University of Southern California Dissertations and Theses

Visualizing email response data to improve marketing campaigns

(USC Thesis Other)

Visualizing email response data to improve marketing campaigns

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content VISUALIZING EMAIL RESPONSE DATA TO IMPROVE
MARKETING CAMPAIGNS

by

Jason Gonzalez

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOGRAPHIC INFORMATION SCIENCE AND TECHNOLOGY)

December 2015

Copyright 2015 Jason Gonzalez

ii

DEDICATION
I dedicate this document to my wife and boys.

iii

ACKNOWLEDGMENTS
I will be forever grateful to my mentor, Professor Swift and the committee. Thanks also to my
family, without whom I could not have made it this far.

i

TABLE OF CONTENTS
DEDICATION ii

ACKNOWLEDGMENTS iii

LIST OF TABLES v

LIST OF FIGURES vii

LIST OF ABBREVIATIONS x

ABSTRACT xi

CHAPTER 1: INTRODUCTION 1

1.1 Research Question and Study Purpose 1

1.2 Study Area 7

CHAPTER 2: BACKGROUND 10

2.1 Effect of Permission-Based Email Marketing 12

2.2 Market Segmentation 13

2.3 Alumni Response 14

2.4 GIS and University Advancement 15

2.5
Normalized versus Non-Normalized Data in Marketing 16

2.6 Ecological Fallacy 17

2.7 Hot Spot Analysis in Business 18

CHAPTER 3: METHODOLOGY 20

3.1 Alumni Data Management 20

3.2 Spatial Analyses 21

3.2.1 Email Data Preparation 22

ii

3.2.2 Normalized Alumni Response Data 24

3.2.3 Alumni Staff Survey 25

3.2.4 Dot Density Analysis 26

3.2.5 Hot Spot Analysis 27

CHAPTER 4: RESULTS 31

4.1 Identifying Patterns in Alumni Locations and Email Responses 31

4.2 Dot Density Maps 33

4.3 Email Respondents – Football Related Email 35

4.3.1 Female 25-34 35

4.3.2 Male 25-34 39

4.3.3 Female 35-44 41

4.3.4 Male 35-44 43

4.3.5 Female 45-54 45

4.3.6 Male 45-54 47

4.4 Email Respondents - Business Partnership Email 49

4.4.1 Female 25-34 49

4.4.2 Male 25-34 51

4.4.3 Female 35-44 53

4.4.4 Male 35-44 55

4.4.5 Female 45-54 57

4.4.6 Male 45-54 59

4.5 Email Respondents – USC Around Town Events Email 61

iii

4.5.1 Female 25-34 61

4.5.2 Male 25-34 63

4.5.3 Female 35-44 65

4.5.4 Male 35-44 67

4.5.5 Female 45-54 69

4.5.6 Male 45-54 71

4.6 Market Segmentation 73

4.6.1 Market Segmentation of the 25-34 Age Range 74

4.6.2 Market Segmentation of the 35-44 Age Range 77

4.6.3 Market Segmentation of the 45-54 Age Range 80

4.7 Normalized Versus Non-Normalized Results 81

4.8 Other Findings 82

CHAPTER 5: DISCUSSION AND CONCLUSIONS 86

5.1 Study Impact 86

5.2 Recommendations for Improving USCAA Email Marketing Lists 88

5.3 Limitations of Data 90

5.3.1 Effects of Normalizing Data 90

5.3.2 Accuracy of Alumni Addresses 93

5.3 Future Work 93

REFERENCES 95

APPENDIX A: Email Content Examples 98

APPENDIX B: Alumni Staff Survey 101

iv

APPENDIX C: Dot Density of Entire Email Populations in Each Category 103

APPENDIX D: Esri Tapestry Segmentation of Alumni Who Opened an Email 121

v

LIST OF TABLES
Table 1 Thesis Population 8

Table 2 Thesis Population Mean, Max, and Min 8

Table 3 Email Open Rate Categories (USCAA 2014) 23

Table 4 USCAA Generational Groups (USCAA, 2015) 24

Table 5 Positive Z-Scores Selected from Hot Spot Polygons (Football Email) 28

Table 6 Positive Z-Scores Selected from Hot Spot Polygons (Corporate Sponsor Email) 29

Table 7 Positive Z-Scores Selected from Hot Spot Polygons (Around Town Events Eamil) 29

Table 8 Female 25-34, Football Email, Not Normalized 75

Table 9 Male 25-34, Football Email, Not Normalized 76

Table 10 Female 35-44, Football Email, Not Normalized 78

Table 11 Male 35-44, Football Email, Not Normalized 79

Table 12 Female 44-54, Football Email, Not Normalized 80

Table 13 Summary of Lifestyle Segmentation 89

Table 14 Recommendations for Improving Response Rates 90

Table 15 - Segmentation of 45-54 males who received football email 121

Table 16 - Segmentation of 25-34 females who received corporate sponsor email 122

Table 17 - Segmentation of 25-34 males who received corporate sponsor email 123

Table 18 - Segmentation of 35-44 females who received corporate sponsor email 124

Table 19 - Segmentation of 35-44 males who received corporate sponsor email 125

Table 20 - Segmentation of 45-54 females who received corporate sponsor email 126

Table 21 - Segmentation of 45-54 males who received corporate sponsor email 127

Table 22 - Segmentation of 25-34 females who received Around Town email 128

vi

Table 23 - Segmentation of 25-34 males who received Around Town email 129

Table 24 - Segmentation of 35-44 females who received Around Town email 130

Table 25 - Segmentation of 35-44 males who received Around Town email 131

Table 26 - Segmentation of 45-54 females who received Around Town email 132

Table 27 - Segmentation of 45-54 males who received Around Town email 133

vii

LIST OF FIGURES
Figure 1 Study Area – Los Angeles County .................................................................................... 3

Figure 2 Workflow Overview ....................................................................................................... 20

Figure 3 All Email Recipients by Census Tract ............................................................................ 32

Figure 4 Dot Density Map of all USC Alumni Email Recipients ................................................. 34

Figure 5 Hot Spot Analysis, Female 25-34 ................................................................................... 36

Figure 6 Male 25-34 Hot Spot Analysis- Normalized ................................................................... 38

Figure 7 Hot Spot Analysis, Male 25-34 ....................................................................................... 40

Figure 8 Hot Spot Analysis, Female 35-44 ................................................................................... 42

Figure 9 Hot Spot Analysis, Male 35-44 ....................................................................................... 44

Figure 10 Hot Spot Analysis, Female 45-54 ................................................................................. 46

Figure 11 Hot Spot Analysis, Male 45-54 ..................................................................................... 48

Figure 12 Hot Spot Analysis, Female 25-34 ................................................................................. 50

Figure 13 Hot Spot Analysis, Male 25-34 ..................................................................................... 52

Figure 14 Hot Spot Analysis, Female 35-44 ................................................................................. 54

Figure 15 Hot Spot Analysis, Male 35-44 ..................................................................................... 56

Figure 16 Hot Spot Analysis, Female 45-54 ................................................................................. 58

Figure 17 Hot Spot Analysis, Male 45-54 ..................................................................................... 60

Figure 18 Hot Spot Analysis, Female 25-34 ................................................................................. 62

Figure 19 Hot Spot Analysis, Male 25-34 ..................................................................................... 64

Figure 20 Hot Spot Analysis, Female 35-44 ................................................................................. 66

Figure 21 Hot Spot Analysis, Male 35-44 ..................................................................................... 68

Figure 22 Hot Spot Analysis, Female 45-54 ................................................................................. 70

viii

Figure 23 Hot Spot Analysis, Male 45-54 ..................................................................................... 72

Figure 24 All Male 25-34 Football Email Recipients in Hotspots and in Pasadena ..................... 83

Figure 25 Male 25-34 Opened a Football Email in Hotspots and Pasadena ................................. 84

Figure 26 All Email Recipients in Long Beach ............................................................................. 85

Figure 27 Current USCAA Email Marketing Report (Harris Connect 2014) ............................... 87

Figure 28 Effect of Normalization ................................................................................................ 92

Figure 29 - Dot density map of all 25-34 females who received football email ......................... 103

Figure 30 - Dot density map of all 25-34 males who received football email ............................ 104

Figure 31 - Dot density map of all 35-44 females who received football email ......................... 105

Figure 32 - Dot density map of all 35-44 males who received football email ............................ 106

Figure 33 - Dot density map of all 45-54 females who received football email ......................... 107

Figure 34 - Dot density map of all 45-54 males who received football email ............................ 108

Figure 35 - Dot density map of all 25-34 females who received corporate sponsor email ......... 109

Figure 36 - Dot density map of all 25-34 males who received corporate sponsor email ............ 110

Figure 37 - Dot density map of all 35-44 females who received corporate sponsor email ......... 111

Figure 38 - Dot density map of all 35-44 males who received corporate sponsor email ............ 112

Figure 39 - Dot density map of all 45-54 females who received corporate sponsor email ......... 113

Figure 40 - Dot density map of all 45-54 males who received corporate sponsor email ............ 114

Figure 41 - Dot density map of all 25-34 females who received Around Town email ................ 115

Figure 42 - Dot density map of all 25-34 males who received Around Town email ................... 116

Figure 43 - Dot density map of all 35-44 females who received Around Town email ................ 117

Figure 44 - Dot density map of all 35-44 males who received Around Town email ................... 118

Figure 45 - Dot density map of all 45-54 females who received Around Town email ................ 119

ix

Figure 46 - Dot density map of all 45-54 males who received Around Town email ................... 120

x

LIST OF ABBREVIATIONS
GIS Geographic Information Science
USC University of Southern California
USCAA University of Southern California Alumni Association

xi

ABSTRACT
Permission-based email marketing is an inexpensive and effective method for companies to
develop an ongoing conversation with customers and to acquire new ones. Almost 80% of all
adult U.S. consumers use email each day. While the process of sending emails to customers is
relatively simple, interpreting email response rates to develop relevant, customized email content
is challenging. Companies who are able to offer a personalized email experience to their
customers report improved response rates and engaged customers. Email response rate metrics in
their current form tend to be one-dimensional and do not provide the location of who opens an
email or to what socio-economic group they belong. Existing studies focus on email response
rate data but do not include visualization of email recipients nor do they assign them to a lifestyle
segmentation profile. The purpose of this thesis project was to utilize hot-spot and dot density
analysis by census tract to identify spatial clusters of USC alumni email interaction in Los
Angeles County, and to conduct market segmentation of alumni who open and do not open
emails. The outcome of this thesis is to provide maps that allow marketing and event managers at
the USC Alumni Association to visualize alumni email recipients that respond to specific
categories of email in order to improve email marketing campaigns and to better position
resources for event planning.

