey with our brands on their phone and they finish it in our stores. Many of our customers begin their journey with our brands in our stores and they finish it on their phone.”12 He digitized the company’s entire product inventory and introduced retail services, such as reserve in store, find in store, and ship from store, which made it easy for customers to browse, purchase, and receive their items seamlessly across channels.

Peck promoted data-driven decision making and pushed his team to utilize big data to learn more about customers’ behaviors, and thereby deliver a better customer experience, “There’s lots of talk out there about big data—to me, big data, personalization is focused on an outcome of relevance. That’s

what we’re working on,” he explained.13 As the company moved into digital, Peck pushed his managers to continuously test and refine its new features as it listened to customers via its voice of the customer initiatives that tracked customer feedback and usage. A surprising finding arose: “Despite the explosive popularity of shopping not just online but via smartphones and tablets, 80% of Gap Inc.

customers still preferred to visit a store to try on the clothes.”14 As a result, Gap was working with Google and Avametric to develop an augmented reality app that allowed shoppers to test out different looks in order to improve their online and mobile shopping experiences.

Data-driven decision making required that customers be trackable and Peck lamented that customers were identifiable online but anonymous when they shopped in a store. He searched for ways to have customers opt-in to self-identify when they shopped in a store. He elucidated,

It is an opportunity to bring our personalization capabilities and customization relevance to bear in a store environment…60% of people visiting the website are recognized as unique visitors, enabling Gap to personalize experiences based on things like browsing and purchase history. Doing so is providing movement on numbers like conversion, time on website, click-through-rate…Good things happen for the customer if they’re willing to self-identify and tell us who they are at the beginning of a shopping experience. They do on the website, they don’t in our stores. If you come into our stores today, we won’t recognize you until you tender, if we recognize you then. This...is about providing…the opportunity to self-identify in order for the company to create a much

more relevant set of experiences compared to when they shop anonymously.15

3

517-115 Predicting Consumer Tastes with Big Data at Gap

Gap developed email programs to provide relevant, personalized messages to consumers. These included rules and conditions that when run through a series of algorithms triggered an email to certain consumers. For example, if a consumer abandoned her cart, an email was sent to remind her of her forgotten goods. If it was a consumer’s birthday, a personalized greeting and promotion was offered. If one of the brands was offering a new product line in a category that a customer had previously purchased, an email notification was sent to her highlighting it. Gap also used personalization in its geosniffing efforts, a term used to describe a company’s ability to determine the physical location of a particular consumer and to send them relevant localized information in real time. Information gleaned from clickstream analysis allowed Gap to reach out to consumers who had visited one of its websites, with customized messaging based on what they were searching previously, or to deliver a different landing page based on a consumer’s browsing history and/or IP address. Peck recognized the importance of allowing customers to opt-in to this type of digital tracking, “the company is carefully walking the line between personalizing a customer’s experience in a way that’s relevant and helpful

without creeping them out…Privacy is a huge concern for us,” he avowed.16

Managing the closing of underperforming stores (200 in 2011, 175 in 2015, and 75 in 2016) was another arena in which Gap used data-driven decision making. The company used the collection of insights from consumers’ online browsing activity and engagement in social media platforms to help understand why consumers were not buying as much from Gap’s physical stores. Peck proclaimed,

Visits to good malls are not down, but the number of store visits inside a mall are down, which says to me that people are planning their store visits as a function of their engagement with the brand, oftentimes expressed on a smartphone. I would argue that nobody’s figured out what exactly the aspirational, holistic, emotional expression of a brand…looks like when it shows up on this device right now.17

This insight drove him to further develop Gap’s digital and mobile e-commerce platforms to drive customer engagement. According to Fast Company, Peck had Silicon Valley developers “camped out at Gap, Banana Republic, and Old Navy stores, incorporating customer and salesperson feedback into

code in real time.”18 Peck’s performance and analytical nature were key to his selection as CEO.

