Mary loves the lines of the song, “Every breath you take. Every move you make … I’ll be watching you. Every single day. Every word you say.” Companies may well be singing that song today, as they track Mary every single moment of her life. Using techniques of data analytics, they know more about her than even her parents.
In the movie Minority Review, as the hero walks into a clothing store, his eye is scanned by a computer and he is immediately greeted by name. The computer also accesses his purchase history to make suggestions and reminds him of the products that he might need.
This is indeed what shopping is evolving into, thanks to a world in which we leave data in everything that we do online. By analyzing this data, companies know quite a bit about us, including our buying habits. Data mining and analytics gives power to companies to identify us and to make product offers to us, just like in the movie.
True, companies have always looked for data about their customers. They want to know who they are, where they live, and analyze their profiles to precisely meet their needs. But getting data about consumer behavior has not been easy. The traditional method consists of arming researchers with questionnaires and getting them filled through surveys or focus groups. This is not a very accurate or dependable method, since consumers do not like to be questioned about their decisions or give their personal data. Worse, they are often not aware of their own motivations for buying things. Brown (2001) summed up the frustration of getting customer data, “Consumers don’t know what they want. They never have. They never will. The wretches don’t even know what they don’t want.”
Indeed, Henry Ford’s famous remark, “If I had asked my customers what they wanted, they would have said a faster horse,” shows that consumers operate from existing mindsets and seldom see things from a more innovative angle.
Companies have looked for better ways to track their customers. Now they have that tool: a data deluge that yields almost everything about a consumer through data mining and analytics.
Data Deluge
Called big data, it is a powerful tool to know not only about the consumers but also about their behavior and state of mind. It consists of the huge volumes of data that are being created at various points in the global economy, such as:
The result is an immense amount of data, streaming in from different sources. “Customers are telling us all about themselves, each day, every day. We now create as much information every 48 hours as we did from the dawn of civilization up to 2003,” says Mckinsey in its report, The Data Driven Life (2013).
How companies use this enormous resource depends on their vision and capabilities. Big data gives information in real time: Every device, shipment, and consumer is providing data on happenings as they occur. Data analytics is the process of inspecting, cleaning, transforming, and modeling data with the purpose of discovering patterns and drawing conclusions. It promises to change the way business is done.
The task is not easy because the data come from varied sources and in differing formats, structured and unstructured. Davenport (2014) explains that companies have to develop abilities to integrate internal transactional data with existing databases, comments on social media, credit and loyalty cards transactions, the sites that people visit, and also with automatic sensors that give information about the context of purchasing. All these data sets will be in different formats and a high level of expertise is needed to clean and integrate them with company databases.
Big data is indeed valuable to companies. Davenport (2006) writes, “They know what products their customers want, but they also know what prices those customers will pay, and how many items each will buy in a lifetime, and what triggers will make people buy more.” Companies can today identify their most profitable customers, increase the likelihood that their products will be liked, and increase customer loyalty, all through analytical methods.
Making Use of Data
The data analysis ability is not easily acquired, simply because the quantum of data available is huge. Loveman (2003) dubs it as a future diamond mine, where millions of individual transactions are available. Barton and Court (2012) explain three steps as to how companies acquire abilities to make use of this enormous data:
Data analysis involves detective work, by uncovering insights from the streaming data. The detective follows the data trails left by people as they do things.
Consumers’ Data Trails
Davenport (2013) writes that powerful data capabilities will apply not just to a company’s operations but also to its offerings—to embed data smartness into the products and services customers buy. Data analytics and optimization can be built into every business decision, leading to better customer engagement and, consequently, to higher profits. McKinsey estimates that a retailer using big data to the full has the potential to increase its operating margin by more than 60 percent. How Apple, Google, and eBay use data trails is illustrated in Exhibit 5.1.
Getting Personal with Data Mining
Companies are already using data mining techniques to offer personalized products and messages pinpointed to customers. This exhibit shows how three companies—Apple, Google, and eBay—effectively use data to understand consumer behavior.
Apple: Apple’s app network connects brands with users where they are most engaged. Its iAd service, which is built into its iOS, helps in precise targeting. Users are analyzed in terms of website preferences, internet usage behavior, apps used, a person’s music, TV and audiobook preferences, devices, demographics, and network. As a result, each ad is shown only to people who will see it, through the apps they like to use. Demographic data and people’s unique interests and preferences are used to connect the most relevant brands.
Google: Google collects and analyzes customer data and uses data mining techniques. Its philosophy, as explained on its website, is to focus on the user and all else will follow. Information about IP addresses, search terms, and browser types are collected, which helps the company to build a precise profile of people. Data provides Google with information about geographical location with a built-in weather forecast. Through its mail analysis, it provides further insights. Data about music and video preferences are collected via YouTube, and Google Maps gets information on travel destinations and plans.
eBay: The eBay stores more consumer data in its central data warehouse, which is used to get an integrated customer view to generate innovations. In-house departments use the data by building sand boxes and compare it with external data.