1

CHAPTER 1: INTRODUCTION
The USC Alumni Association (USCAA) organizes events and programs throughout the year for
the sole purpose of ensuring that alumni, parents, and fans of the University of Southern
California (USC) remain engaged with the university. While the strategies for engagement vary,
the ultimate goal of the USCAA is for alumni to become reacquainted with the university
through these services and programs and, in turn, to foster a culture of ongoing philanthropy with
the university.
The programming used to achieve these goals ranges from educational seminars to career
networking and social events, such as networking mixers and football weekenders. In order to
promote upcoming events and opportunities, and to encourage participation in these events, the
USCAA uses email as its primary marketing tool for reaching approximately 150,000 alumni on
record with email addresses, 100,000 of whom live in Southern California.
1.1 Research Question and Study Purpose
For the purposes of this thesis, USC alumni are defined as USC undergraduate and graduate
students located in in Los Angeles County who graduated with a degree between 1980 and 2014
and were sent an email between February 2013 and February 2014 (Figure 1). This age range is
of primary interest to the USCAA marketing team because the age ranges are used in the
business processes of the USCAA. The research questions addressed in this thesis include the
following:
1. Are the USC alumni who open emails in Los Angeles County significantly clustered, and
do those clusters change based on email subject matter?

2

2. What segmentation lifestyle do alumni who open emails in Los Angeles County belong
to?
This study will use spatial analysis to visualize where USC alumni live in Los Angeles
County, specifically those who have opened an email between 2013 and 2014.The three email
categories in this study are:
1. USC Around Town, a monthly newsletter listing of USC related events throughout
Southern California (USC Alumni Association 2015)
2. USC football weekender emails (USC Alumni Association 2014) sent to promote USC
away football games.
3. Dedicated emails sent on behalf of the USCAA’s corporate sponsors for the purpose of
promoting a product or service.

3

Figure 1 Study Area – Los Angeles County

Hot spot analysis was conducted on both non-normalized data and normalized data in
order to compare the results. The actual counts of alumni who open an email are used as the
input value for the hot spot analysis. The counts of alumni are considered an accurate display of
where alumni receive and where resources need to be targeted. Normalizing data can decrease

4

the differences in data values based on the size of an area, but in some cases the outcome of data
normalization can also present results that do not support the intended goal of the study. A
previous study found that normalizing data of physical tests used to predict draft positions in the
National Football League made almost no difference when compared to the raw data (Robbins
2010). Output of all of the hot spot analyses conducted as part of this research are included in
this thesis for reference. Based on comparison of the results of both non-normalized data and
normalized data, the final maps and recommendations in the thesis are based on analyses
conducted using non-normalized data.
Marketing and maintaining a relationship with university alumni differs from general
retail marketing for which customer acquisition and brand recognition tends to be the primary
focus. In the case of the USCAA, the USC brand has been established as each student has
presumably spent years attending classes and events. Alumni do not need to be acquired since
matriculation and graduation instantly provide membership to the USCAA. The USCAA does
not charge a membership fee to be part of the association, nor does it ever directly ask for
monetary donations. Instead the USCAA is tasked with building relationships “lifelong and
worldwide” as noted in the association’s tagline.
The number of new alumni each year is fixed based on the number of graduating
undergraduates and graduate students at the end of each semester. There is also another group
consisting of non-alumni who have actively decided to align themselves with the university by
being a parent of a student, a donor, or a fan of an athletic team. However, for the purposes of
this thesis only degreed alumni were considered for this study.
While graduating students may value their time at USC and identify well with their alma
mater, there is no inherent reason for them to maintain any type of relationship with their

5

institution after graduation (McDearmon 2013). This is where an alumni association can step in
and work to continue building upon the relationship that was created as a student and promote a
life-long relationship. As such, alumni associations understand that it is easier to contact students
before graduation while they are still enrolled and to effectively develop student specific
programming for the purpose of enforcing the alumni brand (Clotfelter 2003)
A critical channel for maintaining a post-graduation connection for the USCAA is
through email. Upon graduation, student email addresses are transferred from the student
database and entered into the alumni database with the tacit understanding that new graduates
want to stay connected to their university. Six months after graduation, the @usc.edu email
address which students were given is turned off. For this reason and for older alumni who never
had access to an USC email address, the USC Alumni Association offers free lifetime email with
the suffix @alumni.usc.edu. Receiving email from the USCAA is not based on an opt-in system.
Rather, it is up to the alumni to remove oneself from the email list. Alumni already in the online
community are encouraged to update their contact information on a regular basis through event
registrations and memberships to regional clubs.
In 2014 the USCAA sent over 300 distinct emails to some or all of the USC alumni
population. Even with an opt-out system, alumni unsubscribe rates remain under .25% (USC
Alumni Association 2014) further confirming the fact that alumni prefer receiving email from
their alma mater, but only on matters that interest them (Performance Enhancement Group
2013).
The email distribution platform used by the USCAA maintains an open-rate percentage
and click-thru count for each email sent. These metrics provide a general indication of an email’s
effectiveness. However, when the open-rate is analyzed over the period of a year for each email,

6

the change in open-rate percentage varies by only two percentage points in either direction. With
current email statistics providing little distinction between what makes an email successful, the
marketing manager may not have all the information available for improving response rates.
USCAA segments its email distribution lists by zip code and/or graduation year primarily
because managers develop age-specific programming and because alumni regional clubs have
been developed using these same zip code ranges. In some instances, events are marketed to
clubs’ email lists because the event venue is simply located within the boundaries of the regional
club.
Census tracts were used as a boundary for spatially locating the alumni who opened
alumni related emails. Block groups were not used, as the purpose of this study is to provide
guidance to USCAA staff when creating email lists. For practical purposes, the smaller areal size
of a block group and accompanying resolution of demographic information would not provide
any added benefit to the analyses conducted in this study. The data for this thesis is based on Los
Angeles County zip code boundaries, which are in some cases even bigger than the census tracts.
In addition, the use of census tracts assists in maintaining a consistent boundary for market
segmentation using Esri Business Analyst software. Census tract borders also align with county
borders, which is convenient since the focus of this study is Los Angeles County. In the context
of this thesis, census tracts provide the polygon boundary needed to assign the points to polygons
when conducting hot spot analyses described in Chapter 3. While USCAA staff currently utilizes
zip codes for email targeting, zip codes are not unique to a county and the populations of zip
code can vary depending on the area of the zip code. The ability for USCAA staff to visualize
response rates of alumni based on age or gender, or the content of email, will help with
understanding who is reading emails and where event venues might enjoy better attendance,

7

through mapping alumni response rates in Los Angeles County. Lastly, census tract populations
average 4000+ consistently across the county. For all these reasons census tracts were deemed
the most appropriate discrete spatial boundary to be used in this thesis.
1.2 Study Area
This thesis focuses on Los Angeles County, California as the study area because 57% of USC
alumni continue to live in the county after graduation, providing a large alumni population that is
economically and racially diverse. Alumni who live in Los Angeles County make up 62% of the
current email addresses on file. This USC alumni population is geographically dispersed
throughout the 4000 square miles of Los Angeles County and its population of 10 million (U.S.
Census Bureau 2015).
The USCAA creates email lists primarily by zip code and occasionally with graduation
year or a combination of both when creating a list for an email campaign. However, for the
purposes of this thesis, census tracts were determined to be more relevant because the different
land areas that makeup zip codes do not allow the level of granularity needed when creating the
lifestyle segmentation profiles. Zip code segmentation works well for marketing campaigns
where the area of interest is larger, such as a state or region. But because the focus of this study
is Los Angeles County and since the majority of USC alumni continue to live in the county the
census tracts allow USCAA staff to identify individual neighborhoods where alumni email opens
are spatially clustered.
For the purposes of this study, the USC alumni population in Los Angeles County is
52,020. The dataset utilized in this study consists of 26,215 males and the 25,793 females, with
alumni from both groups all between the ages of 25-54. While the actual number of USC alumni
living in Los Angeles County is much larger and includes alumni younger than 25 and older than

8

54, the population studied represents those who have an active email address on file and received
an email from the USCAA in at least one of the email subjects during the study period.
The number of alumni in each age range can be seen in Table 1. The 35-44 and 44-54
year old groups have smaller populations due to the fact that email addresses weren’t available
during their time at USC. The average age of each group is available in Table 2.
Table 1 Thesis Population
Age Range Female Male USCAA Category
25-34 14,366 12,972 Young Alumni
35-44 7,221 7,580 Second Decade
44-54 4,833 5,048 Encores

Table 2 Thesis Population Mean, Max, and Min
Age Range Mean Max Min
Female 25-34 28.4 34 25
Female 35-44 38.6 44 35
Female 45-54 48.9 54 45
Male 25-34 29.6 35 25
Male 35-44 39.2 44 35
Male 45-54 49.0 54 45

1.3 Organization of this Thesis
The remaining chapters in this thesis cover the background research on related efforts, the
methodology developed to perform the relevant spatial analyses, the results of the analyses

9

conducted as part of this thesis presented in tables and maps, and a summary discussion of these
findings as well as recommended future work on this project.
Next, Chapter 2 begins with a discussion on the relevance of marketing and its transition
from a one-way form of communication between companies and consumers to the pursuit of an
ongoing conversation. The chapter then describes the ways in which market analytics, and
specifically market segmentation has proven to make sense of the voluminous amount of data
available to companies and whether ecological fallacy is an acceptable risk. Chapter 2 concludes
with a section discussing the effect permission-based email can have on customer loyalty and the
role GIS has played in supporting university advancement.

