Analysis Report Three - Privacy and Customer Profiling
Author
Jessica Garcia
Executive Summary
Organizations increasingly rely on customer data to personalize products, improve marketing, and create new sources of revenue. The readings show that customer profiling can improve customer experiences and organizational performance, but it also raises concerns surrounding privacy, transparency, consent, and trust (morey2015?; gabel2024?). Companies that collect extensive customer information without clearly communicating how it will be used risk damaging customer relationships and losing long term access to valuable data.
Organizations are finding new ways to use and monetize customer information. Retailers are becoming advertising platforms by using first party purchase data to deliver targeted promotions and advertisements. At the same time, AI companies are developing personalized digital assistants that learn users preferences, values, and behaviors. Both developments depend on the creation of increasingly detailed customer profile.
The Northwind customer analysis demonstrates how routine transaction data can reveal a customers category preferences, spending patterns, and purchasing activity over time. Organizations should therefore treat customer profiling as a responsibility rather than simply a technical opportunity. Businesses should clearly explain their data practices, provide customers with meaningful control, limit profiling to legitimate purposes, and deliver recognizable value in exchange for personal information.
Introduction
Customer data has become an important strategic resource because it allows organizations to understand customer preferences, predict behavior, personalize services, and improve business decisions. Companies collect this information through websites, applications, connected products, loyalty programs, purchase histories, and other customer interactions. When these individual pieces of information are combined, organizations can create detailed profiles that predict what customers may want or do in the future.
Morey, Forbath, and Schoop divide customer information into three categories: self-reported data, digital exhaust, and profiling data (morey2015?). Self reported data includes information that customers intentionally provide, such as an email address or demographic information. Digital exhaust is created through normal technology use and may include browsing history, location, or communication activity. Profiling data is produced when organizations combine and analyze information to make predictions about a person’s interests and behaviors.
The reading argues that profiling data is especially valuable and sensitive because it includes conclusions that organizations create about customers rather than information customers directly provide. Although consumers are generally aware that companies collect information, they often do not understand the exact types of information being captured. The article reports that consumers are concerned about misuse even when they are poorly informed about their complete data footprints.
The authors argue that companies should not approach customer data though secrecy or minimal regulatory compliance (morey2015?). Organizations should instead build transparency, control, and value into their data strategies. Customers are more willing to provide information when they understand how it is used, receive a clear benefit, and trust the organization to protect it. Trust therefore becomes a competitive advantage because trusted businesses may receive greater and more sustained access to customer data.
Privacy vs Profiling
Customer profiling creates a conflict between organizational value and individual privacy. Businesses benefit from understanding customer behavior because profiles can improve recommendations, promotions, product design, and advertising. Customers may also benefit when personalization saves time, reduces costs, or improves their experience. However, profiling becomes more concerning when customers do not understand that it is occurring or when organizations use information in ways that primarily benefit the company.
The transparency and trust reading explains that customers place different values on different types of data. Customers generally place less value on basic information that they intentionally provide and greater value on digital exhaust and profiles generated through analytics (morey2015?). Customer expectations also depend on how organizations use the data. People may accept data collection when it directly improves a product or service, but they expect more value when their information is used for targeted advertising and the greatest value when information is sold to third parties.
Disney’s MagicBand system demonstrates a relatively transparent exchange. Disney collects extensive information about visitors’ movements and purchases, but the system also provides customers with benefits such as park access, hotel entry, reservations, and simplified purchases. Disney explains the system’s data practices during registration, allowing customers to better understand the exchange. Pandora similarly collects stated preferences and listening behavior to personalize music selections. In both cases, profiling supports an identifiable service benefit (morey2015?).
Other examples show the consequences of profiling without sufficient transparency. Target caused public concern after its analysis of shopping behavior allowed it to identify customers who were likely to be pregnant, sometimes before those customers had shared that information with others. The concern was not simply that Target possessed purchase information. The company used ordinary purchases to infer sensitive personal information that customers had not knowingly provided (morey2015?).
The growth of retail media further expands the use of profiling. Retailers such as Amazon, Walmart, Target, and other companies use first-party transaction data to deliver sponsored search results, digital coupons, and advertisements (gabel2024?). Retailers may understand purchasing behavior more accurately than outside advertising platforms because they can observe actual transactions. This creates a major business opportunity, but it also makes customer purchase histories increasingly valuable for purposes beyond completing a sale.
