Analysis Report Three - Privacy and Customer Profiling

Executive Summary

Healthcare organizations and other businesses use customer data to better understand buying habits, improve services, and provide more personalized information. However, collecting and using customer data can create privacy concerns when people do not fully understand what information is being collected or how it will be used. Customers are more likely to share their information when a company is honest about its data practices, gives them some control, and provides something useful in return.

This report uses Northwind transaction data to show how individual purchases can be grouped and summarized to create a customer profile. By using SQL and data visualizations, managers can identify a customer’s most frequently purchased products and the products purchased in the largest quantities. This information can help companies provide better recommendations and more relevant advertising, but it can also become a privacy concern when customers do not know that their purchasing behavior is being tracked and analyzed.

Organizations should only collect the information they actually need and clearly explain how that information will be used. They should also make sure customer profiles are accurate, allow customers to have some control over their data, and use the information in ways that benefit the customer as well as the company. These steps can help organizations use customer data effectively while still protecting privacy and maintaining trust.

Introduction

The main reading explains how important customer data has become to businesses today. Companies use this information to understand their customers, improve products and services, and create more personalized advertising. However, many customers may not fully understand how much information they are sharing or what companies are doing with it. When companies are not open about their data practices, it can create privacy concerns and cause customers to lose trust.

The reading describes three main types of customer data. Self-reported data is information customers provide themselves, such as their name, email address, age, or employment history. Digital exhaust is information collected while people use websites, apps, and connected devices, such as their location or browsing history. Profiling data is created when companies combine and study this information to make predictions about a customer’s interests and behavior. Profiling data can feel more personal because it may reveal things that the customer never directly shared.

Another important point from the reading is that customers expect to receive something useful in return for sharing their information. Customers may be more comfortable with data collection when it improves a product or makes a service more convenient. However, they may expect a greater benefit when their information is used for targeted advertising or sold to another company. Because of this, businesses should clearly explain what information they collect, how they plan to use it, and what the customer will receive in return.

The reading also shows that trust can give a company an advantage over its competitors. Customers are more likely to share information with organizations they believe will protect it and use it responsibly. Companies can build trust by explaining their data practices in a way customers can understand, giving customers control over their information, and providing real value in return. These ideas are especially important in healthcare because patient information is highly sensitive. Healthcare organizations must find a balance between using data to improve care and protecting patient privacy.

Privacy vs Profiling

Customer profiling can help organizations better understand what customers buy, what they are interested in, and what they may purchase in the future. Companies can use this information to improve services, recommend products, and create advertising that is more relevant to each customer. However, profiling can also create privacy concerns, especially when customers do not fully understand what information is being collected or how much a company can learn about them.

The main reading explains that companies collect self-reported data, digital exhaust, and profiling data. Customers may knowingly provide basic information such as their name, email address, or age. However, they may not realize that companies are also collecting information such as their location, browsing history, and online activity. When companies combine all of this information, they can create a much more detailed customer profile. The reading shows that customers may be more willing to share their information when the company clearly explains how it will be used and gives them something useful in return. Disney’s MagicBands are a good example because customers share information but receive convenience, faster service, and a more personalized experience.

The article about retailers becoming advertising platforms shows how valuable customer information has become. Retailers such as Amazon and Walmart can use purchase histories, online searches, and loyalty-program information to help companies target advertisements. This can make advertising more useful, but it also creates concerns about transparency. Retailers may operate the advertising platform and also report whether the advertisements were successful. Manufacturers may not know exactly how those results were measured, and customers may not realize that their shopping habits are being used to create advertising profiles.

AI chatbots create another concern related to customer profiling. Technology companies are trying to make chatbots feel more personal and friendly so people will continue using them. To provide this kind of personalization, a chatbot may collect information from conversations about a person’s interests, beliefs, habits, and emotions. This information may help the chatbot give better responses, but it could also be used to keep the person engaged or influence how they think. Companies should clearly explain whether conversations are saved, how the information is analyzed, and whether it is shared with other organizations.