Peck as CEO: The First Two Years

Peck was appointed CEO in October 2014. He faced some key challenges:

1. Slow growth in core markets: Gap Inc. competed in the $3 trillion global apparel industry, which accounted for 2% of the world’s gross domestic product (GDP). The U.S. and Canadian markets accounted for over $250 billion and were expected to grow annually by 2% through

2025.19 These two markets accounted for 84% of Gap’s sales. Millennials were spending less on apparel. Speaking to investors at a retail conference, Peck claimed that “there are no compelling [fashion] trends driving the business” and lamented that there had been a change in consumers’

buying habits such that there was a lack of need to replenish her closet.20

2. Competition: The mid-tier apparel landscape (see Exhibit 3) was highly fragmented, overcrowded, and competitive.

3. Rise of e-commerce: Consumers were shifting their purchasing from brick and mortar stores to online channels. In the U.S., 19% of apparel was sold through online channels in 201621 and, in 2015, clothing became the bestselling online sales category, driven by Amazon’s increasing strength in apparel. Amazon, the world’s largest multi-line, multi-brand Internet-based retailer,

4

Predicting Consumer Tastes with Big Data at Gap 517-115

was on track to become the largest seller of apparel in the U.S. by the end of 2017. As online sales grew, brands did not need the same number of storefronts. Empty stores lined the American shopping malls, as both specialty retailers and department stores simultaneously faced pressure to close locations. Gap had over 3,000 physical stores. By 2017, Gap Inc.’s online sales exceeded $2.5 billion.

4. Rise of Fast Fashion: New competitors, such as H&M and Zara compressed supply chains, delivering low priced looks knocked off from luxury fashion runways within weeks of their unveilings. With an average product cycle time of ten months, Gap lagged competitors such as Zara that could deliver products to stores within four weeks due to their consumer-responsive and decentralized buying process that allowed individual stores to order small batches of product, wait to see how consumers responded to it, and then airlift additional products to backfill the store’s inventory within days. The speed and pace of the fashion cycle was dizzying,

with new styles appearing in stores on a weekly basis in a constantly renewing fashion cycle.22

5. Heavy and frequent discounting: Clothing was increasingly commoditized as consumers viewed the lower quality fast fashion offerings as disposable, yielding a need for low prices and heavy discounting. Retail analysts were concerned about an overabundance of price promotion at Gap, where 40% discounts were common.

6. Gap’s size and ubiquity was transforming from asset to liability: Consumers, looking to forge a unique identity were moving away from Gap’s classic offerings.

Given these challenges, Peck believed that product assortment was key and that Gap’s model for selecting the right assortment was failing. The market seemed to agree. In January 2015, a retail analyst commented “They flip flop between a little trending, a little Euro, a little strip, whatever. It just gives you a headache…They’ve been redesigning the clothes for a decade because there is a total lack of clarity around who they are designing for. Who do you think their shopper is? I think it depends on

the week.”23 The flagship brand was struggling to find its place, wedged in the awkward middle between competitors’ value and premium brands (see Exhibit 4). Consumers, particularly millennials were cooling to Gap’s brands (see Exhibits 5 and 6).

While Peck knew that he was facing a 15-year-old problem that could not be fixed overnight, the results for 2015 and 2016 were disappointing (see Exhibits 7, 8, and 9 for recent financial performance).

Comparable salesa had declined for eight quarters before growing by 2% in Q4 2016 to deliver a -2% sales decline for the year, despite a 4% increase in marketing expenditures. Gap Inc.’s market cap had

dropped to $9.2 billionb and the board was looking for longer term solutions.

Peck’s Product Strategy: Big Data In, Creative Directors Out

Even prior to becoming CEO, Peck was skeptical of Gap’s creative directors. Creative directors were tastemakers, classically trained in design and using their unique eye, attitude, and personality to shape tomorrow’s fashions. They were arbiters of taste and provided legitimacy and credibility to new trends with their stamp of approval. “The creative director is God,” proclaimed a major fashion brand

executive.24 Rather than sensing or spotting existing trends, creative directors imagined and birthed them, “Creative directors are there to bring the magic to brands and product, and the magic to the

a Comparable sales include the results of Company-owned stores and sales through online channels. A store is included in the calculations when it has been operated by the Company for at least one year.

b As of January 31, 2017.

