Big Data Analysis in Marketing: Harnessing Customer Data for Better Business Strategies

E-book Services Use Data to Tailor Content to Readership

With the amount of data available to companies doubling every year, new sources of data, and innovations in data collection, possibilities for marketers to identify market niches and finely tune campaigns are boundless. In the e-book market, for example, three reading subscription service startups—Scribe, Oyster, and Entitle—aim to turn a profit by discovering exactly what makes readers tick.

A flat monthly fee gives users unlimited access to a broad selection of titles from these companies’ digital libraries. Like Barnes and Noble and Amazon, the newcomers will collect an assortment of data from their customers’ digital reading devices (e-readers, tablets, smartphones), including whether a book is completed, if pages are skimmed or skipped, and which genres are most often finished. These subscription services intend to disseminate what they have learned. The idea is that writers can use it to better tailor their work to their readership, and book editors can use it to choose which manuscripts to publish.

When customers sign up with these services, they are informed that some of their data will be collected and used but assured that their identities will be protected. Large independent publisher Smash words is enthusiastic about the value of such data to the authors who use its platform to self-publish and distribute their work. Many contemporary authors have already explored the feedback opportunities available through their own Web sites, social networking sites, and Goodreads, a user-populated database of books, annotations, and reviews now owned by Amazon. The subscription services will take this type of market research to a more quantifiable level.

Preliminary data analysis has already revealed that as the length of a mystery novel increases so does the likelihood that a reader will skip to the end to discover the resolution. Business books are less likely to be finished than biographies, most readers complete just a single chapter of a yoga book, and some of the quickest reading is recorded for romance novels, with erotica leading the pack. Shorter chapters entice readers on e-readers, tablets, and smartphones to finish a book 25 percent more often than books with long chapters.

But does book completion translate to book sales? And how will this knowledge impact the creative process? Will quality be negatively impacted to satisfy reader preferences? Before any of these questions can be answered, authors will need access to comprehensive data. And that depends on the large publishing houses signing deals with the subscription services. After nearly two decades of market disruptions spearheaded by Amazon, publishers are not flocking to supply titles. So far, only Harper Collins has signed with Oyster and Scribe, while Random House, Penguin, and Simon & Schuster remain on the sidelines.

In the airline industry, nearly all carriers collect passenger data, but some are aggressively pursuing data mining to personalize the flying experience. Previously unlinked data sets can now be consolidated to build comprehensive customer profiles. Cabin crews equipped with tablets or smartphones can identify the top five customers on the plane, passengers with special diets or allergies, seat preferences, newlyweds embarking on their honeymoon, and customers whose luggage was misplaced or who experienced flight delays on their previous flights. In-flight browsing history and Facebook likes are even used to fashion relevant marketing pitches.

This “captive audience” aspect of air travel in conjunction with the sheer volume of information airlines collect presents a unique opportunity to marketers. Allegiant Travel company has already been able to sell show tickets, car rentals, and helicopter tours to Las Vegas travellers. United Airlines’ revamp of its Web site, kiosks, and mobile app, along with its data integration initiative, have enabled it to target flyers predisposed to upgrading to an economy plus seat.

Not all customers are pleased. A user on Delta’s FlyerTalk forum complained that a link from the new DL.com Web site led to a personal profile that included a lot more than her miles accumulated and home airport. Annual income, home value, and the age ranges of her children were included along with expected data such as amount spent on airfare, hotel preference, and type of credit card. The resulting negative publicity prompted Delta to apologize, but it defended its use of demographic data and data not covered under its privacy policy. Credit-card partner American Express had supplied some data, as allowed under the policy. Global information services group Experian supplied the rest, unbeknownst to consumers.

These data-driven marketing approaches are not flawless. Even customers who accept the inevitability of profiling are miffed when they receive unsuitable offers based on faulty personal information. A Qantas survey of frequent fliers found that most customers want a line drawn between data collection to facilitate useful offers and data collection that is too intrusive. British Airways crossed the line with its “Know Me” program. Google Image searches were used to identify VIP customers as they entered the airport and first class lounge. The practice has since been discontinued. Customers can opt out of British Airways personalization services—but not its data collection. Upon request, a note is added to the customer profile, which nonetheless continues to grow. None of the carriers currently allow customers to opt out of their data programs.

