5個零售業大數據帶來巨大收益的實例(譯文)

5個零售業大數據帶來巨大收益的實例(譯文)

原文作者：Chuck Schaeffer 翻譯：馬延超、秦何煜、 王嫣、崔鳳焦 校對：馬延超

大數據正在為零售商們傳遞一些可觀的成果。

Macy說他們的大數據程序是一個關鍵的競爭優勢，指出大數據作為一個強有力的貢獻因素，將零售店的銷售額提高了十個百分點。Sterling Jewelers把上個休假期49%的銷售額增長歸功於大數據。Kroger的執行總裁David Dillon把他們的大數據程序視為自己的秘密武器。

麥肯錫通過五年多來對超過250個業務案例的分析，揭示了將大數據作為銷售和營銷策略中心的公司，他們的投資回報率提高了15到20個百分點的事實。但是儘管有一些驚人的回報和改變零售業遊戲規則的可能，我們仍然要面對許多存在的阻礙。

從和許多零售業高管的會議中，我發現大數據正在得到越來越多的關注，但是同時大部分的高管都在為了共同的挑戰而作鬥爭，像如何將大數據與實用案例結合，如何確定新的類型的數據（通用結構），還有如何得到提高決策質量的大數據。

大數據可以用來做任何事卻並不能被直接使用。它是一種沒有打包解決方案的顛覆性科技。當然，你可以獲取大數據技術，但是如果沒有理解和設想之前的那些隱藏數據是怎樣被得到和應用於商業過程，商業挑戰和機遇，大數據就會變成另外一個無法帶來預期回報並且使用壽命短的冷板凳軟體。

就我的經驗來說，成功部署一個大數據解決方案應該首先從新信息中確定使用實例以及可收益商業決策。

當零售業大數據實例是你的創造性思維的所展現的一個函數時，這比做起來容易一些。為了促進這種思維，來一起看看下面的零售業大數據實例。

連鎖酒店利用大數據來增加預訂量

糟糕的天氣會使旅行的人數減少，當然同時也使得晚上的住宿人數減少了。對於在酒店行業工作的人來說這並不是一個好消息。然而Red Roof酒店把這種趨勢轉化成了他們的優勢。他們意識到被取消的航班會使旅客們處於困境並且需要一個地方來睡覺過夜。這個公司免費提供隨意可得的天氣和航班取消信息，這些信息是聯合酒店和機場位置歸納所得。該公司還建立了一套演算法，從眾多變數中提取天氣惡劣程度，旅遊環境，當日時刻以及航班取消率作為因子計入該演算法。 通過大數據的洞察力以及對於實例中將要使用移動設備旅客的識別，該公司利用Search，PPC，SoMoLo等搜索引擎給滯留旅客的目標移動設備發送廣告，使消費者能更容易的訂到周圍的酒店。

大數據所帶來的回報是引人注目的，利用大數據我們可以統計出平均每天有1%—3% 的航班會取消，摺合平均每天有150到500個航班取消，或則說每天約有25000到90000的滯留旅客。紅屋頂酒店運用大數據和地理的移動營銷活動，使它的業務與從2013年到2014年增長了10% 。

披薩連鎖店在天氣不好的時候能賺更多的錢

類似於上面的例子，當消費者在惡劣天氣或者停電的情況下無法做飯時，一個披薩連鎖店就可以使用一個移動應用程序和移動營銷技術提供優惠券來刺激他們消費。這個移動和定位營銷活動大概得到了20%的反應率。

音樂經銷商運用大數據來指定需求計劃

EMI唱片公司可以用大數據來測量和預測產品需求。唱片公司發布音樂以後，可以利用自己的社交網路統計音樂點擊率，另外也可通過流行音樂流媒體伺服器，歌曲識別應用程序或「第二屏」社交媒體排序獲得第三方監聽模式數據。這些數據是人口，位置和亞文化群的聚合數據，音樂發行商可以很放心利用這些數據提供精確定位廣告以及預測產品需求。對於其他零售商也可以通過社交網路獲得聚合數據，以此來了解他們的新產品在新的或現有的市場的銷售情況，甚至可會獲悉公眾對他們的產品評價以及他們公司的信譽度。