Data analysts follow trails left by people, picking up details of their habits and purchases. It is like following a fireworks display—the eye knows the light left behind by a bright light as it moves across the sky. Following this data trail in a manner of detectives—dubbed as digital breadcrumbs by Alex Pentland (2014)—can give accurate pictures of consumer habits. Data analytics help to link behavioral, transaction, and customer interaction insights into the minds of consumers. Pentland calls this social physics—the combination of data reveals how humans interact and how ideas spread. The millions of data points measured over a long period of time in real settings are the living laboratories that help us monitor human behavior like never before. For example, Pentland finds how behavior changes when people fall ill; using data from websites and phones, companies can tell that certain people were going to get flu even before they get it. Similarly, advanced analytics help companies to know when purchase decisions are most likely to be made. Some companies have already made strides in this capability (Exhibit 5.2).
Predictive Analytics
Retailers have always collected information on their customers. Target too does that, assigning each shopper a unique code that lets the company track a person’s purchase history. The company tracks all interactions—such as using a credit card, filling out a survey, calling the customer helpline, opening an e-mail, or visiting its website—to workout detailed consumer profiles. This data is linked to demographic and geographic information along with bought data about ethnicity, job history, magazines subscribed, banking information, social networking data, and so on to work out detailed product preferences. This information is analyzed and throughpredictive analytics the company can figure out what you want and when. Specific mailers are then sent to customers, and sales of the products advertised skyrocket as a result.
Target could link purchase of certain items like unscented lotion, health supplements, scent-free soap, and extra-big bags of cotton balls to pregnant ladies. The company could also figure out a lady’s due date.
New Scientist (Rutkin 2014) reports an algorithm can track flu cases across the United States by mining data. The program is able to predict flu outbreaks in the United States by monitoring what people search for on Wikipedia. The program downloads publicly available information every hour about how many people across the country accessed the pages. By doing so, they could accurately predict the number of cases in the country two weeks earlier and with a difference of just 0.27 percent.
Google Analytics helps website owners to know important facts about visitors. Users can set up their own dashboards and track which online campaigns bring the most traffic and conversions, where the visitors are located, and find out what people are searching for and see what they click on. Google Analytics is a virtual powerhouse to understand website traffic and to identify users.
This is not all. Researchers say they can pinpoint personality traits by merely studying user-generated text such as e-mails and social media posts. Eben Haber and his team at IBM has developed software that studies streams of tweets from social media and correlates them to personality, values, and needs. The software can map someone’s personality from just about 200 tweets, which helps predict their purchases. Extroverts, for instance, are attracted to advertisements that offer excitement by buying a mobile phone. They also prefer Coca-Cola to Pepsi and Maybelline cosmetics to Max Factor. Pepsi, on the other hand, is preferred by agreeable people.
Technologies are being refined all the time to get deeper insights. Whereas earlier, computers could only look at quantitative data, now all kinds of data are analyzed.
The Economist (2014) reports of face-recognition technologies which can extract information about people by tracking the images available online. The Elastic Bunch Graph Matching technique (EBGM), creates a 3D model from two-dimensional images, which can be used to match with any other images. The technology is used to go over video feeds from the growing number of cameras from restaurants, streets, traffic lights, offices, and other sources. “We may not be far from a world in which your movements could be tracked all the time, where a stranger walking down the street can immediately identify exactly who you are,” says another report in The Economist (2013a).
Another technology is IBM Watson. It helps companies track the natural language that is being posted all the time, as shown in Exhibit 5.3. This represents a giant step toward analyzing the millions of comments that are posted by consumers daily.
IBM Watson: Big Data Analysis for Marketing
So far, computers have crunched numbers to help in marketing decision making. Managers and researchers must have wondered how nice it would be if a computer could go through the enormous natural language data produced daily—on blogs, Internet conversations on various platforms, Facebook comments, and so on—and produce intelligence about the mood and concerns of consumers. Reading and tracking all this data by humans is impossible, simply because of the size involved. Companies have no way to track what is written on each and every site globally.
IBM Watson does just that. Named after IBM founder Thomas J. Watson, it uses natural language processing and analytics, and processes information just the way people think. This ability represents a major advantage in an organization’s ability to analyze and respond to big data. Watson’s ability to answer complex questions posed in natural language quickly is a great boon for companies to help assess and predict consumer needs based on what they share on the Internet.