As car companies explore their Big Data opportunities, customer privacy will become an issue for them as well. Ford Motor Company began exploring how integrating databases and using complex algorithms could lead to increased sales three years ago when it developed a program for its dealerships to more closely match car lot inventory to buyer demand. Using buying trends, local and national vehicle supply, and current car lot inventory, Ford devised a program to make purchasing recommendations to dealers. Not only did vehicle turnover rate improve, but net price—the price a consumer pays minus the manufacturer subsidy—rose, fuelling an upturn in Ford’s profits.

But Ford is thinking even bigger. Performance monitoring using vehicle Internet connections to collect fuel economy, mechanical failure, and other safety and performance metrics could soon be used to improve product engineering. What’s more, on board connections can be used to message drivers about potential breakdown issues, perhaps heading off an expensive recall. Since Ford estimates that by 2016, up to a third of all its consumer communication will occur inside vehicles, possibilities abound. Leased vehicle usage data could inform end-of-lease marketing pitches; driving pattern, schedule, and driving manoeuvre data could suggest routes most compatible to a driver’s habits; car location data could be sent to traffic management systems to control stop lights; data from networked cars could alert other drivers to hazardous conditions and traffic jams, and current car value and payment data can advise drivers of their optimal trade-in date.

It’s not hard to foresee the privacy issues that could come into play as drivers realize that not only their location, but their every movement inside their vehicle is being tracked. There are implications for law enforcement—traffic tickets and accident blame attribution. Balancing privacy dilemmas with convenience, security, and expediency of transactions will be the challenge going forward for all companies as they explore emerging big data analysis capabilities.

Sources: Tim Winship, “Big Brother Unmasked as ... Delta Air Lines,” smartertravel.com, January 28, 2013; Jack Nicas, “When Is Your Birthday? The Flight Attendant Knows,”Wall Street Journal, November 7, 2013 and “How Airlines Mine Personal Data In-Flight,” Wall Street Journal, November 8, 2013; David Streitfeld, “As New Services Track Habits, the E-Books Are Reading You,” New York Times, December 24, 2013; Ian Sherr and Mike Ramsey, “Drive into the Future,” Wall Street Journal, March 7, 2013.

Case Study Questions

1. Describe the kinds of data being analysed by the companies in this case and describe any methods, techniques and IT solutions to analyse the data.

2. Provide an assessment of how this fine-grained data analysis improve operations and decision making in the companies described in this case and the business strategies they support.

3. Are there any disadvantages to mining customer data? Provide an analysis of your answer.

4. Critically assess your feelings about airlines mining your in-flight data. Is this any different from companies mining your credit card purchases or Web surfing? Discuss.

Secondary Research Level HE4 - It is expected that the Reference List will contain between five and ten sources. As a MINIMUM the Reference List should include one refereed academic journal and three academic books

Specific Assessment Criteria

First class- 70%+: Demonstrates an exceptional knowledge/ understanding of theory and practice for this level through the identification and analysis of the most important issues.  Makes exceptional use of appropriate arguments and/or theoretical models. Presents an analysis of the material resulting in clear, logical and original conclusions.

Second class Upper 2:1 60 - 69%: Demonstrates a very good knowledge/understanding of theory and practice for this level through the identification and summary of key issues.  Uses sound arguments or theoretical models. Presents a clear and valid discussion of the material. Clear, logical conclusions.

Second class Lower 2:2 50 - 59%:  Demonstrates a good knowledge/understanding of theory and practice for this level through the identification and summary of some key issues.  Presents largely coherent arguments. Some issues and theoretical models expressed in simplistic terms. Conclusions are fairly clear and logical.

Third Class 40 -49%:  Demonstrates an adequate knowledge/understanding of theory and practice for this level. An attempt is made to identify key issues.  Presents basic arguments, but focus and consistency lacking in places. Some issues may lack clarity, and/or theoretical models expressed in simplistic terms. Conclusions are not always clear or logical