金融服務公司運用大數據獲得新客戶

由於在新客戶群體中獲得贏利率較低，金融服務公司也開始轉向運用大數據，以便更好地識別哪些新客戶會有更大的投資可能性。金融公司客戶人口統計數據以及攜帶的第三數據是從eBureau購買的。數據服務提供消費者的職業、收入、年齡、零售歷史和其他相關因素，以此得知他們會在本公司消費的幾率。增強的數據集應用於一個演算法，該演算法可以識別對哪些新客戶有必要再進行額外的投資，對哪些新客戶不必再進行投資。應用該方法取得的成果是新客戶獲得率增加11%,同時公司銷售相關費用降低了14.5%。

零售商創建懷孕檢測模型

最近一個非常有名的零售商運用大數據的例子：零售商Target使用寶寶派對註冊以及客戶身份認證，來確定客戶是否懷孕。Target客戶身份認證是一個獨特的消費ID，它可以追蹤購買歷史，信用卡使用，調查回復，客戶支持事件，郵件點擊率，網站訪問等。該公司通過購買人口數據來補充消費者活動，這些數據包括：年齡，種族，教育，婚姻狀況，孩子數量，估計收入、工作歷史和生活事件 如你最後一次移動或如果你已經離婚或宣布破產。通過比較在嬰兒淋浴註冊的顧客的購買歷史，零售商發現她們在懷孕期間購物習慣的改變情況。例如，在第一個20周，孕婦開始採購補充鈣、鎂和鋅的產品。在懷孕中期，孕婦開始購買更大的牛仔褲和更大數量的洗手液，無香味的乳液、香水免費肥皂和棉花球，通常是買大袋的棉花球。總的來說，零售商可以確定孕婦經常購買的25種產品。

通過查看所有顧客目標的消費行為可以識別孕婦甚至是那些未發現懷孕的女性目標。目標是用這一發現創建了懷孕預測模型可以給顧客提供懷孕預測。然後零售商就可以傳播嬰兒產品根據懷孕的不同時間階段推廣給特定的細分顧客，這樣不僅是那些女性購買嬰兒產品也可以知道重要的生活事件可以改變顧客的整體消費習慣，目標是可以讓經濟消費從分析的時間開始，2002年的440億美元增長到2010年的670億美元。當零售商們還沒有公開地關注這一計劃。目標的負責人，GreggSteinhafel已經記錄分享投資者的數據「著重關注項是可以吸引特定顧客段的項目和分類 ，如媽媽和孩子」很大程度上決定了零售商的成功。

是大數據還是舊經驗

這些零售商的大數據的樣例可以被多種方式推測，使用天氣模型預測、商店銷售結合互聯網搜索數據、網頁瀏覽模式、社交網路和工業預報去預測產品趨勢，預告需求，精確定位顧客和優化價格和促銷。

理解你的產品銷售與其他未發現的因素之間的關係，這些因素有很多，比如天氣、潮流文化、社會媒體潮流、你的競爭者和消費者的觀點，這可以讓你充分利用和挖掘有特定行為環境中的事件以提升財務績效。

零售商利用大數據可以設計更能被消費者接受的產品，更好地參與和應對市場變化，更好地吸引消費者。這意味著更好地數據可以有更少的缺貨，更高的訪問購買率，更好地改變和採取商品結構的規劃。

大數據並不只是服務大公司

零售業的思想領袖加里霍金斯認為，大數據可以創造一個零售壟斷。在哈佛商業評論中，霍金斯指出了大數據可能「擊潰除了最大的零售商所有其他零售商」。

他認為大型零售商有更大的IT預算和資源，可以利用大數據的機會，增加市場的優勢，把較小的零售商基本上壓迫到「便利店的角色」。

儘管霍金斯支持的論點以及大數據的確是切實提高營銷地有利條件，可以提高產品的可用性和用戶體驗，從而超越零售業的其他競爭對手，但是我堅信新的零售強弱順序將更少取決於零售商的IT預算而更多地傾向創新性和敏銳度。零售業正在發生深刻的變化，小企業往往比大的零售商更敏捷。正如達爾文教我們「不是最強也不是最聰明的物種可以生存。而是最能適應改變的那個可以笑到最後」。

原文部分

Big Data in Retail Examples

By Chuck Schaeffer

5 Retail Big DataExamples with Big Paybacks

Big data is deliveringsome big results for retailers.