Retail media also introduces transparency challenges because retailers may both execute advertising campaigns and report their performance (gabel2024?). Advertisers may not be able to independently observe the impressions purchased, the retailer’s margins, or the attribution methods used to connect advertisements with sales. The reading argues that retail media platforms need transparent measurement and experimentation capabilities so organizations can credibly demonstrate the effectiveness of their advertising activity.
Recent research from the Federal Trade Commission demonstrates that the conflict between personalization and privacy continues to affect major digital platforms. In their 2024 examination of social media and video streaming companies, the FTC found that many platforms collected and retained extensive amounts of user information in order to support targeted advertising and other revenue generating activities. The report concluded that these businesses incentives were frequently in tension with user privacy and recommended stronger data minimization, retention, transparency, and consumer control practices. this supports the arguments from the other readings that organizations should not collect information simply because it may become valuable later. Customer profiling should instead be limited to clearly defines purposes that customers understand and reasonably expect (book?).
The article about personalized AI companions presents another form of profiling. Companies want conversational AI systems to learn users’ personalities, preferences, values, and communication styles so interactions feel more personal. These systems may provide companionship and emotional support, but personalization also requires continuous collection and interpretation of user behavior. An AI designed to reflect a specific worldview may reinforce a user’s existing beliefs rather than expose the person to balanced information. This creates the risk of highly personalized echo chambers.
The article concerning corporate “trust words” provides an additional warning. Companies that repeatedly use words such as ethical, integrity, and responsibility do not automatically become trustworthy. The research described in the article found that the use of these terms was associated with several negative outcomes, although the study did not establish causation. The implication is that organizations cannot replace transparent behavior with reassuring language. Trust must be supported through actual policies, controls, and accountable decisions.
Together, the readings suggest that customer profiling is not automatically harmful. The ethical and strategic concern depends on what data is collected, what is inferred, how the profile is used, whether the customer understands the process, and whether the exchange creates fair value. Profiling can strengthen customer relationships when it produces useful services and respects customer choices. It can damage those relationships when it is hidden, intrusive, manipulative, or primarily designed to extract value from customers.
Profiling Example
To demonstrate how transaction data can produce a customer profile, this section examines Hanari Carnes, customer ID ‘HANAR’. Northwind does not contain the same sensetive personal information collected by modern digital platforms. Its transaction records can still reveal purchasing preferences, spending levels, order frequency, category interests, and changing behavior over time.
Customer Background
Query identifies customer and provides basic account information.
CustomerID CompanyName City Country
1 HANAR Hanari Carnes Rio de Janeiro Brazil
This query confirms that ‘HANAR’ represents Hanari Carnes, located in Rio de Janeiro, Brazil. This information provides the basic context for analyzing the customers transactions.
Visual One: Spending by Product Category
The first analysis identifies the product categories that generate the most revenue from Hanari Carnes. Query joins ‘Orders’, ‘Order Details’, ‘Products’, and ‘Categories’ tables. It then uses ‘GROUP BY’ to organize the transactions by category.
category_profile <-dbGetQuery( mydb,' SELECT c.CategoryName, COUNT(*) AS purchase_lines, SUM(od.Quantity) AS total_units, ROUND( SUM(od.UnitPrice * od.Quantity * (1 - od.Discount)), 2 ) AS total_revenue, ROUND( AVG(od.Discount) * 100, 2 ) AS avg_discount_percentFROM Orders AS oINNER JOIN "Order Details" AS od ON o.OrderID = od.OrderIDINNER JOIN Products AS p ON od.ProductID = p.ProductIDINNER JOIN Categories AS c ON p.CategoryID = c.CategoryIDWHERE o.CustomerID = "HANAR"GROUP BY c.CategoryNameHAVING SUM(od.UnitPrice * od.Quantity * (1 - od.Discount)) > 0ORDER BY total_revenue DESC;')category_profile
ggplot(data = category_profile,mapping =aes(x =reorder(CategoryName, total_revenue),y = total_revenue )) +geom_col(fill ="pink") +coord_flip() +scale_y_continuous(labels = scales::label_dollar() ) +labs(title ="Hanari Carnes Spending by Product Category",subtitle ="Revenue includes quantity, unit price, and order-line discounts",x ="Product Category",y ="Total Revenue" ) +theme_minimal()
Hanari Carnes total discounted revenue by product category.