Another article shows that companies cannot prove they are trustworthy simply by using words such as ethical, integrity, or responsibility. Customers and investors are more likely to judge a company by its actions. An organization must earn trust by being open about its practices, protecting customer information, and taking responsibility when problems occur.

Outside research supports many of these same ideas. One study found that customers may feel more vulnerable when companies have access to their personal information. The study also found that transparency and customer control can help reduce some of those concerns (martin2017data?). Another study found that highly personalized advertising worked better when customers already trusted the retailer. When customers did not trust the retailer, the advertising felt more invasive and created stronger privacy concerns (bleier2015importance?).

Research on the personalization paradox also found that customers react differently depending on how their information is collected. Customers were more accepting of personalized advertising when the company was open about the data collection. When information was collected without the customer’s knowledge, customers felt more vulnerable and became less interested in the advertisements (aguirre2015unraveling?). This shows that collecting more information does not always lead to better results. A company may create a detailed customer profile, but that profile may lose its value if the customer feels watched, misled, or manipulated.

These concerns are especially important in healthcare because patient information may include medical conditions, prescriptions, appointments, payment information, and personal conversations. Healthcare organizations can use this data to improve care, send reminders, and provide more personalized services. However, patients should understand what information is being collected and how it will be used. Healthcare organizations should only collect the information they need, protect it carefully, give patients reasonable control, and make sure the use of the data provides a real benefit. Customer profiling can support better decisions, but it must be handled in a way that protects privacy and maintains trust.

Profiling Example

For this customer profiling example, I chose QUICK-Stop from the Northwind database. QUICK-Stop has enough purchasing activity to show how regular transaction data can be grouped and summarized to better understand a customer’s buying behavior. By looking at the products they purchase, how much they purchase, and how often certain products appear in their orders, a business can begin to develop a customer profile.

This type of information could help a company prepare for future orders or recommend products that may interest the customer. At the same time, it shows how much information can be learned from normal transaction records, which is why companies should be open about how customer purchasing information is collected and used.

Visualization One: Total Quantity Purchased by Product

The first visualization looks at the total quantity of each product purchased by QUICK-Stop. The query connects the Orders, Order Details, and Products tables. The WHERE statement limits the records to QUICK-Stop, while SUM() adds together the quantities purchased for each product. The GROUP BY statement creates a separate group for each product, and the HAVING statement keeps only products with a total quantity greater than 2,550 so the graph focuses on the products purchased in the largest amounts.

SELECT Products.ProductName,
       SUM(CAST("Order Details".Quantity AS INTEGER)) AS total_quantity
FROM Orders
INNER JOIN "Order Details"
ON Orders.OrderID = "Order Details".OrderID
INNER JOIN Products
ON "Order Details".ProductID = Products.ProductID
WHERE Orders.CustomerID = "QUICK"
GROUP BY Products.ProductName
HAVING total_quantity > 2550

ggplot(data = myquery1,
       aes(y = ProductName, x = total_quantity)) +
  geom_col() +
  labs(
    title = "Total Quantity Purchased by Product",
    x = "Total Quantity Purchased",
    y = "Product"
  )

This visualization shows which products QUICK-Stop purchased in the greatest total quantities. Products with longer bars represent products that QUICK-Stop purchased in larger amounts overall. Since SQL already grouped the records and calculated the totals, I used geom_col() instead of geom_bar().

This type of information could help a business identify products the customer may need again in the future. It could also be used for recommendations or targeted promotions. However, it is also an example of how a company can use normal transaction data to learn more about a customer’s purchasing preferences.

Visualization Two: Frequently Purchased Products

The second visualization looks at QUICK-Stop’s purchasing behavior in a different way. Instead of adding together the quantities purchased, this query uses COUNT(*) to count how many purchase records there are for each product. I used the alias n for the count, similar to the example used in Practice Three.

The query again uses GROUP BY to create a separate group for each product. The HAVING statement keeps only products that appear in more than 95 purchase records. This helps focus the graph on the products QUICK-Stop purchased most frequently.