10

CHAPTER 2: BACKGROUND
Although an array of digital communication channels such as Facebook and Twitter have
become popular and widely used, email is still highly relevant as a marketing tool for businesses
and organizations (Laroche, Habibi, and Richard 2013). Email is still used by 79% of all U.S.
adult customers each day (Fisher 2014). Email is also a cheaper medium than conventional forms
of marketing such as direct mail, radio, and television. In fact, over the next three to five years,
25% of businesses in the United States plan to increase spending in the areas of permission-
based email marketing and marketing analytics (eMarketer 2014).
Marketing analytics research has shown that collecting relevant customer data points and
using this information to develop actionable customer profiles can assist companies in
understanding their customers (German 2014). Market segmentation moves beyond collecting
and sending numerous emails to customers in a company database to interpreting customer
expectations and identifying their life experiences (Coelho and Jörg Henseler 2012). Companies
and organizations have found that the more they know about their customers, the better they can
be at adjusting their advertising tactics to learn the preferences, modes of communication, and
products most enticing to their customers.
In an effort to gain an edge on competitors, businesses and organizations have turned to
marketing analytics to distinguish themselves from their competitors in the hope of making a
lasting connection with customers (Greene and Greene 2008). Ironically, because customer data
has become so readily available, some companies are finding that the competitive edge that
marketing research used to give has narrowed (Kiron 2012). In this new environment innovative
strategies such as relationship marketing, marketing segmentation, and geodemographics are

11

used in conjunction with a company’s historical data to enhance the outcome of marketing
campaigns.
In recent years relationship marketing has become popular because it allows for the one-
on-one communication that businesses believe resonates most with potential customers (Bulger
1999). Relationship marketing is now possible because of technological advances and the
abundance of consumer data. Data from disparate sources, such as point of sale, online
transactions, mobile devices, and the U.S. Census and surveys can now be stored in vast
quantities and queried in support of decision-making platforms. According to a recent study, the
number of contact occurrences was a prime indicator of whether a person would become or
remain a customer (Rodrigue, 2012). In other words, in the case of a university, the more an
alumnus was contacted, whether by phone, email or mail, the more likely and that alumnus was
to give a donation.
When email marketing at the USC Alumni Association first began in the 2000’s, the
emphasis was on the number of viable email accounts, and obtaining more. As email tools and
staff became more sophisticated there was an ongoing attempt to segment email lists based on
the employee’s personal knowledge of the event and the belief that sending more emails was still
the best strategy. Over time, there were signs that alumni might have been suffering from email
fatigue as click-thru and open rates have remained flat for the past five years. To address the
stagnant response rates, the alumni association staff began to segment their lists the best they
could with the information available.
Modern marketing techniques must be able to make sense of the vast and varied customer
information collected in order to promote a company’s brand rather than a singular product.
Visualizing alumni email response rate is the main focus of this thesis, specifically as it relates to

12

the USCAA marketing team. A review of past similar efforts follows in this chapter, synthesized
and organized to assist in answering the research questions stated in Chapter 1.1. During the time
frame of this thesis literature, associated with visualizing customer response rates from emails
was found to be scarce.
2.1 Effect of Permission-Based Email Marketing
Marketing is often associated with the notion of selling a product to a customer.
However, non-profit organizations engage in marketing to develop relationships (German 1997)
in the hope that that the relationship will yield some sort of commitment, whether as monetary or
other forms of future engagement. The ultimate goal in relationship marketing as it pertains to
non-profit organizations is to create an ongoing conversation between the constituent and
organization (Andreasen 2012). New strategies involve creating a dialogue between the
consumer and the company and transforming it into a partner relationship where both sides feel
they benefit from engaging with one another.
In a non-profit organization such as the USCAA, the customers, or alumni in this case,
have already been customers for some time. Alumni made their purchase years before as a
university attendee. While an individual alumna might be a graduate of more than one university,
it is more than likely that their loyalty to each school is compartmentalized. It is the perceived
relationship between the individual and the entity that encourages continued loyalty.
As an example of a study similar to the focus of this thesis, in a study of Sony Ericsson
Club members, researchers set out to find whether Sony loyalty cardholders were more likely to
make a purchase if club members clicked on links in emails sent by Sony Ericsson (Hasouneh
and Alqeed 2010) The campaign focused on its registered club members for which it had general
data about their customer’s name, address, email address and recent purchases. The study is

13

significant because the company began to experiment with relationship development and email
responses as a method to encourage ongoing communication in lieu of direct marketing.
The relationship marketing that Sony Ericsson created included targeted emails and
campaigns to specific groups within the loyalty card program based on their purchasing history.
This study focused on customer responses to emails sent by Sony Ericsson and tracked the email
click-thru rates and attached those responses to customer data. Although Sony Ericsson did not
include a spatial analysis or visualization (such as a map), their study did find that customers
who clicked on an email link were more likely to stay engaged in some manner with Sony
Ericson in the future.
2.2 Market Segmentation
Market segmentation places customers into broad but homogeneous groups for use in more
efficient marketing campaigns and to improve the return on investment. A byproduct of market
segmentation is the need for better customer data that will allow the organization to place their
customer base into smaller groups in which general assumptions about their purchasing habits or
lifestyle preferences can be made (Jarratt and Fayed 2012). Market segmentation as it applies to
email is not always practiced because permission-based email marketing is a low-cost medium
with little financial reason for not sending emails to as many customers as possible. This is
especially relevant in the case of a university where alumni have indicated they are not contacted
enough even though they receive an email almost every day (Performance Enhancement Group
2013).
With the growth of technology, market segmentation gained popularity in the early
1970’s. In the late 70’s and 80’s companies began early attempts at segmentation that were rigid
and static, mainly using zip codes as the boundary for geographic segmentation (Lien 2005). To

14

identify potential markets to target, processes were developed so that data not normally
associated with identifying customers, such as socio-economic indicators, spatial and temporal
data,was appended to visualize potential customers and retail locations, as well as early attempts
at customer segmentation.
The addition of spatial data enhanced customer data and allowed organizations to locate
prime areas for new businesses, improve delivery routes, and visualize where customers live
(S.M. Musyoka et al. 2007). Spatial data can be used to make better informed decisions about
retail locations siting, optimal locations for outdoor advertisements, and customer location to
name a few. In this way, a geographic information system (GIS) can become a decision support
tool for businesses and organizations (Ghita 2014).
2.3 Alumni Response
During a fund-raising campaign, universities rely heavily on their alumni base as the source for
meeting their advancement goals. In 2014, colleges and universities in the United States received
$37.5 billion in private donations, of that total, alumni gave 26 percent (Case Advancement
Study 2014). Institutional or university advancement consists of programs and services designed
to develop ongoing relationships with alumni and friends. Developing those relationships relies
on marketing and communication throughout an alumni lifecycle that begins at graduation and
hopefully continues for decades thereafter.
While the goal of any university is for alumni to give back, it may take years after
graduation for an alumnus to actually do so. In those interim years where there is no financial
interaction between the alumni and the institution it’s often up to the alumni association to
initiate an ongoing conversation where alumni are urged to participate, either through attending
events or providing donations.

15

Alumni are more likely to continue to be engaged with their university if they have a
positive view of the university and the alumni association (Newman and Petrosko 2011). When
graduates of a university are many years out of college, their only connection with their alma
mater might be email communications. Whether those communications have a meaningful
impact on alumni is a factor in alumni participation.
Alumni giving is most associated with an older alumni population who are more likely to
have the financial means to donate. And based on preliminary mapping presented in this study,
the areas and neighborhoods in Southern California in which they reside are fairly predictable.
However, there is a far greater number that don’t respond to emails and the ongoing
communication which the university tries to develop through email is eventually lost. Since
education giving was the second highest recipient of all donations in 2010, any improvement in
targeting alumni will be a benefit to the USCAA (Rodrigue 2012).
2.4 GIS and University Advancement
In the context of this study, GIS is used for university advancement in understanding where
potential donors are clustered within the LA Country. According to Steven German (1997),
individuals get involved with non-profit organizations for three reasons; characteristics of the
individual, a perceived respect of the institution, and personal gain. Of these three factors,
marketing and communication has the strongest effect on perceived respect of the institution.
Although this study identifies the individual characteristics age, location, household income and
other socio-economic identifiers (German 1997), it does not expand on whether those most likely
to get involved were spatially clustered in any way. In the context of this thesis, clustering is
defined as a geographical area where a significant number of emails are opened by USC alumni
and the reason those emails were opened is not random.