Macys says that its bigdata program is a key competitive advantage and cites big data as a strongcontributing factor in boosting the department stores sales by 10 percent.Sterling Jewelers attributes a 49 percent increase in sales during the lastholiday season to big data. Kroger CEO David Dillon refers to his big dataprogram as his "secret weapon."

McKinsey analysis of morethan 250 engagements over a five year period revealed that companies that putdata at the center of the sales and marketing decisions improved theirmarketing ROI by 15 to 20 percent.

But despite some impressivepaybacks and what may be a game changer in the retail industry, plenty ofobstacles remain.

In meeting with a numberof retail executives Ive found that Big Data is getting a lot of interest, butmost of these executives struggle with some common challenges – suchas how to align big data with use cases, how to identify new types of(generally unstructured) data and how to harvest big data for improved decisionmaking.

Big data is anything butout of the box. This is a disruptive technology without packaged solutions.Sure, you can acquire big data technology, but without understanding andhypothesizing how previously hidden data can be harvested and applied tobusiness processes, challenges or opportunities, big data becomes anothershelfware solution with a disappointing payback and short lifespan.

In my experience,successfully deploying a big data solution begins by identifying use cases andbusiness decisions which benefit from new information. This is easier said thandone as retail big data use cases are a function of your creative thinking. Tostimulate that thinking, consider the following retail big data examples.

Hotel Chain Uses Big Datato Increase Bookings

Bad weather reducestravel, which then reduces overnight lodging. That』s notgood news if you』re in the hotel business. However, Red Roof Inn turned thistrend on its head. The hotel chain recognized that cancelled flights leavetravelers in a bind and in need of a place to sleep overnight. The companysourced freely available weather and flight cancellation information, organizedby combinations of hotel and airport locations, and built an algorithm whichfactored weather severity, travel conditions, time of the day and cancellationrates by airport and airline among other variables. With its big data insights,and recognition that travelers will be using mobile devices for this use case,the company used Search, PPC and SoLoMo mobile campaigns to deliver targetedmobile ads to stranded travelers and make it easy for them to book a nearbyhotel.

This big data paybackis compelling. Flight cancellations average 1-3% daily, which translates into150 to 500 cancelled flights or around 25,000 to 90,000 stranded passengerseach day. With its big data and geo-based mobile marketing campaigns Red RoofInn achieved a 10% business increase from 2013 to 2014.

Pizza Chain Earns MoreDough in Bad Weather

Somewhat similar to theabove example, a pizza chain uses a mobile app and mobile marketing techniquesto deliver coupons based on bad weather or where power outages leave consumersunable to cook. This mobile and location-based marketing campaign achieves a20% response rate.

Music distributorApplies Big Data for Demand Planning

Record label EMI usesbig data to measure and forecast product demand. After distributing or leakingmusic, the company measures consumption on its own social networks andadditionally acquires third party listening pattern data from popular musicstreaming services, song identification apps or second screen social mediacollators. The data is aggregated by demographics, locations and subculturesand helps the music distributor deliver pinpoint advertising and forecastproduct demand with a high confidence level. This concept is applicable toother retailers who can also aggregate feeds from social networks to build anunderstanding of how new products will be received by new or existing markets,or even how their products and company reputation are perceived among thepublic.

