Customer Segmentation: A Scientific Approach to Marketing

Learn tools to segment the market and earn credit toward a certificate in Data-Driven Marketing from Cornell University
Instructed by eCornell
  • Lectures 22
  • Video 2 Hours
  • Skill level intermediate level
  • Languages English
  • Includes Lifetime access
    30 day money back guarantee!
    Available on iOS and Android

How taking a course works


Find online courses made by experts from around the world.


Take your courses with you and learn anywhere, anytime.


Learn and practice real-world skills and achieve your goals.

Course Description

In this course from the Johnson School of Management, you will use cluster analysis to divide the market based on customer needs and preferences. This helps you identify and target the segments with the greatest potential for profitability. Through dynamic activities, you will analyze data similar to those typically provided by market research firms and answer segmentation and targeting questions; for instance, you will analyze the data provided by a firm’s website browsing history and predict which segments are most attractive for a firm to target.

What are the requirements?

  • None

What am I going to get from this course?

  • Over 22 lectures and 1 hour of content!
  • Over 5 hours of activities, practical exercises, and quizzes!
  • Get started on your pathway to a certificate in Data-Driven Marketing from Cornell University
  • Articulate a strategic rationale for customer segmentation
  • Articulate how segmentation fits into the process of developing marketing strategy
  • Use data on bases and descriptor variables to group customers into segments
  • Identify some of the most common algorithms used in segmentation

What is the target audience?

  • Anyone interested in learning about using data to segment and better target their customers
  • Anyone interested in pursuing a full Cornell University Certificate in Data-Driven Marketing via eCornell

What you get with this course?

Not for you? No problem.
30 day money back guarantee

Forever yours.
Lifetime access

Learn on the go.
Desktop, iOS and Android

Get rewarded.
Certificate of completion


Section 1: Exploring Segmentation Bases and Descriptors
As marketers, we want greater market share and increased profitability. To achieve these goals, we use data analysis to answer two critical questions: Who is your target customer? And what unique benefit does your product offer relative to competing products? In this course, developed by Professor Sachin Gupta of the Johnson School of Management at Cornell University, you examine the process of segmenting the market and targeting the right customers so your marketing strategies yield better returns. In the context of segmentation and targeting, you also want to develop an understanding of "big data," a popular concept that's currently attracting much attention. Analyzing big data requires specialized software, but the underlying theories of how to approach and analyze it are the same as those used here to analyze smaller data sets. You'll hear from Cornell University Tisch Professor Johannes Gehrke about why big data are important, and how an analysis of big data can help marketing professionals with segmentation and targeting.

Professor Gupta's research focuses on analytical models of marketing phenomena, including discrete choice models of consumer behavior, marketing mix models, measurement of returns on marketing investments, pricing, promotions, and advertising decisions, channel relationships, and so forth. He is especially interested in the prescription drug and consumer goods industries.

In 2008 one of Professor Gupta’s papers received the O’Dell award of the American Marketing Association. This award is given to the authors of the best article published in the Journal of Marketing Research five years before. Professor Gupta also received the Paul Green award of the American Marketing Association in 2003. In 2007, he received the Cornell Hospitality Quarterly's best paper award for his article on customer satisfaction in the restaurant industry. Professor Gupta serves on the editorial boards of Marketing Science and the Journal of Marketing Research.

At Johnson, Gupta teaches the core Marketing Management course, as well as a popular elective course called Data Driven Marketing. In 2009, he received the Stephen Russell Distinguished Teaching Award, given by the Johnson class of 2004, at their fifth reunion. The 2007 graduating MBA class selected him to receive the Apple Award for Teaching Excellence. Gupta previously taught at the Kellogg School of Management at Northwestern University, where he received the Sidney Levy Award for teaching excellence.


In this video, Professor Gupta explains that because customers vary, it is difficult to market products or services that try to be all things to all people. Instead, marketers try to see how certain groups of people are similar (homogeneous) and others are different (heterogeneous). These groups should be fairly homogeneous within each group, but heterogeneous from one group to the next. People will often pay more money for things that exactly meet their needs or that have a specific appeal (and when the additional amount that they pay is greater than the cost to find the exact match).

For existing brands, it is especially useful if the needs and benefits used for segmentation can be mapped to the product’s value proposition. For instance, if you have a drug whose primary differentiating advantage is that it works quickly, the segmentation will be useful if you use that characteristic as one of the needs or benefits that you use to segment.


In this video, Professor Gupta discusses bases variables and descriptor variables, which help define segments. You must understand these two terms to successfully complete this module. "Bases" are needs, motivations, and preferences; "descriptors" refer to demographics, psychographics, geographic locations, and other characteristics. Bases tell us why customers will respond differently to a given offering—for example, they may have different needs and wants. Bases are often hard to observe, except via market research done with a sample. With some effort, however, descriptors can be observed.

The most appropriate basis for segmentation depends on the managerial reason for the segmentation. For example, for positioning studies for existing products, appropriate bases may be benefits sought, product use, or attribute preferences; for a new-product concept, reaction to the new concept may be the appropriate basis. The general approach is to create a segmentation framework using a sample, and then apply it to the population at large by using descriptors.