16

In a study similar to this thesis work, the State University of New York (SUNY) used
GIS to generate maps of each state in the US showing percentage of total giving and average of
total giving by state. The SUNY maps also depict where alumni are located by state, county, and
congressional districts. (Jardine 2003). SUNY used the maps to visualize where alumni were
located and how to best position staff and marketing resources to reach the alumni. The results
from the SUNY study indicated that the output they obtained was beneficial to their university
advancement staff and provided encouragement that visualization of alumni by age could be
useful.
According to a more recent study of alumni donors, the number of contact occurrences
was a prime indicator of whether a person would give (Rodrigue, 2012). In other words, the
more an alumnus was contacted, whether by phone, email or regular mail, the more likely and
that alumnus was to give a donation.
2.5 Normalized versus Non-Normalized Data in Marketing
In another study to map population densities of census data, data was normalized to account for
the changes in census tract population over time (Holt, Lo, and Hodler 2004). That study
intended to demonstrate a new technique for mapping dasymetric population densities using
census data and areal interpolation over three decades. To account for the variations in alumni
populations per census tract, the input data for this study was normalized by the total number of
alumni who received an email by census tract. Also to account for the changes in population over
time, census tract populations had to be normalized so later years in the study could be compared
with earlier years. Thus normalizing by census tract allowed tracts with very few alumni to be
compared with higher populated tracts across the time span of this study.

17

2.6 Ecological Fallacy
A study by William S. Robinson in 1950 introduced the idea of ecological fallacy, an important
consideration in this study. His paper suggests that the primary difficulty lies is making
assumptions or developing inferences about a subset or individuals within a larger group.
Because market segmentation essentially aggregates census and other available data to make
inferences about individuals in a larger group, the concept of ecological fallacy is a concern
(Magliozzi, Berger, and Clancy 2003). Yet, from a business perspective it’s not economically or
logistically feasible for a company to create an individual profile of each customer in a targeted
area. There has to be a distinction between ecological fallacy as a real concern in areas such as
social and epidemiological studies, for example, and as an acceptable risk when making business
decisions. The question isn’t whether an organization is going to spend marketing dollars, but
rather, what is the best way to spend those dollars.
In many cases, companies and organizations generally know the kind of customer they
want to attract. Thus marketers use segmentation to make inferences about customers and to be
more strategic in their planning and execution. Both private and public organizations use market
segmentation data to better understand the customers they serve. Market segmentation has been
referred to as the classification of people based on where they live (Singleton 2004). This
practice allows organizations to make more accurate assumptions about groups of people living
in a certain locations at the neighborhood level.
In order to create customer profiles, most segmentation software generally uses U.S.
Census data as its foundation (Esri 2012) Segmentation data typically covers the entire
geographical area of the United States, and is broken out at many levels, including county, block
and tract. Given that the data is used as a critical resource in a variety of fields - from education

18

and housing to transportation and healthcare - it is also of significant value in assembling
segmentation profiles.
Another important factor when determining the risk associated with ecological fallacy is
to remember that profile segmentation is considered a tool to assist the user in making informed
decisions about customer lifestyles based on their location in the socio-economic strata. In
essence, the entire field of market segmentation and geodemographics is associated with
ecological fallacy (Singleton 2004). In this study, segmentation is utilized to enhance the typical
email list stored in a database and to better organize alumni inside groups that have similar
interests and lifestyles. Thus in this thesis, market segmentation is presumed to not be prone to
ecological fallacy since the primary goal of segmentation is to place customers into broad groups
to make accurate as possible assumptions about them.
2.7 Hot Spot Analysis in Business
In a study of districts in Istanbul, Turkey, hot spot analysis was used to assist the city
planning commission with identifying clusters of customer and transportation activity which
eventually led to improvements in logistics and delivery of good and services (Wang and
Varady 2005). In this study the authors were able to visualize where the most deliveries
occurred and where trucks and loading facilities were located within the city. By identifying the
hot spots the maps were useful to planners as a tool used to centralize goods and resources. . In
another similar study, using historical taxi pickup location data researchers proposed analyzing
clusters of customer pickup requests in Taiwan to see how demand changed in certain parts of
the city based on temporal transportation data and past weather forecasts. (Han-wen Chang
2010).

19

This literature review indicates that researchers can use hot spot analysis for predicting
requests from customers in order to serve them more efficiently. Of the studies discussed herein,
the Sony Ericson by Hasouneh described in Chapter 3 provided some guidance in the sense that
it validated that the study of email response rates can be useful in supporting a business process.
In the following Chapter 3 Methodology, insight gained from the literature review was utilized to
develop the analyses discussed in detail herein.

20

CHAPTER 3: METHODOLOGY
Since previous studies which involve visualizing customer response rates based on email were
found to be scarce in the literature, the methodology described in this chapter is based the
author’s experience in communicating with the USCAA marketing team to determine the output
needed to create better email marketing lists for staff. This thesis utilizes GIS to visualize where
alumni in Los Angeles County open emails through the use of dot density and hot spot analysis.
Alumni are first located as living in either in a “hot” or “cold” census tract, then are assigned a
lifestyle segmentation profile based on the census tract they live in. Figure 2 depicts an overview
of the workflow used in this study, described in detail in the following sections of this chapter.

Figure 2 Workflow Overview

3.1 Alumni Data Management
The USCAA sends emails to alumni that span a wide range of subject matters. For the purposes
of this study, emails which were sent between 2013 and 2014 to the entire USC alumni
population in Los Angeles County will be used. Furthermore, those emails were categorized into

21

the following groups: an email promoting USC football weekenders, a USCAA event newsletter
sent to Southern California residents (Around Town), and emails sent on behalf of corporate
sponsors. This study utilized spatial analysis to visualize where USC alumni live in Los Angeles
County, specifically those who have opened an email between 2013 and 2014.
Alumni data for this thesis is maintained in a SQL Server database that captured email
responses as well as updates to alumni records between 2012 and 2014. In all cases, the email list
was modified to remove any recipients without a Los Angeles County address, allowing only the
county to be sampled and visualized. Alumni names and any other personal information were
removed from this database prior to importing into Esri ArcMap, and subsequent data processing
steps required for this thesis work.
3.2 Spatial Analyses
The first step in this research was to visualize where alumni are living in Los Angeles County. A
county shapefile (Los Angeles County 2014) was used as a container to visualize alumni
residential locations by census tracts. All alumni living in Southern California were geocoded
using the Geocode Addresses tool in Esri ArcGIS.com (2013) and spatially joined to the county
shapefile of census tracts to create a point layer of alumni home locations.
Addresses used in this study are based on the home address and, consequently, assumes
that all email activity took place at the home of the recipient. It is quite possible, however, that
the user opened the email at a location other than his residence. But for the purpose of this study,
the assumption is accepted, because the home address provides the best opportunity to evaluate
the socio and demographic characteristics of the users (Hasouneh and Alqeed 2010) utilizing
current census (U.S. Census Bureau 2010) and segmentation data (Esri 2012).

22

3.2.1 Email Data Preparation
For this study, alumni email data was organized in multiple tables. Whenever email is sent, a
new table is created and the response data is captured in that table. Email response data is then
joined to each alumnus record using the unique identification number. The geodatabase contains
a table with one row per alumnus and another table with the email response data; the tables are
associated using the unique identification number that links the individual alumnus with the
actions they took while viewing the email. In order to perform multiple queries without creating
additional tables, a Microsoft SQL Server connection was created within ArcMap to connect the
database table with the GIS mapping capabilities. The output from these queries created
temporary views of the table that were exported to the geodatabase in ArcMap for analysis.
The email open data is captured in separate tables created for each email. For example,
when the monthly newsletter, Around Town is sent out, a separate activity summary is created in
the USCAA system. The activity summary organizes the email bouncebacks, click-thru counts,
and open rates in three separate tables.
The open-rate percentage is defined as the rate at which someone opens or views the
email and is calculated by dividing the total count of opens by the total count of those who were
sent the email. To achieve the rate metric, a small transparent image is placed in the body of the
email. When the email is opened the image, along with the unique identifier, is requested from
the server and the request is captured by the database as a single open of a given email. For this
metric to be counted, the email client must be able to download the images contained within the
email.
Rather than create a table for each email content type and associated “opens”, a single
table is created which contains all of the individual email “opens” with two additional columns

23

containing the email content type and the unique alumni identification number. The table
contains count data for any user who opened an email; however, if any duplicate entries for a
given email (based on subject, date and time sent) exist within a single email content type, they
were removed.
Email activity as it relates to a given organization is somewhat unique since it is based on
the levels of interaction amongst end users (Hasouneh and Alqeed 2010). For example, a
company with higher click-thru rates and open rates will have a different threshold for success as
compared to an organization with less activity. In the case of the USCAA, email open rate
consistently measures in the 18%-21% range regardless of content.
For the purposes of this thesis, open rate participation will be categorized by high,
medium, and low ratings. By comparison, open rates for non-profit educational organizations
average 33.5% (Constant Contact 2015). The percentages defining open rate activity for the
USCAA can be seen in Table 3.
Table 3 Email Open Rate Categories (USCAA 2014)
Low
Medium High
<=17% 18-22% >=23%

While open rate is applied to a single email, for this thesis, a single open is associated
with a single alumni and visualized in terms of the type of email sent. For example, the emails
with corporate partner related content are combined together and clusters of activity across age
groups are visualized. Additionally, another layer containing the counts from football related
email content is compared to see if there are any changes in interest by age group or location.

