Hello. I'm Jang Saetbyeol. Here we are exploring a number of cases to understand how big data is changing our jobs and how we can utilize data in the future. Over the previous chapters, we have followed the flow of business to create products or services, ranging from planning to production and the distribution and logistics we discussed in chapter 3. Today, we will take a look at cases of utilizing big data to respond to and understand customers, who consume the products or services we have provided through a number of processes and efforts made by various staff members.

 In fact, "customer" has been mentioned in all of the areas we discussed in the previous three chapters. This is the proof that all business elements in the business flow agonize over internal and external customers.

 As big data has been placed in the spotlight, one of the areas to which people have paid attention is customer management. As we had already experienced the fever of CRM (customer relationship management) that once swept the entire industries before big data, there were also, most likely, many who considered big data to be useless. Some years after the advent of big data, however, it is still regarded as being an important foundation for the Fourth Industrial Revolution, featuring hyperconnectivity and superintelligence.

 What makes big data important is that it gives advantages in identifying and understanding customers. What does it mean by saying that you know someone? It could mean that you know them simply by their face or name, or that they were introduced to you, or you are very close enough to open your hearts to each other. However, knowing customers in terms of business includes deeper considerations than those in normal relationships.

 There are so many factors in knowing customers but what we learn about customers from big data could be summarized as follows.

 Even before big data, the simplest customer information and history, including purchase history, had been collected and managed systematically. However, we are now in the new era where we are technically ready to see how customers behave and observe what they want rather than imagining it. What we should do is consider and choose how we do it rather than whether to do it or not.

 Among the above questions, the one about communication methods requires more contemplation. It could be simpler to allow the customer to choose their desired method (SMS, e-mail, DM, etc.) but even if the customer chooses or refuses any of them, it is still impossible to judge whether they prefer or dislike the specific method. In addition, we cannot conclude that a customer who has not experienced online shopping, using the given information we know, also purchases other products such as daily necessities only at offline shops.

 Despite such difficulties, however, we should not choose what is not done, and thus we need to make various attempts to figure out the communication method preferred by customer. Our company has also made efforts to better understand our customers through various tests and learning attempts. I hope our case becomes the best practice in the area of understanding customers, but today, we will explore general cases where we utilize data to understand customers and suggest proper benefits or products, rather than specific areas.

 Have you ever tried to decide whether to order some food after watching a fried chicken commercial on TV or a night snack ad on a food delivery app on your smartphone? Have you ever watched any fried chicken commercial at a movie theater, which appears just with sound or images for less than 5 seconds, and ended up visiting some place to grab fried chicken immediately after the movie?

 The advertisement industry, which has an interest in curation utilizing data, has begun to utilize big data to provide personalized advertisements customized for specific customers rather than the existing broadcast method. Even before big data, the advertising firms tried to optimize their product based on their own standards and basis. This though, was practiced at a basic level. For example, they would place confectionery or toy commercials primarily to cartoons or on children's channels, and place insurance, banking and nighttime snack commercials before and after news or current affairs programs.

 Now, they identify customer interests based on personal behavioral data and information, and provide personalized advertisements. You might have seen the hotels or tourist spots you recently researched for your vacation or the products you clicked on at online shopping malls repeatedly appear in your Facebook news feed.

 This method is an example of very simple personalization as it is an example of re-targeting advertising about what the customer expressed interest in before (by keyword search, etc.). However, there is also a higher level of personalized advertising that exposes even products in which the customer has not shown any interest, based on prediction. In this method, a person who searched a certain makeup brand is considered to be likely to be a woman in their 20s or 30s, and thus ads on other products frequently searched by customers of that age and gender are also exposed to the person. If you look closely at the ads that frequently appear on websites, you might see how data thinks of you.

 There is a practice that has gradually spread and now is familiar to us. Have you felt that the frequency of seeing the word "discount" in car insurance commercials has been rising for some time now? The existing car insurance ads focus on improving insurer reputation or mentioning basic characteristics of insurance, such as avoiding loss, or talk about what can appeal to every customer, such as the low cost of direct insurance services. Recently, however, insurer advertisements have put more emphasis on premium discounts based on personal data.

 Progressive, a U.S. insurance company, analyzes customer's driving patterns based on the various data sent from the sensors attached to the customer's vehicle. Based on the data, they estimate the risk of accidents and reflect such probability values to the calculation of premium. This system titled "Pay as You Drive" considers not only driving habits, but also personal environmental factors, such as time or region of driving.

 Recently, Korean insurance companies also provide personalized product design by differentiating premium based on driving distance or reflecting driving habits, such as eco-driving, compliance or children, or even the presence of a dashboard camera, in calculating premium.

 VISA, a global card company, provides its RTM (Real Time Messaging) service that collects data on time, location and items of deals made by its card users in real time and analyzes this based on the customer's history and preference, and sends discount coupons for partner stores near the customer. Since starting this service, VISA has seen both the frequency of card use and the number of customers increase.

 AMEX also offers personalized marketing based on location data and customer data. It connects the customer's SNS account with their card account to provide something that is likely to be preferred by the customer. If you click "Like" on Facebook for a photo of a restaurant in Itaewon, the card company sends you the discount benefits or information for the restaurant or other partner restaurants located in Itaewon.

 Recently, Samsung Card has also been promoting its LINK service, which also utilizes region/customer data. This service recommends a group of affiliates and benefits (discount, 2 for 1 promotion, etc.) that are likely to be preferred by the customer, who in turn store the benefits in their card with a simple click. If the customer uses the card at the affiliate, they will automatically get the benefits without having to produce a coupon. Since Samsung Card launched this service, it has seen a dramatic rise in card usage as well as the number of new customers.

 Now being used by almost every card company, a system that detects illegal use of cards and prevents fraudulent use also utilizes big data. This is called a fraud detection system. If one attempts to use a card in a region different from those where the card is usually used or attempts to pay many times at a single affiliate or pay an unusually large amount, the transaction may not be approved until the user's identity is verified.

 Some of you might have experienced being denied when using your credit card on your summer vacation or overseas trip for the reason of suspected illegal use. This is in fact the proof that the system is in place, so I hope you consider it to be a small inconvenience for reliable financial transactions.

 In this chapter, we explored cases of utilizing big data to better understand customers and offer personalized benefits. You might now realize that you have already received personalized recommendations many times from advertisements or card companies even though you were not aware of it. With the evolution of machine learning or A.I., some day we might live in a world where we plan our day with recommendations from somebody who knows me better than me.

 In the next chapter, we will discuss how customer behaviors are observed, understood and predicted. Understanding how customers behave and what they will purchase or do is something that attracts interest and investments from both online and offline companies. Nobody can reverse the wave of this omni-channel era. I will come back in the next chapter with more interesting cases of big data, including how businesses observe, record and utilize customer behaviors. 