Visualization Two: Purchasing Patterns Over Time
Evaluates when Hanari Carnes purchases products and which categories contribute to spending during each period. The query groups purchases by year, quarter, and product category.
quarterly_category_profile <-dbGetQuery( mydb,' SELECT strftime("%Y", o.OrderDate) AS order_year, CASE WHEN CAST(strftime("%m", o.OrderDate) AS INTEGER) BETWEEN 1 AND 3 THEN "Q1" WHEN CAST(strftime("%m", o.OrderDate) AS INTEGER) BETWEEN 4 AND 6 THEN "Q2" WHEN CAST(strftime("%m", o.OrderDate) AS INTEGER) BETWEEN 7 AND 9 THEN "Q3" ELSE "Q4" END AS order_quarter, c.CategoryName, SUM(od.Quantity) AS total_units, ROUND( SUM(od.UnitPrice * od.Quantity * (1 - od.Discount)), 2 ) AS total_revenue FROM Orders AS o INNER JOIN "Order Details" AS od ON o.OrderID = od.OrderID INNER JOIN Products AS p ON od.ProductID = p.ProductID INNER JOIN Categories AS c ON p.CategoryID = c.CategoryID WHERE o.CustomerID = "HANAR" GROUP BY order_year, order_quarter, c.CategoryName HAVING SUM(od.UnitPrice * od.Quantity * (1 - od.Discount)) > 0 ORDER BY order_year, order_quarter, total_revenue DESC ')quarterly_category_profile
Quarterly Hanari Carnes revenue divided by product category.
Recommendations for Industry
Organizations should first practice transparency by clearly explaining what information they collect, how they create customer profiles, and how those profiles influence customer experiences (morey2015?). Companies should not depend only on lengthy privacy policies or terms and conditions. Data explanations should be understanable and presented when information is collected or used.
Organizations should give customers meaningful control over their information. Customers should be able to review privacy settings, make choices regarding targeted ads, correct inaccurate information, and withdraw access when appropriate. Providing control reduces the imbalance between the organization creating the profile and the person represented by it.
Customer profiling should provide recognizable value. Customers are more likely to accept data collection when it produces better recommendations, greater convenience, improved service, or financial savings. Organizations should be able to explain the relationship between the information collected and the customer benefit delivered.
Organizations should follow data minimization principles (book?). A company should not collect information simply because it may become useful later. It should identify a purpose and limit collection to information needed for that purpose. Derived information should recieve special attention because profiles may reveal conclusions that customers never directly disclosed.
Companies operating media platforms should create transparent and independent measurement processes (gabel2024?). Retailers should clearly explain how advertisements are targeted, how sales are attributed to campaigns, and how performance is measured. Separating campaign execution from evaluation can increase accountability and advertiser confidence.
Organizations that are developing AI systems should prevent personalization from becoming manipulation. AI systems should not use detailed profiles to exploit emotional dependence or reinforce only the opinions users already hold. Companies should establish boundaries for personalization and regularly evaluate whether automated systems produce biases, unsafe, or misleading interactions.
Finally, organizations should demonstrate trustworthiness through behavior rather than promotional language. Teams like: ethical, responsible, and trustworthy are only meaningful when supported by clear policies, secure systems, accountable leadership, and consistent treatment of customer information. In a data driven economy, customer trust is not simply a public relations objective. Its necessary for maintaining long term success to the information that makes effective personalization possible.
References
(book?){ftc staff report finds large social media and video streaming companies have engaged in vast surveillance of users with lax privacy controls and inadequate safeguards for kids and teens_2024, url={https://www.ftc.gov/news-events/news/press-releases/2024/09/ftc-staff-report-finds-large-social-media-video-streaming-companies-have-engaged-vast-surveillance?utm_source=chatgpt.com}, journal={Federal Trade Commission}, year={2024}, month=sept }
(gabel2024?){gabel_duncan_timoshenko_2024, title={How Retailers Became Ad Platforms}, author={Gabel, Sebastian and Duncan, Duncan and Timoshenko, Artem}, year={2024}, month=june }
(morey2015?){morey_forbath_schoop_2015, title={CUSTOMER DATA: DESIGNING FOR TRANSPARENCY AND TRUST}, author={Morey, Timothy and Forbath, Theo and Schoop, Allison}, year={2015}, month=may }