24

In order to visualize the difference in activity between multiple demographic groups, the
email data was further categorized into age groups. The groupings match the generational groups
the USCAA uses to plan programs and events for alumni as depicted in Table 4.
Table 4 USCAA Generational Groups (USCAA, 2015)
USCAA Generational Groups Age Range
Young Alumni 25-34 years old
Second Decade Society 35-44 years old
Encores 45-54 years old

While the Around Town and football emails primarily advertise upcoming general USC
and alumni events, the corporate sponsor emails contain almost no content directly related to
USC or the alumni association. Although the corporate sponsor emails always contain USCAA
branding, all of the content is related to the corporate sponsor and any links in the email forward
the user to the website of the sponsor (Appendix A).
3.2.2 Normalized Alumni Response Data
In preparation for this analysis, the alumni response data were normalized since the total
alumni email recipient population counts and the actual counts of alumni who opened an email in
each census tract varied between 2012 and 2014. Normalization takes into account the
differences in scale or varying sizes of areas of the census tracts so that the email open rate of
large census tracts with large populations and hence large numbers of alumni can be compared to
smaller tracts with smaller populations (Holt, Lo, and Hodler 2004). Thus to provide a better
picture of alumni response, the normalized values were obtained by dividing the alumni response
of each census tract by the total number of alumni who were sent the email:

25

Females, between the ages 25-34 who opened an email from corporate sponsors
All females, between the age 25-34 who received a corporate sponsor email

Normalizing the alumni data essentially evened out the population counts of the census
tracts based on area. Hot spot analyses of census tracts with very low alumni populations but
higher open instances may show up in the results as hot spots. While this result would be
accurate based on the values being used, in the end identifying hot spots in areas where there are
very few alumni turned out might not to be helpful to alumni staff seeking to attract larger
numbers of alumni to events. To compare results and to ensure that resulting maps are useful to
USCAA staff, hot spot analysis was also performed using the actual counts of alumni who
opened an email (non-normalized) as the value for the incident aggregation.
3.2.3 Alumni Staff Survey
A survey (Appendix B) was created and given to seven marketing and communications
employees at the USCAA to elicit thoughts on the current email list segmentation process. The
survey was created and delivered to seven USCAA employees online.

26

The survey attempted to answer the following questions in order to provide guidance when
creating the final maps for this thesis:
1. Which criteria are most useful when segmenting an email list?
a. Age
b. Location
c. Preferred Graduation Year
d. Preferred School
2. If you could use a map to help in segmentation, what kind of map would it be?
a. Choropleth Map
b. Hot Spot Map
The survey results indicated that the most valuable trait when creating email lists is the age of the
alumni, followed by zip code. Census tracts garnered less interest but upon interviews conducted
after the survey indicated that staff were unfamiliar with census tracts. When asked about which
maps would be most useful, alumni staff selected the cholorpleth maps over the hot spot maps.
The survey respondents said the chloropleth maps were easier and faster to understand, however
after explaining how hot spot maps were created alumni staff thought they might be useful as
well.
3.2.4 Dot Density Analysis
Dot density analysis was performed on the entire population of email recipients as well as those
alumni who opened an email in order to visualize the concentration of alumni email opens and
compare that with the output from the hot spot analysis (Esri 2014b). Dot density analysis uses a
one-to-many map type to represent the density of the geographic points in the data based on the
count field, in this case number of alumni, in each polygon or census tract. Instead of showing

27

one point per location, a single dot aggregates a predefined grouping of alumni. The purpose of
dot density mapping is to consider whether the results of the hot spot analysis are indicative of a
normal spatial distribution of alumni living in Los Angeles County. While the output from the
dot density maps will show where alumni live via the dot values, it could be that the dot values
are merely representative of the general populations of the census tracts.
3.2.5 Hot Spot Analysis
Hot spot analysis was conducted using the Esri ArcGIS Optimized Hot Spot tool found in Esri
Arcmap (Esri 2014a). Hot spot analysis is used to identify the neighborhoods surrounding the
features, in this case the census tracts, where alumni open alumni emails. Identification of hot
spots determines where the clustering of those opening an email is not a random occurrence.
Determining the causality of clustering is not the purpose of this thesis, but rather to discover
where spatial clustering exists.
The Optimized Hot Spot tool collects the parameters needed to execute the tool from the
alumni and email response data. In order to aggregate the incidents of alumni open rate data, the
Esri-named variable Count_Incidents_Within_Aggregation_Polygons was used to ensure that
more than 30 features were available. Points representing alumni home addresses of those who
opened an email were aggregated in order to create weighted features, which in this case are the
census tracts. Each analysis contained at least 1,800 features (census tracts). Because the values
were incidents of emails being opened there is no entry needed in the Analysis Field. This tool
aggregates the points into counts for each polygon. Rather than have the software provide
polygons based on the average nearest neighbor (ANN) distance or median nearest neighbor
(MNN), a polygon feature layer of Los Angeles County census tracts was used as an overlay to
contain and count the point data in each polygon (census tract).

28

The Optimized Hot Spot tool defines the appropriate scale of analysis by examining
where the Z-score reaches its highest point and the distance associated with it. A Z-score is a
standard deviation (Esri 2014a). The tool uses Global Moran’s I statistic and examines ever
increasing distances via Incremental Spatial Autocorrelation. By examining different distances
and comparing Z-scores with each distance, the tool will identify a point of high clustering along
with a distance where spatial clustering is most pronounced. For example, the optimal fixed
distance band based on peak clustering for the Male 45-54 group who opened the Around Town
newsletter is 5,558 meters. Because one goal of this thesis is to determine where hot spots occur
in Los Angeles County, a distance band of 5,000 meters, for example, is considered applicable
since the size of study area is not at the individual household level but rather at the census tract
level.
Hot spot analyses executed on each age group of non-normalized and normalized alumni
counts and email content type resulted in 18 output features classes in the form of polygon
layers. The Z-scores from the hot spot analysis are provided in Tables 5, 6 and 7:
Table 5 Positive Z-Scores Selected from Hot Spot Polygons (Football Email)
Z-Score
Female 25-34 2.0123-8.7771
Female 35-44 2.224-7.151
Female 45-54 2.000-7.170
Male 25-34 2.000-9.460
Male 35-44 1.864-9.014
Male 45-54 2.093-6.654

29

Table 6 Positive Z-Scores Selected from Hot Spot Polygons (Corporate Sponsor Email)
Z-Score
Female 25-34 1.901-9.315
Female 35-44 2.064-8.377
Female 45-54 2.011-8.080
Male 25-34 1.969-8.981
Male 35-44 1.953-10.195
Male 45-54 2.1554-6.426

Table 7 Positive Z-Scores Selected from Hot Spot Polygons (Around Town Events Eamil)
Z-Score
Female 25-34 2.004-8.607
Female 35-44 2.016-6.982
Female 45-54 2.00-7.302
Male 25-34 2.026-8.387
Male 35-44 2.013-7.730
Male 45-54 1.987-6.661

Almost all groups had Z-scores ranging from +2.00-10.195. A Z-score of 2.064 indicates
that the female aged 34-44 group who opened a corporate sponsor email are 2.046 standard
deviations from the mean. To put this into perspective, the level of statistical significance at the
95% confidence rate is less than or equal to 1.96. This means that there is a 95% probability that
the feature in question will have a Z-score between -1.96 to +1.9.6. Yet in this study, there are Z-
scores as high as 10.195 with very low p-values. The higher Z-scores and the smaller the p-

30

value, or probability, suggest that the spatial distribution of the points in the dataset are not
random and that there is an underlying reason for clustering in these areas.
As the purpose of this thesis is to provide guidance for alumni staff when creating email-
marketing campaigns, developing broad profiles of alumni is necessary for creating group-
specific content. To accomplish this, each of the alumni associated with an email open within the
census tract polygons identified as hot spots (Tables 3, 4, 5, 6, and 7) were input into Esri
Business Analyst Tapestry Segmentation for viewing the lifestyle segments. The results of the
hot sport analyses and segmentations for these alumni are discussed in detail in Chapter 4.

31

CHAPTER 4: RESULTS
This chapter describes the location of the alumni population examined in this study as well as a
discussion about the use of census tracts as a boundary in such an analysis. A baseline
visualization of where alumni with active email addresses live in Los Angeles County is
provided. This chapter reviews the dot density and hot spot analysis results and profile
segmentation applied to alumni census demographic information. The alumni data used in the
hot spot and dot density analyses were mapped using age and gender data. USCAA event
programming is determined by the stage of life that alumni are in. In this case, the age groups
correspond to the group classifications created by the USCAA, including: Young Alumni (25-
34), Second Decade (35-44) and Encores (45-54). A summary of how these detailed results can
be aggregated and used by the USCAA for marketing campaigns is then provided in Chapter 5.
4.1 Identifying Patterns in Alumni Locations and Email Responses
To provide a baseline of where alumni reside in Los Angeles County, alumni home locations of
all male and female records with email addresses are visualized in Figure 3. In general, census
tracts with greater than 70 (count) alumni are found on the west side along the coast and in the
San Gabriel Valley region. Areas with fewer than 70 alumni per census tract are located on the
southeastern and the most northern sections of the county.