5 slides
As we've seen, bases and descriptors provide valuable insight to marketers. Review the following scenario related to bases and descriptors and answer the questions that follow.
Effective segmentation studies rely on good underlying data gathered via surveys, conjoint analysis and so on. 
Course Project Part I: Interpreting Clusters
6 questions
Section 2: Ask the Expert: Johannes Gehrke on Big Data and Segmentation
Now that you've had a chance to explore segmentation studies and the information they yield, you are ready to consider how that analysis may be executed on a broader scale. What do marketers need to know to discuss big data knowledgeably in terms of segmentation? The work you're doing in this course will provide you with the foundation you need to understand how analysts measure and analyze big data. Professor Gupta invites Cornell University professor Johannes Gehrke to discuss some of the most commonly asked questions about big data and how they relate to the study of advanced marketing research. 

Question for the expert: What is a transaction? 

Johannes Gehrke is the Tisch University Professor in the Department of Computer Science at Cornell University. Gehrke's research interests are in the areas of database systems, data science, and data privacy. He has received a National Science Foundation Career Award, an Arthur P. Sloan Fellowship, an IBM Faculty Award, the Cornell College of Engineering James and Mary Tien Excellence in Teaching Award, the Cornell University Provost's Award for Distinguished Scholarship, a Humboldt Research Award from the Alexander von Humboldt Foundation, the 2011 IEEE Computer Society Technical Achievement Award, and the 2011 Blavatnik Award for Young Scientists (from the New York Academy of Sciences). He co-authored the undergraduate textbook Database Management Systems (McGraw-Hill, 2002, currently in its third edition), used at universities all over the world. He is also an adjunct professor at the University of Tromsø in Norway. Gehrke was co-chair of the 2003 ACM SIGKDD Cup, program co-chair of the 2004 ACM International Conference on Knowledge Discovery and Data Mining (KDD 2004), program chair of the 33rd International Conference on Very Large Data Bases (VLDB 2007), and program co-chair of the 28th IEEE International Conference on Data Engineering (ICDE 2012). From 2007 to 2008, he was chief scientist at FAST, a Microsoft subsidiary.
Question for the expert: What are key properties of a transaction?
Question for the expert: What's the difference between transaction processing and a data warehouse?
Question for the expert: What is a market-basket analysis?
Question for the expert: What is an example of a market-basket analysis?
Section 3: Analyzing Data to Divide the Market

When we talk about segmenting and targeting customers, we're looking at analytical methods of classifying consumers into groups based on similar needs and preferences. How do we perform this statistical analysis? And how do we look for patterns in the data that will be meaningful from a business perspective? In this module, developed by Cornell University professor Sachin Gupta, you examine how you can divide the market meaningfully based on customer needs and preferences, so that you can identify and target the segments with the greatest potential for profitability.

You'll also hear again from Professor Gehrke, who explains how big data can be an important part of this conversation. You'll learn what to consider in terms of selecting the right attributes, what "data preprocessing" is, and what "representative subsets" are. Increasingly, marketing professionals are expected to have a well-rounded awareness of the terms being used in discussing big data.

As Professor Gupta explains in this video, cluster analysis is a statistical technique commonly used for segmentation. It classifies a set of "observations" (customers or prospects) into mutually exclusive, unknown groups based on several variables or shared properties.

Measuring the Euclidean distance could be thought of as measuring the distance between two points on a graph as if you placed a ruler between them; it is computed by the square root of the sum of shared differences.
To further your understanding, download the additional reading, Clustering Algorithms, that Professor Gupta has provided.
Hierarchical clustering seeks to build a hierarchy of clusters. "Agglomerative" hierarchical clustering can be thought of as a "bottom-up" approach; pairs of clusters are merged as we move further up the hierarchy.
In this video, Professor Gupta introduces some of the considerations in choosing a clustering method, and he describes how a hierarchy of clusters can be reached. You want to develop a well-rounded understanding of how analysts use consumer data to segment the market. You may not be the person who actually performs the computations described here, but you may be in a position to commission this analysis from a vendor. You should understand what the data can do for you and how these analyses inform segmentation and targeting decisions.
What are the methods that are available to relate the descriptor variables to cluster membership?
Section 4: Ask the Expert: Johannes Gehrke on Big Data and Data Preprocessing
Question for the expert: What's data preprocessing?
Question for the expert: Can you tell us more about data preprocessing?
Question for the expert: What do you do when you have missing values?
Question for the expert: How do you select the right attributes?
Question for the expert: What are representative subsets?
Section 5: Course Project and Wrap-up
Course Project Part II: Profiling Clusters
6 questions
Course Completion

Instructor Biography

eCornell is a subsidiary of Cornell University that provides online professional and executive development to students around the world. We offer more than 30 award-winning professional certificate programs in a wide variety of disciplines. eCornell‘s unique approach to online learning combines the most effective elements of an Ivy League classroom with the flexibility of an online learning environment. eCornell courses are all developed by Cornell University faculty, and often include practical insights from other industry experts. All eCornell course content comes from top-rated programs with proven curricula.

Join the biggest student community


Hours of video content


Course Enrollments




Average Rating
  1. 5 Stars
  2. 4 Stars
  3. 3 Stars
  4. 2 Stars
  5. 1 Stars
    • Meghan Vincent

    The New Content Was Good...

    But if you took the other class offered by eCornell (Data Driven Marketing), then you would be sitting through two sections of repeated material. The new content, though, was very thorough and helpful.

    • Audrey Heinesen

    Nicely taught

    The animations and graphics make it easy to understand the concepts and topics covered.

    • Frank Visciano

    Great course!

    A good introduction to a few powerful tools to help truly identify and target your customer.

Ready to start learning?
Preview this course