5

517-115 Predicting Consumer Tastes with Big Data at Gap

consumer experience...They are the conjurers of that incredible feeling we get when we buy something in a store or online that we really don’t need or didn’t know we needed until we saw it.” said Daniel

Marks, chief creative officer at The Communications Store.25 Without them, a company risked its brand asset, “You knew what Gap stood for when Mickey Drexler was running it...When you don't have a creative visionary leading a company, you can’t really establish a consistent look over a period of time

and reinforce a brand’s purpose,” declared Garret Bennett, a retail consultant.26

One of Peck’s first moves when he was appointed president of Gap North America in 2011 was to fire Gap’s head of design, Patrick Robinson. Robinson, who had designed for Giorgio Armani, Perry Ellis, and Paco Rabanne, led the design team from 2007-2011. He was a fashion insider, a friend of Anna Wintour, Vogue’s editor-in-chief, and a bit of a celebrity himself, dispensing advice in Glamour and Teen Vogue. He had been excited for his new role at Gap, “We needed to redefine all those American classics…for today. Not for 15 years ago. Not for 10 years ago.”27

After Robinson’s designs missed the mark, he blamed the poor retail execution of the company’s merchants or merchandisers. Merchants, or merchandisers, in the fashion industry were closer to the market with more of a commercial orientation. They were responsible for selecting products to craft a coherent assortment for each store to reach a particular target consumer at a particular price position. Peck explained the difference, “Design’s job is to push creatively and merchandising’s job is to

counterbalance that with a commercial orientation.”28 Merchants were market-responsive, while creative directors were market-leading. Gap’s head of merchandising, Michelle DeMartini elaborated, “I am representing the consumer, and [the creative director] is representing the future. And sometimes that creates conflict about what risks we want to take.”29

Robinson’s replacement, Rebekka Bay, was hired in 2012 following her successful launch of Cos, a modern, upscale brand designed for H&M, a leading fast fashion retailer. The Gap team had high hopes that Bay would bring her Scandinavian minimalist aesthetic and understanding of fast fashion to bear. Bay was a traditional designer, governed by her gut rather than by market research. Steve Sunnucks, global president for the Gap brand was excited by what she had to offer, “Her great skill is that while she is a trained designer, her experience in trend prediction means she takes a much broader view and thinks about the brand, the product, and the customer experience holistically.”30 Said Bay, “I’m intrigued by the process of fashion, the collective mind, how we all suddenly have a taste for the same things.”31 She explained her approach as head of 160 designers at Gap:

My role is to balance creativity and commerciality. Good design is less about taste and more about integrity…You need a very strong foundation. You have boundaries, and you can only – and I’m kind of rigid about this – you can only work within them. First, you design the most iconic piece. Then you can maybe create a seasonal version of that. If anyone is going to go beyond that, I have to agree to it.32

In January 2015, as Peck transitioned into his CEO role, he dismissed Bay, judging her design aesthetic--unadorned, simple, structured with a loose, ultramodern fit and somber black and gray palette—to be inconsistent with Gap’s optimistic brand. Bay saw it differently, claiming that “Gap is

not a design-led company and thus I had very little say in what ended up in the store.”33

At Banana Republic, creative director Marissa Webb, owner of her own eponymous fashion label, was hired in April 2014 to leverage her sensibility and credibility with younger consumers. Peck was disappointed by her first effort, “It’s had a couple of very positive impacts in terms of reestablishing some fashion credibility for the brand, but we didn’t get it 100% right...The color palette was pretty

stark...we’re still working to buy an assortment that is both commercial and fashion-oriented.”34 Webb stepped down in October 2015 after only eighteen months on the job.

6

Predicting Consumer Tastes with Big Data at Gap 517-115

Neither Bay nor Webb were replaced. Instead, Peck’s solution was to eliminate the position of creative director and spread the responsibility for design of the brand’s seasonal lines to a collaborative team informed by hard data. At an investor conference, he explained his decision,

We need great design. We need great creative talent. But we need that talent to be part of a highly collaborative team every season…Where we have gone wrong oftentimes as a company is when we have put the burden of running these brands season after season on the shoulders of an inspired individual. That’s not the model for success…These are

businesses, global in scale that require a highly collaborative team to be success.35

Two former employees voiced disapproval. “Anything that has to become a consensus is an equation for dilution…Without a distinct point of view, you become like everyone else,” said Todd

Oldham, a creative director at Old Navy.36 “There are not many retailers with more resources than Gap to create the next trend...In this retail environment, you have to take risky bets to even have a chance,” said Rajiv Malik, vice president of Gap global product operations. Retail analysts were skeptical. “There’s really no fashion direction…Right now, they’re a ship without a captain,” said one.37

Peck formulized his approach in what he called Product 3.0 (detailed below).