32

Figure 3 All Email Recipients by Census Tract
As described in detail in Chapter 3, all hot spot analyses were run using the count field
for each data set. The count field is the field which contains the number of emails opened per

33

census tract. Because each age group and gender was contained in separate tables in ArcMap, a
count of alumni who opened an email could be used as the analysis field. The first round of hot
spot analysis was conducted using normalized data. After reviewing these results, a second round
of analyses was done using non-normalized data because normalized data resulted in cold spots
where a high number of alumni received alumni emails but did not open them. This occurred
because the normalization resulted in a ratio where the denominator is larger than the numerator,
in most case resulting in a number of less than one. In comparison, normalized cold spot census
tracts may contain very few alumni yet if only a few alumni in these tracts open an email the
resulting output is a number equal to one. In these cases, there were multiple census tracts with
one or two alumni where each time the alumni opened the email.
4.2 Dot Density Maps
In this study each dot in Figure 4 represents approximately 50 email recipients, one
representative example map from the dot density analysis. Appendix C beginning on page
Error! Bookmark not defined.103 contains the rest of the dot density maps for all USCAA
email recipients from each age group and email content type considered in this study. Each dot
represents ten alumni who were sent an email and have an address on file with the USCAA. The
results shown in the dot density maps display where alumni are clustered by count throughout
the county by census tract. The resulting maps are fairly consistent in their output, with little
variation between maps. The densest locations of alumni are on the west side of Los Angeles
County, running along the coast from Santa Monica to Palos Verdes. Denser areas continue west
through Hollywood, Downtown Los Angeles and into the San Gabriel Valley.
Based on the quantity assigned to the dot in the map, there are many census tracts along
the southern central area of the county that appear to have almost no representation. One purpose

34

of these maps is to assist marketing managers in making informed marketing campaign
decisions. The purpose of the density maps is simply to visualize where alumni live. Thus the dot
density maps are considered a first step for alumni staff to visualize where there emails are being
sent.

Figure 4 Dot Density Map of all USC Alumni Email Recipients

35

4.3 Email Respondents – Football Related Email
The following discussion covers the results of the analyses of the non-normalized and
normalized alumni data grouped in the age groups described in Tables 1 and 2.
4.3.1 Female 25-34
The female 25-34 year old age group shown in Figure 5 consists of 14,188 alumni, of which
4,010 opened a football related email. Of those who opened the email, 379 census tracts appear
as hot spots in Santa Monica, continuing across Culver City, Beverly Hills and into West
Hollywood then northeast to the 101 Freeway. Another significant hotspot occurs in Manhattan
Beach, Hermosa Beach and Redondo Beach. The only other hotspot areas for this age group
were in downtown Los Angeles and in the Pasadena, South Pasadena and La Canada areas.
While the hotspots were essentially confined to these three areas, the cold spots were dispersed
throughout the county with very few were in contiguous census tracts. Significant cold spot areas
are located in North Hollywood, Inglewood, El Monte and Covina.

36

Figure 5 Hot Spot Analysis, Female 25-34

Both the male and female populations in the 25-34 age groups have the largest number of
alumni compared to the two other age groups in this study. The ratio between those who received
an email versus those who actually opened an email is much greater and creates large areas of

37

cold spots. In this case, the normalized data results are essentially the reverse of the non-
normalized data since a large number of census tracts are identified as cold spots extending along
the coast and into Santa Monica, all the way to eastern San Gabriel Valley. Hot spots are
dispersed along the southern border of the county and in the Pacoima area, as seen in Figure 6.
The reason for the reversal is assumed to be due to census tracts with comparatively small
alumni populations along the southern and eastern sections of the county, yet those same tracts
also record high open rates yet are not indicative of where alumni are actually opening emails in
larger numbers. Thus hot spot analysis maps of non-normalized data are included in the
remainder of Chapter 4.

38

Figure 6 Male 25-34 Hot Spot Analysis- Normalized

39

4.3.2 Male 25-34
The male 25-34 year old age group in Figure 7 shows 12,259 alumni of which 3,529 opened a
football related email during the study period. While the significant hotspots in Santa Monica,
Venice, Century City and West Hollywood are similar to the female age group (Figure 5), the
male 25-34 population also exhibits a cluster in the Mid City area bordered by Vermont and La
Brea Avenues, and another large cluster still in the University Park Campus and downtown.
Unlike the female group, there are no significant hotspots in the Pasadena, South Pasadena or La
Canada.

40

Figure 7 Hot Spot Analysis, Male 25-34

Conversely, when analysis is performed on normalized data, the cold and hot spots
essentially reverse similarly to Figure 6. Large swaths of cold spots exist from Santa Monica,
through Mid City and into Downtown. The cold spots reappear in Pasadena and also in Hermosa

41

Beach and Palos Verdes. Hotspots of normalized data appear in Southeast Los Angeles, North
Hollywood and Covina.

4.3.3 Female 35-44
The non-normalized data depicts hotspots in Santa Monica, Culver City and Venice as seen in
Figure 8. However, this group has a larger representation in Torrance, Hawthorne, Manhattan
Beach and Redondo Beach. This group also displays hotspots in Pasadena, but the hotspots
extend further east past Pasadena, Sierra Madre, San Marino and the western section of Arcadia.
While there are cold spots, there is not one well-defined area of contiguous census tracts.
The areas with the most census tracts identified as cold spots are El Monte, Covina, Mission
Hills and Pacoima.

42

Figure 8 Hot Spot Analysis, Female 35-44

Again conversely, the normalized data for this same age group displays cold spots in
Pacific Palisades, Santa Monica, Brentwood, Bel Air, Beverly Hills and the some tracts in West

43

Hollywood. Cold spots also occur in Pasadena and La Canada. Hotspots of normalized data
occur in Inglewood, Downey and Santa Fe Springs.
4.3.4 Male 35-44
This population consists of 2,945 alumni who opened a football related email. The non-
normalized hotspots again appear on the west side of the county and in Pasadena, San Marino
and La Canada (Figure 9). However, this is the first age group where more hot spots appear in
the Palos Verdes peninsula, albeit in the northern most portion. This is also the first group who
received a football related email and where a large contiguous grouping of 291census tracts
appears as cold spots. These cold spots run along the southern border of the county from Long
Beach, Norwalk, Hacienda Heights, West Covina and into San Dimas. Prior groups exhibited
cold spots with the non-normalized data, but the cold spots in this group suggest that these are
areas not populated at all by this age group.

44

Figure 9 Hot Spot Analysis, Male 35-44

When the data on males 35-44 is normalized there are very few hot spots. The few hot
spots found are mainly centered around Downey and into West Covina, but the cold spot area is
quite large as it stretches from Santa Monica all the way into West Hollywood. This means that

45

although a large number of alumni live in these areas, when compared to surrounding groups
they are less inclined to open an email.
4.3.5 Female 45-54
This population is composed of 1,316 alumni who opened a football related email (Figure 10).
There were 199 census tract areas that registered as hotspots. In this population, the hotspot areas
are less widespread and mainly focused in the Palos Verdes, Pasadena, Altadena, San Marino
and Arcadia areas. The number of hotspot census tracts from the Santa Monica area are
significantly smaller when compared to the younger age groups.
The number of census tract cold spots is 198, and there appears to be no single area
where cold spots are concentrated. The overall smaller areas in terms of both hot and cold spots
can be attributed to the fact that less than a third of the alumni population in this group appeared
interested in a football related email.

46

Figure 10 Hot Spot Analysis, Female 45-54

When the analysis is run on normalized data for female alumni age 35-44, the hot spots
all but disappear. There are a few scattered census tracts in Norwalk, Palmdale and Northridge

47

but no significant areas where a group might be targeted for marketing purposes. The cold spots
appear in Santa Monica, but unlike all other groups so far, there are no cold spots in Pasadena,
San Marino or Arcadia. Instead the only cold spots are the areas in upper Altadena and part way
into Sierra Madre.
4.3.6 Male 45-54
The male age 45-54 alumni population consists of 1,883 alumni who opened a football related
email in 181 census tracts. Hotspots that appeared along the coast towards and including Palos
Verdes and the Pasadena area appear in this group as well. However, the significant difference in
this group, is that Santa Monica is not as widely represented as it had been with younger age
groups, as seen in Figure 11. The clustering for this group moves slightly north of Santa Monica
and is concentrated in Brentwood, Bel Air, Pacific Palisades and the Topanga Canyon.
The cold spots appear in the Wilshire and Los Feliz area and downtown Los Angeles
between 2
nd
Street and Pico Boulevard to the north and south, and from the 110 to Los Angeles
Street.

48

Figure 11 Hot Spot Analysis, Male 45-54

49

4.4 Email Respondents - Business Partnership Email
4.4.1 Female 25-34
The results of the hot spot analysis of non-normalized data for the female population age 25-34
continue to be centered in Santa Monica, the south bay area of Manhattan Beach and the San
Gabriel Valley areas of Pasadena, South Pasadena and La Canada (Figurer 12). Cold spots are
observed in North Hollywood, Paramount, Lakewood, Signal Hill and San Dimas as seen in
Figure 12.
Conversely, the normalized data displays the cold spots span from Santa Monica, Venice,
Marina del Rey into West Hollywood, and then a large cluster in the Pasadena area. The hotspots
occur in Lakewood, Carson, Huntington Park, San Dimas and Chatsworth.

50

Figure 12 Hot Spot Analysis, Female 25-34

51

4.4.2 Male 25-34
Figure 13 displays the male 25-34 male population consisting of 12,259 alumni, of which
4,337 expressed interest by opening a business partnership email. Similar to the male age 25-34
football email group, this group has no hotspots in the Pasadena and surrounding cities. Almost
all of the interest in email is located in Santa Monica, Culver City, West Hollywood and into the
Mid-City neighborhoods ending in Downtown Los Angeles. There are only 75 cold spots census
tracts, which is a lower number compared to the other age groups in this study. Cold spots are
clustered in Northridge and El Monte.