Another advantage of IBM Watson is that users can get insights through visual representations without the need for advanced analytics training, through its natural language interface. The service removes impediments in the data discovery process, enabling business users to quickly and independently uncover new insights in their data. Watson Analytics automatically prepares the data, finds the most important relationships, and presents the results in a visual format for managers.
The ability of Watson to evaluate social media data, publicly available data, and proprietary data from clients and continuously learn from that information is becoming invaluable in many industries. In marketing, a company could train Watson to understand its customers, and then use predictive models to recognize new products or services that their customers will buy. It can analyze and predict consumer sentiment and needs, evaluate new products, or predict which advertising will be effective. For example, it can figure out whether a movie trailer is going to positively affect an audience or not. IBM is now allowing customers to use Watson as a service and has opened it up to developers to build Watson apps.
Source: Courtesy of International Business Machines Corporation, © International Business Machines Corporation. Reprinted with permission from IBM Corporation.
The natural language processing and learning ability of IBM Watson can:
Predict new trends and shifting tastes: Watson processes enormous amounts of consumer data and learns as it goes along. It does not forget anything: data from credit cards, sales databases, social networks, location data, web pages, and it combines all the information to make high-probability predictions and even sentiment analysis. Interestingly, Watson can recognize irony and sarcasm—and find out the intended meaning. This ability helps it to quickly analyze large sample sizes to determine whether a product offering or clothing line will be accepted by consumers.
Analyze social conversations and thus generate leads: Watson can predict what information is most important and go through reports about the industry, competitors, partners, and customers and make recommendations on how to act on it. For example, if it finds some people discussing problems, it can match those with the company’s product and notify the sales team.
Determine whether a new innovation will sell or not: Watson can learn from one domain of knowledge and make high-probability predictions in another, and this helps it to understand whether a new innovation will sell or not. It crunches the company’s current market and customer base to provide success probabilities, as well as provides a sharp picture of the opportunities and threats.
Computer-calculated and automated growth hacking: Growth hackers seek to maximize conversions on e-mails, websites, social media, online content, or other digital media. Watson does this to measure and optimize digital content, ads, website pages, and even a company’s product to maximize customer growth.
Measurement of ad effectiveness and media planning: Nielsen has partnered with IBM Watson to improve measurement of ad effectiveness and media planning. Its data-parsing system is used for CRM, customer call centers, and other purposes through its Watson Engagement Advisor offering. Questions like who are my best prospects in a category and how much should I budget for next year are answered by taking into account data that might be normally ignored. By identifying prospective customers, ad effectiveness will become more efficient.
Data trails are analyzed in a series of five steps. Starting with finding out what people need, it also helps to identify focused advertisements, tailor-made promotions, reinvents purchase experience, and finally encourages positive WOM (Figure 5.1).
Figure 5.1 Analyzing data trails
How these are achieved is discussed in detail here.
Identifying Consumer Needs
Big data helps right from the beginning, that is, by identifying consumer needs. Predictive analytics track data from various sources to tell companies what people need before they know it.
The information that earlier took years to collect and analyze is available easily today. Take a look at how data trails lead to the consumer’s mind:
All it requires is a good detective, or a data analyst, who can make sense of these trails and combine them with data streaming in from various connected devices.
Focused Advertising
In advertising, companies create microfocused targeted messages that are delivered at the perfect moment, when the consumer is receptive toward receiving such ads. Data analytics help discover which ads will be most liked and should be served to individual customers.
This is a big help to companies because traditional ways of advertising do not work anymore. Mass media is fragmented today and losing its effectiveness, which is described by Bob Garfield as The Chaos Scenario (2009) in the book of the same name. “Newspapers, magazines and especially TV as we currently know them are fundamentally doomed,” he predicts as audience shrinks globally. Connected consumers prefer to fiddle with their connected devices rather than listen to what the media is telling them. How marketing communications can be made more relevant is discussed in Chapter 6.
Devising Offers
Just as ads can be precisely targeted, so also can products. Companies track what a consumer likes across the Internet and are able to modify their offers. Using consumer data with online tracking, nowadays companies are able to create customized offers that immediately attract customers. Davenport, Mule, and Lucker (2011) call these next best offers (NBOs). Based on social, mobile, and location information of customers, “companies are beginning to craft offers based on where a customer is at any given moment, what his social media posts say about his interests, and even what his friends are buying or discussing online,” they write.
New methods of understanding customers are being developed. One such method is Real-time Experience Tracking (RET), which helps companies to know how and which touch points influence the consumer decision journey most significantly. The method involves asking a consumer panel to send four-character text messages through their mobile phones when they come across a brand or a competitor during their purchase process. The short message sends the codes for brand, touch point type, positivity or the feeling of the customer, and a rating of persuasiveness. Qualitative details are captured when respondents are surveyed about their brand-attitude changes.