Big Data and Predictive Analytics in Marketing

Digital data streams allow companies to observe their consumers’ purchase journeys and collect a detailed trail of data about their online behavior. The mining of big data could yield many actionable insights to inform managerial decision making, such as identifying consumers who were more loyal to brands, matching consumers to products they might prefer, or predicting the behaviors or characteristics that could cause consumers to churn. By uncovering patterns in past customer behavior, companies could develop heuristics or algorithm-driven protocols to customize how they treated future customers to maximize satisfaction and/or profitability. It allowed remarketing or retargeting: as companies observed that a particular visitor viewed an item online but failed to purchase it, they could immediately serve up customized digital advertising that appeared as customers surfed other websites to entice them to return and complete the purchase. As digital data streams became more accessible and robust, companies were exploring how to use datamining and machine-learning to induct consumer preferences and predict future behaviors.

Utilizing predictive analytics to sell existing products: E-commerce companies, such as Amazon and Netflix, used predictive analytics to mine data to generate personalized product recommendations for their users. These suggestions were often based on aggregate data from other users, usage patterns of similar users, or a user’s own purchase history or expressed preferences, generally gathered through reviews of existing purchases or through preference polling. Offline retailers used purchase histories, accessed as customers swiped a loyalty card at checkout, to drive algorithms that determined which consumers should receive coupons or promotions. In 2012, to the dismay of her father, a teenage girl received coupons for baby clothes from Target. The retailer’s data algorithms predicted that she was pregnant even before she herself knew that she was.38

Amazon had recently patented “anticipatory shipping.” The idea was to move beyond merely providing recommendations to consumers, and instead, anticipate, based on the consumer’s historical behavior, when the consumer would need an item. Using information such as previous orders, product searches, wish lists, shopping-cart contents, returns, and even how long an Internet user’s cursor hovered over an item, Amazon would preemptively ship products to a distribution center close to the

7

517-115 Predicting Consumer Tastes with Big Data at Gap

consumer, in anticipation of an incoming order. This would reduce the time lag between ordering and receiving a package to dissuade consumers from feeling the need to visit physical stores.39

Utilizing predictive analytics for new product development: Beyond making viewing recommendations, Netflix used data to make decisions on which new series and movies to develop. However, its CEO Reed Hastings cautioned, “We start with the data but the final call is always gut. It’s informed intuition. Data science simply isn’t sophisticated enough to predict whether a product will

be a hit.”40 Stitch Fix, an online styling service that delivered a personalized shopping experience by curating outfits for consumers based on their expressed preferences, aggregated all of its consumer preference data to learn which fashion elements were popular and then used that insight to design its own private-label fashion products.

Some companies utilized secondary data to anticipate market trends. L’Oreal Paris analyzed data from Google searches, social media sites (YouTube, Facebook, and Instagram), and fashion magazines to create a new product, the Do-It-Yourself Ombré hair coloring kit, that leveraged the ombré trend that was surging in popularity. Ocean Spray used Twitter streams to unveil flavors consumers traditionally associate with cranberries to inform novel flavor combinations.

Predicting Consumer Preferences

Predicting consumers’ future fashion tastes was a difficult proposition. Traditional market research methods, such as surveys, focus groups, and interviews, were often inadequate, as consumers were notoriously poor at predicting their future behaviors. Consumers were often unable to imagine changes in fashion, so conducting research with them was futile. Automotive pioneer Henry Ford proclaimed:

“If I had asked people what they wanted, they would have said ‘a faster horse.’“ Or, as the innovative

chef Ferran Adrià put it: “Creativity comes first. Then comes the customer.”41

Relying upon past purchase behavior was also problematic as research in consumer psychology showed that consumers’ preferences were constructed rather than revealed, subject to marketers’ manipulation, unstable over time, and therefore unpredictable. While most consumers believed that they were cognitively in charge of their decisions and thus master of their own tastes and preferences, countless experimental manipulations demonstrated that one’s choices could be swayed by elements of the decision or social context, information framing effects, and the knowledge, ability, goals, biases, and emotional state of the decider.