Financial ServicesCompany Scores New Clients

After incurring low winrates for new client acquisitions, a financial services firm turned to big datain order to better identify which new client opportunities warrant the mostinvestment. The company supplemented its customer demographic data with thirdparty data purchased from eBureau. The data service provider appended saleslead opportunities with consumer occupations, incomes, ages, retail historiesand related factors. The enhanced data set is then applied to an algorithmwhich identifies which new client leads should receive additional investmentand which should not. The result has been an 11 percent increase in new clientwin rates while at the same time the firm has lowered sales related expenses by14.5%.

Retailer CreatesPregnancy Detection Model

In a near infamousretail big data example, retailer Target correlated its baby-shower registrywith its Guest ID program in order to determine when a shopper is likelypregnant. Targets Guest ID is a unique consumer ID that tracks purchasehistory, credit card use, survey responses, customer support incidents, emailclick-throughs, web site visits and more. The company supplements the consumeractivities it tracks by purchasing demographic data such as age, ethnicity,education, marital status, number of children, estimated income, job historyand life events such as when you last moved or if you have been divorced orever declared bankruptcy.

By comparing shopperswho registered on the baby shower registry with the purchase history from theirGuest ID, the retailer discovered changes in shopping habits as the womanprogressed through her pregnancy. For example, during the first 20 weeks,pregnant women began purchasing supplements like calcium, magnesium and http://zinc.In the second trimester, pregnant women began buying larger jeans and largerquantities of hand sanitizers, unscented lotion, fragrance free soap and cottonballs; often extra-big bags of cotton balls. In total, the retailer identifiedabout 25 products purchased by pregnant women.

By applying thesepurchase behaviors to all shoppers Target was able to identify women who werepregnant even though these women had not notified Target – oroften anybody else – they were pregnant. Target used this discoveryto create a pregnancy prediction model which assigned a pregnancy predictionscore to shoppers. The retailer was then able to distribute baby productpromotions to a very specific customer segment, timed to stages of pregnancy,and the financial results were off the charts. Not only did these women makenew baby product purchases, but knowing that significant life events change aconsumers overall shopping habits, Target was able to grow its revenues from$44 billion in 2002 when the analysis started to $67 billion in 2010. While theretailer does not publicly comment on this program, Targets president, GreggSteinhafel, is on record sharing with investors that the companys"heightened focus on items and categories that appeal to specific guestsegments such as mom and baby" heavily contribute to the retailerssuccess.

Notwithstanding theconsumer privacy and public relations considerations which must be deliberated,this is a powerful lesson for retailers.

Go Big or Go Home

These retail big dataexamples can be extrapolated in many ways — from using weatherpatterns to predict in-store sales to combining data from web search trends,website browsing patterns, social networks and industry forecasts to predictproduct trends, forecast demand, pinpoint customers and optimize pricing andpromotions.

Understanding thecorrelation between your product sales and otherwise undetected factors such asthe weather, pop culture, social media trending, your competitors and consumersentiment can allow you to tap into these environmental events with specificactions that lead to improved financial performance.

Retailers that leveragebig data will design products that are more embraced by consumers, betteranticipate and respond to market shifts, and engage consumers with predictableresults. This means fewer stockouts, higher visit to buy ratios, bigger basketsizes and other performance measures that can be improved with better data.

Big Data Not Just ForBig Companies

Retail thought leaderGary Hawkins suggests that big data may actually create a retail oligopoly.Writing in the Harvard Business Review, Hawkins poses the likelihood that bigdata may "kill all but the biggest retailers." He suggests that largeretailers, with their larger IT budgets and resources, can capitalize on thebig data opportunity, increase market dominance and essentially relegatesmaller retailers to "the role of convenience stores."

Notwithstanding Hawkinswell supported argument as well as big datas very real opportunity to improvemarketing, product availability or the customer experience, and therebyoutperform retail competitors, its my strong belief that the new retailpecking order will be less determined by the size of the retailers IT budgetand more by the retailers propensity toward innovation and agility. The retailindustry is incurring profound change and smaller businesses often show moreagility than larger retailers. As Darwin taught us "It is not thestrongest of the species that survives, nor the most intelligent that survives.It is the one that is most adaptable to change."