SELECT Products.ProductName,
       COUNT(*) AS n
FROM Orders
INNER JOIN "Order Details"
ON Orders.OrderID = "Order Details".OrderID
INNER JOIN Products
ON "Order Details".ProductID = Products.ProductID
WHERE Orders.CustomerID = "QUICK"
GROUP BY Products.ProductName
HAVING n > 95

ggplot(data = myquery2,
       aes(y = ProductName, x = n)) +
  geom_col() +
  labs(
    title = "Frequently Purchased Products",
    x = "Number of Purchase Records",
    y = "Product"
  )

This visualization shows how often different products appear in QUICK-Stop’s purchasing records. Products with longer bars appear more frequently in the customer’s transaction history. This gives a different view from the first graph because a product could be purchased frequently in smaller quantities or purchased fewer times in larger quantities.

Together, the two visualizations give a better picture of QUICK-Stop’s purchasing behavior. The first graph shows how much of each product the customer purchased, while the second graph shows how often certain products were purchased. A business could use both types of information to better understand the customer and possibly predict future purchasing needs.

At the same time, this example connects back to the privacy concerns discussed in the readings. QUICK-Stop is simply making normal purchases, but those individual transactions can be grouped and analyzed to reveal patterns about the customer. Organizations should be open about how this information is being used and make sure that customer profiling provides value to the customer as well as the business.

Recommendations for Industry

Organizations that use customer profiling should make sure they are collecting and using customer information in a way that is useful, but also respectful of privacy. The research shows that customers are more comfortable sharing information when they understand what is being collected, how it will be used, and what benefit they will receive in return.

One recommendation is for organizations to only collect the information they actually need. Collecting large amounts of customer data simply because it may be useful later can create unnecessary privacy and security concerns. Companies should have a clear reason for collecting each type of information and should avoid gathering information that does not provide a real benefit to the customer or the organization.

Organizations should also be more transparent about how customer information is being used. Customers may willingly provide basic information without realizing that their purchases, searches, locations, or other activities can later be combined to create a much more detailed profile. Companies should explain these practices in a way that customers can easily understand and give them reasonable control over how their information is used.

Another recommendation is to make sure customer profiling provides value to the customer and not only to the company. The Northwind example shows how normal transaction records can be grouped and analyzed to identify the products QUICK-Stop purchases most often and in the largest quantities. A company could use this information to prepare for future orders, recommend related products, or provide more useful promotions. Using the information in this way gives the customer something in return for allowing the company to learn from their purchasing behavior.

Companies should also regularly review customer profiles to make sure the information is accurate and is not being used in ways that customers would consider unfair or misleading. Retailers are using customer purchasing information more often for targeted advertising, which makes transparency and accountability even more important. Organizations should be able to explain how customer data is being used and how decisions or recommendations are being made from that information.

Finally, organizations need to remember that trust has to be earned through their actions. Simply telling customers that a company values privacy or is trustworthy is not enough. Companies should protect customer information, be honest about their data practices, give customers reasonable choices, and use profiling in ways that provide a clear benefit. Taking these steps can allow organizations to gain useful information from customer data while still respecting privacy and maintaining customer trust.

References

(article?){martin2017data, title={Data Privacy: Effects on Customer and Firm Performance}, author={Martin, Kelly D. and Borah, Abhishek and Palmatier, Robert W.}, journal={Journal of Marketing}, volume={81}, number={1}, pages={36–58}, year={2017}, doi={10.1509/jm.15.0497} } (article?){bleier2015importance, title={The Importance of Trust for Personalized Online Advertising}, author={Bleier, Alexander and Eisenbeiss, Maik}, journal={Journal of Retailing}, volume={91}, number={3}, pages={390–409}, year={2015}, doi={10.1016/j.jretai.2015.04.001} } (article?){aguirre2015unraveling, title={Unraveling the Personalization Paradox: The Effect of Information Collection and Trust-Building Strategies on Online Advertisement Effectiveness}, author={Aguirre, Elizabeth and Mahr, Dominik and Grewal, Dhruv and de Ruyter, Ko and Wetzels, Martin}, journal={Journal of Retailing}, volume={91}, number={1}, pages={34–49}, year={2015}, doi={10.1016/j.jretai.2014.09.005} }