While taste was defined as an individual’s attitude toward an aesthetic object, fashion, was a social construct that relied upon collective behavior of many people carrying out the same or similar tastes at the same time. A consumer’s individual tastes developed within the context of social influences, including the tastes of others around them, their membership in a variety of subgroups, and the prevailing fashions of the time.42 Distinctions in taste helped mark members of different social classes

and people who occupied the same group tended to share aesthetic preferences.43

What was in and out of fashion was constantly changing, driven both by a self-dynamic process and by tastemakers. Changes happened naturally as people craved newness when yesterday’s fashion had become boring or commonplace. Because people rely on fashion to both fit in with and stand out from others, as soon as a fashion trend broadly permeated society, it stopped being fashionable.

Sociologists theorized a ratchet effect in tastes, where persistent movements in one direction were

suddenly and unexpectedly reversed and followed by movements in the other direction.44 In the short term, new tastes were generally based on existing tastes; thus, year-to-year shifts in fashion were often

8

Predicting Consumer Tastes with Big Data at Gap 517-115

modest. But, suddenly and unexpectedly, the taste changed significantly, ratcheting in a non-linear step change to another direction. Hemlines are an illustrative example. Women’s skirts get progressively shorter as each season embraces the miniskirt, but tries to make it look different from the previous season. However, once miniskirts become ubiquitous, short skirts appear unfashionable, so the next season’s look might suddenly feature long, floor sweeping skirt lengths.

Fashion cycles were often initiated by designers, artists, fashion innovators, and other creative gatekeepers. These tastemakers constantly swept the culture looking for inspiration and ideas to combine in new ways. Their creative inventions often became the raw materials for changing fashion.

Product 3.0 at Gap

The place where Peck was hoping big data could make the biggest difference was in product development. In a strategy he dubbed Product 3.0, Peck promised to,

Combine a clear brand vision with a common operating model...The new brand vision governs every decision in design, merchandising, inventory, and production so that [Gap Inc.] can identify trends, make them relevant to its customers, test them in stores, and respond to demand -- buying more of those that sell and quickly moving away from those that don’t with a goal of fewer fashion misses and markdowns.45

In imitation of his fast fashion competitors, Peck wanted Gap to increase its competence at “combining spotting trends with reading real-time performance and acting faster on that,” using real time data from its registers and e-commerce purchase data to inform what the company produced for inventory going forward. Peck clarified, “[We need to] move forward very quickly in how we bring product to market and the speed that we bring it to market and the flexibility in our inventory…being able to be more predictive and demand driven...[being] more commercial around things that are starting to move up the curve, or [getting] out of a product that is no longer relevant to the customer.”46

This new process was fundamentally different from the traditional process that included creative directors, explained Stefan Larsson, global president of Old Navy,

…The old school variety of designing was to send [the designer] over to Europe, and have them buy samples high end…come back, and then a year and half later you would see it in our stores…this doesn’t work anymore…what we do [now] is that no one creates chance. So no one in our brands believes that they are trend creators. And so what we have design do is to work in a very systematic way to funnel down all of the trends...And

then once you have funneled down the trend, you apply unique design.47

In place of a creative director, each brand’s vision statement served as a filter so that trends could be incorporated consistent with its image. Explained Jeff Kirwan, global brand president, Gap,

Our product aesthetic filters…What are the questions that we are going to ask about every single piece of product that goes into a Gap assortment…which are anchored back to who we are as a brand and what our lifestyle is…and the authenticity of who we are as a brand? If they don’t get through those filters, they don’t show up in the store.48

In this way, each trend cascaded through the entire brand portfolio, showing up in Banana Republic, Gap, and Old Navy simultaneously but interpreted through each brand’s unique prism. Managers across the brands were encouraged to share trend information across the portfolio. Described Peck,