RET helps companies to track effectiveness of ads, track competitor stimuli, and more importantly, helps understand how consumer attitudes change over touch points during the purchase journey. The sequence of text messages of RET reveals insights that traditional surveys cannot reveal.
Dialog, access, risk–benefits, and transparency are emerging as the basis for interaction between the consumer and the firm, which helps in reinventing the purchase experience for customers.
Reinventing Purchase Experience
Traditional retailers had a big advantage over today’s companies—they knew about their consumers. A person running a small store could recognize most customers, engage them in small talk, and make gentle product recommendations. Companies and retailers do not have this luxury today, but now big data promises to give them the same advantage. As customers walk into a store, their mobile phones give away their identity and location. If they are already using the store’s app, the company immediately knows their profiles and purchase history. This data can be used to enhance consumer experience and also to generate loyalty. Companies can now deliver convenience just like the traditional small retailer in the following ways:
The retail store then becomes much more than a place to browse and buy. It becomes an opportunity to deliver a personalized experience; the store and the online presence become one entity. Argos, a UK retailer, has a digital first strategy through which it places self-service iPads which customers use to browse product videos and reviews. The store provides free Wi-Fi to help customers use their own devices to check prices and products. While customers are happy to get free Wi-Fi, the store benefits as it gets tracking data to help in personalized promotions. Dynamic digital screens have replaced traditional posters and display screens. It offers a 60-s fast track service for customers who have ordered online.
Big data provides an indispensable arsenal for companies consisting of:
Postpurchase: Encouraging WOM
WOM is one of the most important factors in purchasing decisions, especially for a first-time purchase. Its influence is growing with the digital revolution: Today every experience or comment is shared with friends. Digital WOM is read by many people, and product reviews and opinions play an important role in purchase decisions: It is no longer an intimate, one-on-one conversation, but one-to-many communication because clever reviews and comments have the tendency to go viral.
Social data analytics involves real-time tracking of what is posted on social sites. Moreover, it requires going beyond the surface and analyzing the context in which the comments are made: The analysis has to uncover sarcasm and statements using double meanings. This calls for sophisticated data and text analytics. IBM reports that such tools are available today, which are as follows:
Social media is the single largest source of WOM. But monitoring it and encouraging consumer engagement is tough because it spans the three dimensions: Gartner calls big social data as consisting of three Vs: high-volume, high-velocity, and high-variety information. Social data is huge: more than a billion consumers spending about a quarter of their time on social media platforms. The speed at which the social data flows is remarkable too: It is available in real time or nearly real time. It is constantly refreshing, providing quick and accurate insights into what people are doing or liking, if only we have the means to listen. And then, there is variety. Social media provides different kinds of data such as images, videos, and text. Companies can encourage WOM through the following ways:
While data analytics is a goldmine for companies, there are some limitations to it as well. It is important to keep in mind what data cannot reveal as well.
What Data Cannot Reveal
Many writers explain that big data analysis is a paradigm shift and enables companies to move away from gut feelings and toward data-driven decision making. While this may be true, can the human element in marketing be truly ignored?
Lee and Sobol (2012) write that human behavior is nuanced and complex, and data can provide only part of the story. Desire and motivation are influenced by a number of factors that have to be understood in context and conversation. “Data can reveal new patterns that point a firm in the right direction, but it can’t indicate what to do once there. It reveals what people do, but not why they do it,” they write.
For instance, if you notice a person with an eye movement, you can probably make out whether it was a wink or a twitch. But data cannot. When customer intelligence is applied to consumer behavior, companies are likely to find that knowing about someone is not the same as knowing them. Customer data can be used to strengthen customer relationships, but those actions have to build values and trust. Enamored with the predictive power of data, we often lose out on how to deal with people. Systems generate e-mails, SMS, and telephone calls to customers by millions, which are merely seen as intrusion in personal lives. Sending a birthday greeting to a customer might seem good CRM by a company, but for a customer getting such a message from a nameless and faceless company may mean nothing at all.
Secondly, many companies do not have access to big data analytics. Meer (2014) writes that small companies, companies in emerging markets, and business-to-business (B2B) industries operate in data-sparse situations. Such companies have to adopt a series of little data techniques to develop solutions for their needs.
A word of caution is given by Ross, Beath, and Quaadgras (2013). They warn about the hype surrounding big data which has made companies expect more than they can. Very often managers can gain insights merely by observing people. Moreover, using data analytics requires huge changes in business models and organizations that companies are either incapable or reluctant to do.
In the ultimate analysis, data analysis cannot substitute for common sense and observation. Companies also have to learn to use little data as well, which requires three things, writes Meer:
In the next chapter, we discover the marketing communications that affect modern consumers like Mary. Data techniques, farming social media information, help to create targeted communications which people do not block out.
3.129.247.196