52

Figure 13 Hot Spot Analysis, Male 25-34

Because the number of alumni in this population is high, the cold spots produced from
analysis of normalized data are much larger than for other age groups in this study. The colds

53

spots stretch from Santa Monica all the way into Arcadia. This is the first grouping of
normalized data where the cold spot census tracts run uninterrupted. In addition, the cold spot
intensity is greatest in the Pasadena area, which is significant because the non-normalized data
indicated that there were no hotspots in this area. This is interpreted as meaning that although
there are a large number of male aged 25-34 alumni living in this location, very few show any
interest in alumni partner emails.
4.4.3 Female 35-44
The hot spot analysis of the non-normalized female age 35-44 population data is depicted in
Figure 14. This group contains 2,754 alumni who opened an email, it is the group with the
largest number of census tracts identified as hotspots at 492, and 347 cold spots. Large areas of
interest are created because of the increased number of those opening an email over a more
dispersed area. The hotspots in this group begin in Santa Monica and continue down to the
northern tip of the Palos Verdes peninsula. As with most other groups in this study, the hot spots
occur in Pasadena and the surrounding areas.

54

Figure 14 Hot Spot Analysis, Female 35-44

Again conversely, when analyzing the normalized data, both the higher number of alumni
in the area and those opening an email led to a centralized cold spot in Santa Monica, Beverly

55

Hills and West Hollywood. And unlike the non-normalized data the cold spots do not extend
down the coast but only show as a cluster around Hermosa Beach and the northern tip of Palos
Verdes. The cold spots are also east of Pasadena in Sierra Madre, Arcadia, Monrovia, and south
in Alhambra and San Gabriel.
4.4.4 Male 35-44
Similar to the female age 35-44 group, Figure 15 depicts hotspots beginning in Santa Monica and
moving down the coast to Palos Verdes. Hot spots are also observed in Pasadena, Sierra Madre,
and Arcadia. There are 410 census tracts that are cold spots when examining the non-normalized
data. This group shows a large contiguous area of cold spot census tracts beginning in Torrance
and moving eastward through Norwalk and into Covina. Another significant area of cold spots
occurs in North Hollywood and Northridge.

56

Figure 15 Hot Spot Analysis, Male 35-44

The normalized data shows large areas of cold spots through Santa Monica and into
downtown Los Angeles. Again, this is because the actual number of alumni living in these areas

57

is relatively high, yet when the data is normalized the ratio of alumni opening an email is
comparatively small. While there are hot spots in Torrance and Norwalk, the majority of cold
spots are centered in El Monte, Covina and West Covina.
4.4.5 Female 45-54
This group shown in Figure 16 consists of 629 alumni who opened a business partner email in a
population of 1,522 alumni. Because the population is small the census tracts with hot or cold
spots are also small in count. There are 209 census tracts that registered as a hot spot and 119
that are cold spots (Figure 16). The hotspot locations begin further up in Pacific Palisades and
move slightly east and south to Brentwood, Westwood and Santa Monica. Also in this group,
almost all of the census tracts in Rancho Palos Verdes are hotspots. Other significant hotspots
occur in Pasadena, South Pasadena, San Marino, Sierra Madre, Arcadia and most cities in the
San Gabriel Valley. There are very few clustered cold spots in this age range except for
Bellflower and Artesia, which borders the Los Angeles County line.

58

Figure 16 Hot Spot Analysis, Female 45-54

When analyzing the normalized data, the only cold spots occur Topanga Canyon,
Brentwood, Westwood and Santa Monica. There are very few hot spots in the normalized data,
located in Downey, Santa Fe Spring, Bellflower, and six tracts in the Pacoima area.

59

4.4.6 Male 45-54
There are 2,081 alumni who opened an email with alumni business partner content. While there
are some hotspots in the Topanga Canyon and Brentwood areas, these are significantly less than
the female population of the same age range. Instead, the census tracts shown in Figure 17 with
the most clustered hot spots are in Hermosa Beach, down to and including the entire Rancho
Palos Verdes peninsula. Also, there is not a large representation of alumni who opened an email
in Santa Monica and surrounding areas, which had been the case in just about every group,
regardless of age. There are only 71 cold spots in the male 45-54 year old group mostly around
the Los Feliz and Wilshire areas.

60

Figure 17 Hot Spot Analysis, Male 45-54

The normalized data contains a smaller number of hot and cold spot census tracts. Again,
the cold spots appear in Santa Monica, Venice and Pacific Palisades. There are very few clusters
of hotspot census tracts except for in Compton, Lakewood and Downey.

61

4.5 Email Respondents – USC Around Town Events Email
4.5.1 Female 25-34
There are 14,188 alumni records in this population who received an Around Town email. Of that
total, 4,195 opened an Around Town email, which are represented as hotspots in 421 census
tracts and 321 cold spot census tracts in Figure 18. Similar to other populations in this study,
alumni residing in western Los Angeles are well represented. One noticeable difference is that
while a hot spot shows alumni clustered in Santa Monica, the census tracts extend further north
and east than other female populations in this study, into Pacific Palisades, Brentwood, Bel Air
and more tracts above Santa Monica Boulevard, compared to the 25-34 female age groups who
received the football and corporate sponsor emails. This group is also significant in Pasadena,
Alhambra and San Gabriel and generally north of the I-10 freeway. The other significant areas
are Redondo and Manhattan Beach.
In contrast, the normalized data shows a cold spot running from Santa Monica all the way
to Arcadia with only a handful of census tracts not represented. The large area of cold spots is
due to the large number of alumni living in those areas yet most do not open an email. The
hotspots associated with the normalized data are clustered in Granada Hills, Lakewood, Downey,
Glendora and San Dimas.

62

Figure 18 Hot Spot Analysis, Female 25-34

63

4.5.2 Male 25-34
There are 12,259 alumni in this population who were sent an Around Town email. Of this group
in Figure 19, 3,809 opened an email about upcoming USC events and are represented in 329
census tracts that are hotspots and 102 cold spot census tracts. The hotspots for this group appear
in Santa Monica, Culver City, West Hollywood, downtown Los Angeles and in neighborhoods
surrounding the University Park Campus. While other groups are clustered in the south bay
region, only 14 census tracts registered as hot spots in the Redondo Beach area.
The cold spots are not clustered in one area but spread throughout the eastern sections of
the county, and in North Hollywood and central Los Angeles.

64

Figure 19 Hot Spot Analysis, Male 25-34

While the clustered tracts of hot and cold spots is small in number, the normalized data
shows cold spots again stretching across Los Angeles county from Santa Monica to Sierra Madre
and at the northern end of Palos Verdes. Large cold spots of normalized data suggests more

65

alumni are not opening emails. Hot spots are interspersed in central Los Angeles, Covina and the
North Hollywood area.
4.5.3 Female 35-44
Of the 7,949 alumni in this group, 2,572 opened an Around Town email is shown in Figure 20.
The clustered hot spots of non-normalized data existed in 300 census tracts and in 172 cold spot
census tracts. The hot spots are in Santa Monica, Venice, Culver City, Marina Del Rey,
Pasadena, La Canada, Hermosa and Manhattan Beach. The cold spots are spread around the
county but appear most concentrated in central Los Angeles, Northridge, Pacoima and Covina.

66

Figure 20 Hot Spot Analysis, Female 35-44

The normalized data of cold spots is centered in Beverly Hills, Brentwood, Bel Air, West
Hollywood , Santa Monica and Culver City. There is a gap starting from at Highway 101 and
reappearing in La Canada, Glendale and Pasadena.

67

4.5.4 Male 35-44
Of the 8,869 alumni in this population, 3,153 opened an email about Around Town in 301 hot
spot census tracts and 194 cold spots shown in Figure 21. The hotspots in this group are clustered
in West Los Angeles, mainly along the coast from Santa Monica to Redondo Beach. The inland
clusters are in the Pasadena and San Gabriel Valley area. Whereas the cold spots are clustered in
North Hollywood, Northridge, Downey and El Monte.

68

Figure 21 Hot Spot Analysis, Male 35-44

The normalized data does not extend down the coast as with the non-normalized data but
instead shows a gap between Hermosa Beach and El Segundo. The cold spots in this group begin
to extend into the northern top of Rancho Palos Verdes. From Santa Monica, the cold spots

69

extend eastward into Los Feliz. The cold spots also appear in eastern Pasadena, Arcadia, Sierra
Madre, San Marino, South Pasadena and Alhambra.
4.5.5 Female 45-54
There are 3,622 alumni in this population. Of this group, 1,385 opened an Around Town email in
212 census tracts identified as hotspots and 170 census tracts identified as cold spots, shown in
Figure 22. Because the population is smaller in count, the hot spots are clustered into just two
sections of the county. The first section is on the west side, in Santa Monica and down the coast
through Rancho Palos Verdes and stopping in San Pedro. The cold spot tracts are not contiguous
but begin in the north in Santa Clarita and run southward in the center of the county to Norwalk.

70

Figure 22 Hot Spot Analysis, Female 45-54

The normalized data shows just 10 tracts as hotspots and six tracts as cold spots. The
remaining tracts are identified as not significant. Review of the actual alumni locations does not

71

show areas where points are heavily clustered; again this is due to the smaller number of alumni
who exist on the email list in these areas and who actually opened the email.
4.5.6 Male 45-54
The male population for this group is 5,086 of which 1,985 opened an Around Town email.
There are 252 census tracts that identified as a hot spot and 219, which are cold spots. Similar to
the female population of the same age group, the hotspots run southward along the coast to
Rancho Palos Verdes but stop at Beverly Hills and do not reappear until Pasadena and
surrounding communities shown in Figure 23. The cold spots are spread out but are mainly
centered in Santa Clarita, downtown Los Angeles, Downey, Commerce, Walnut and Covina.