9

517-115 Predicting Consumer Tastes with Big Data at Gap

It starts with what we think is a really good process right at the beginning which filters trend, really systematically across a wide variety of sources, filters trend down to the right ones that we feel are brand right and appropriately commercial…It’s allowed us to be in trends that are happening at the same time in the designer and the premium

contemporary space, it’s allowed us to be into those trends in Old Navy.49

According to Larsson, transforming trends into saleable products quickly was essential in the new marketplace, “you see aspirational trends becoming aspirational much, much faster…suddenly the

value customer is more on trend than any of the brands out there.”50 Via an in-season open program, Gap tried putting a small quantity of goods into stores, waiting to see how customers responded to them, and then quickly producing significant quantities of well-performing products to get them into stores before the end of the same season. Clarified Peck,

We…have traditionally bought the year one season on a grand reveal at a time. And that means making large commitments well in advance of when the product is going to be in the store and well in advance of knowing what the consumer really wants…we have been re-engineering the front end of the business, so that we can buy on a much more continuous basis...every month versus on a quarterly basis...We remain ‘open’ as we get closer to the season and can pivot to buy it to the most meaningful trends.51

Product 3.0 relied heavily on the analysis of customer purchase data. According to Peck, “we’ve also substantially increased our testing of product whether that’s crowd source testing, which we now have validation results in better commercial outcomes, or testing physically in our stores, oftentimes

in stores that are seasonally ahead of where we are so that we can that to inform our buys.” 52 Google Analytics data was also a source of inspiration. A recent fashion trend, men’s jogging pants, was identified early as Gap’s managers noticed that customers were using the search term on its websites, and its progressive adoption across North America was predicted based upon observation of the geolocations of various people using the search term.

To implement Product 3.0, Peck shifted some manufacturing from Asia to the Caribbean to receive items faster. He implemented fabric platforming, buying large quantities of fabric and holding it in inventory so that designs could be quickly created in response to of-the-moment trends. He shortened the time it took for items to go from design to stores and postponed making the final decision on orders until he could incorporate the most recent data trends from limited quantity early releases designed to test the waters. Cutting the development cycle down to 8-10 weeks in some categories enabled Gap to

be much more nimble and responsive to consumer purchasing data.53

Through the changes, Peck was listening closely to data from the voice of the customer program, “I spend a lot of time reading reviews of our products online. And our customers are very clear in telling us what we’re doing well and what we’re not doing well…And those are the things the teams are acting on in both Gap and Banana [Republic], to get the product back to where it needs to be.”54

Peck’s vision to reinvigorate Gap was first and foremost focused on fixing the product. Time and time again, when questioned about the amount of money the company was spending on marketing, he emphasized that the best marketing was a good product. As the company struggled to get its product offering right, he cut back on television advertising and store window merchandising and increased the investment in the company’s digital platforms, explaining, “When you are not proud of our product, you are not going to go out there putting a lot of marketing behind the business...we’ve pulled back and we will continue to do that until we feel like there is an opportunity to really tell the story.”

10

Predicting Consumer Tastes with Big Data at Gap 517-115

He tightened inventory as a way to reduce the need for deep discounting, “The second worst place to be in this business is over-bought [in inventory]. The first worst place to be in this business is over- bought with product that she is not responding to and that yields too many 40% offs…to reduce the promotional expense of this business, the promotional depth and frequency, we will start with product

that we buy tightly that she loves.”55 He continued, “Scarcity is a good thing…the simple reality of pulling the promotion needle out…you have product that she loves and then she finds out that if she didn’t buy it when she went in, it’s no longer available.” However, he recognized the risks, “When you start tightening up on promotion, you are playing a game of chicken with your customers and they try to wait you out.”56

Even once he felt the product had improved, Peck was slow to open the tap on traditional advertising, believing that a discovery process was imperative to Gap’s success: “We don’t plan on making huge marketing investments...I think that’s imprudent. You improve your product, the customer discovers it, you start to see conversion move on the basis of the traffic that you have in the store. Word of mouth, these days, it’s a super powerful form of marketing. If you just think about Instagram and Pinterest and the other social media, that’s a very powerful form of marketing. And that will start to bring traffic in.”57

Shifting the Distribution Model

Another key decision that was on the horizon was whether to partner with Amazon and allow it to sell Gap’s branded products through its online platform. In 2016, 55% of online shoppers started their product search on Amazon,58 which offered over 350 million different products on its platform, about

10% of those in the apparel category.59 Amazon traditionally charged its third party sellers a commission rate of 15%; however, given its size and brand strength, Gap might be able to negotiate a lower fee.