72

Figure 23 Hot Spot Analysis, Male 45-54

The normalized data shows hotspots in the southern part of the county close to the
Orange County border and in the north in Northridge and Granada Hills. This is the only group

73

that had any census tract register as either a hot or cold spot in Agoura Hills, Westlake Village
and Calabasas that were all cold spots.
4.6 Market Segmentation
Segmenting customer lists into broad profiles so a company or organization can infer a general
level of understanding of their customers at reasonable cost is common practice in marketing and
business. This practice allows a company to at least make an educated assumption about
customer behavior through the concept of segments of one, where customers with similar
attributes are assigned a classification so a company can better respond to the needs of
customers. (Dibb 2001). Market segmentation at its most basic level is about understanding what
your customers look like in a broad sense that provides insight into their current and future
purchasing habits and interests. For example, the female age 25-34 group scored high in a
lifestyle segment titled Laptops and Lattes. Knowing this, the USCAA staff can use this insight
to tailor events that don’t involve children since this segment has been identified as being mostly
childless. Another example of the benefit of market segmentation is knowing that a constituent
segment prefers to get their email on a mobile device rather than a computer. This suggests that
the marketing staff when communicating with the Laptop and Lattes group should develop
mobile friendly emails to increase readership.
The profile segmentation in this study uses the Esri Tapestry Segmentation found in Esri
Business Analyst (Esri 2012). There are 65 segments based on private and public data sources
including the U.S. Census, Esri’s Updated Demographics, InfoBase-X consumer database and a
collection of customer surveys including the Survey of the American Consumer (Esri 2012).

74

4.6.1 Market Segmentation of the 25-34 Age Range
The female and male 25-34 populations in Tables 6, 7, 8 and 9 proved to be part of the same
first three profile segments: Trendsetters, Laptops and Lattes, and Metro Renters. Trendsetters
account for 18.5%-20.2%, and in the context of this thesis are defined as alumni who are 25 and
older, median household income of $53,000, are health conscious and are attached to their
electronic devices for much of their news, shopping, and staying connected with friends and
entertainment. Laptop and Lattes account for 14% and are considered more affluent than
trendsetters with an $86,000 median household income. This group has more disposable income
for dining out, going to live shows and sporting events and has a median age of 37. Metro
Renters account for 10.2%-12.3% of the population and are approximately 32 years old with a
household income of $48,000. They tend to be single and go online for many reasons including
shopping, entertainment, travel and research.
Since the study focuses on Los Angeles County it is expected that all three of these
segments are part of the Principle Urban Centers I Urbanization Summary Group that Esri has
identified as being the most populated, affluent and in the largest metropolitan cities in the
United States. Essentially the segments tell us what can already be perceived about this group;
that they’re young, urban, educated, more likely to live in a condo or apartment and more than
half are not married. However the Preferences section of Trendsetters provides usable insight
into where they like to shop, what their hobbies are and how they prefer to consume media. This
kind of insight is helpful when developing university event programming and communication
tools that will be most likely resonate with this age group.

75

Table 8 Female 25-34, Football Email, Not Normalized
Segment Name Count Percent
Trendsetters 361 20%
Laptops and Lattes 265 14%
Metro Renters 241 13%
Urban Chic 182 10%
Top Tier 174 9%
International Marketplace 107 5%
Dorms to Diplomas 73 4%
City Lights 62 3%
Downtown Melting Pot 42 2%
Enterprising Professionals 36 2%
Las Casas 22 1%
Pacific Heights 20 1%

76

Table 9 Male 25-34, Football Email, Not Normalized

Segment Name Count
Percent
Trendsetters 322 19%
Laptops and Lattes 279 17%
Metro Renters 209 12%
Dorms to Diplomas 161
9%
International Marketplace 135
8%
Urban Chic 119
7%
Top Tier 86
5%
NeWest Residents 41
2%
Las Casas 31 1%
City Lights 30 1%
Young and Restless 24 1%
Downtown Melting Pot 23 1%
Fresh Ambitions 23 1%
College Towns 23 1%

The top three segments of the 25-34 cold spot alumni population who opened an email
were identified as being in the segments Urban Villages, Pleasantville and Las Casas. Unlike the
hotspot population, these segments lean toward a family oriented structure. Urban Villages is
made up of 61% Hispanics and 11% are Asian. Interestingly, in the Urban Villages segment
nationwide, only 20% have attended college, yet 100% of the populations in this thesis attended

77

and graduated from USC. The Las Casas segment represents a group where 84% are Hispanic
and the education level is lower than the national average. Just outside of the top three is a
segment titled Dorms to Diplomas. As the name implies this group recently graduated from
college with a bachelors or masters degree. The median household income is $24,047 and most
are employed in part-time service oriented positions.
While the cold spots represent a section of alumni who are less interested in opening
emails from the alumni association, this group can also be viewed as an opportunity to engage
through different programs and messaging. Coordination with the multicultural alumni
associations offers the possibility of generating more interest among these groups by creating
alternative events and programs.
4.6.2 Market Segmentation of the 35-44 Age Range
The female 35-44 top two spots are Urban Chic (15-18%) and Trendsetters (15-16%). However,
the female group who opened an alumni partner email had Top Tier as their third segment and
Laptops and Lattes was fourth as seen in Table 10. The Urban Chic population is more than half
married, yet less than half of those who are married have children. The U.S. median age of the
Urban Chic group is 42.7, whereas our population in this study is under 44. The median
household income is $82,524 and according to Esri are clustered in areas along the Southern
California coast. The Urban Chic alumni tend to be aware of their surroundings and what their
surroundings say about themselves. This group prefers to live in exclusive areas, visit upscale
restaurants and shopping, and they tend to be concerned about their fitness and beauty.

78

Table 10 Female 35-44, Football Email, Not Normalized

Segment Name Count Percent
Urban Chic 178 18%
Trendsetters 149 15%
Laptops and Lattes 147 15%
Top Tier 121 12%
Metro Renters 79 8%
City Lights 48 5%
Pacific Heights 33 3%
Enterprising Professionals 26 2%
Exurbanites 25 2%
Pleasantville 24 2%
International Marketplace 24 2%
Downtown Melting Pot 21 2%
Urban Villages 17 1%
Savvy Suburbanites 14 1%

The male population also had Urban Chic as the top segment for those who opened and
Around Town or corporate sponsor email. Yet the male group that opened a football email is the
first segment to be categorized as Top Tier (Table 11). The Esri Community Tapestry rating for
this group makes up less than 1% of the all U.S. households, yet alumni in the Top Tier appear in
at least some manner in all groups. In fact, the male 35-44 group who opened a football email
accounted for 16.7% of their group, and 13.3% of the Around Town group and 16.2% of the
alumni partners group. The lifestyle segmentation profile for this group is generally employed in

79

management, legal services, medicine, entertainment, and the financial sectors. The median
household income of the Top Tier group is $168,876.
Table 11 Male 35-44, Football Email, Not Normalized

Segment Name Count Percent
Top Tier 239 16%
Laptops and Lattes 234 16%
Urban Chic 233 16%
Trendsetters 220 15%
Metro Renters 135 9%
City Lights 77 5%
Pacific Heights 46 3%
Enterprising Professionals 38 2%
Pleasantville 34 2%
International Marketplace 32 2%
Exurbanites 31 2%
Downtown Melting Pot 18 1%

The female cold spot population for the corporate sponsor and football emails had
identical segments as Urban Villages (18-20.1%), Pleasantville (15.4-17.5%), and Pacific
Heights (10.3-15.7%) made up the top three spots. Those who opened an Around Town email
had City Lights (8.6%), International Marketplace (7.0%) and Pacific Heights (6.1%) as their top
three segments.

80

4.6.3 Market Segmentation of the 45-54 Age Range
In the age range seen in Table 12 all six groups have the same top two segments: Top Tier and
Urban Chic. The male Top Tier group ranges from 34.7%-39.7% while the female Top Tier
group ranges from 27%-29.7%. The third segment for four of the groups is Laptops and Lattes,
except for the females who opened a football and Around Town email and fall into the segment
Pacific Heights. As mentioned previously, Top Tier is the wealthiest group in the Esri Tapestry
Segmentation model. Again, this group accounts for less than 1% of all U.S. households but
makes up 27-39% of the USC alumni population who opened an email in this age range.

Table 12 Female 44-54, Football Email, Not Normalized
Segment Name Count Percent
Top Tier 160 29%
Urban Chic 102 18%
Pacific Heights 54 10%
Laptops and Lattes 39 7%
City Lights 39 7%
Trendsetters 29 5%
Exurbanites 25 4%
Pleasantville 25 4%
Enterprising Professionals 13 2%
Metro Renters 12 2%
Savvy Suburbanites 7 1%
International Marketplace 7 1%

81

In the cold spots the segment Trendsetters makes its first appearance in the top three
spots in three of the six groups with Pleasantville making up the remaining three. Nationally,
Trendsetters have a median age of 38, whereas this group is all at least 45 years old. The
Trendsetters inhabit the number one spot in the male and female populations who opened an
email about corporate sponsors.
The remaining Esri segmentation of alumni by age group and email content type can be
found in Appendix D on page 121.
4.7 Normalized Versus Non-Normalized Results
The results of this research provide the following answers to the original research questions:
1. Are the USC alumni who open emails in Los Angeles County significantly
clustered, and do those clusters change based on email subject matter? When both
normalized and non-normalized data are analyzed, alumni who open emails are clustered.
However, the meaning of the hot spots differs because of the effect normalization has on the
count value. The hot spots associated with normalized data identify tracts with very few
alumni but almost all open the email. These census tracts were clustered along the eastern
and southern areas of the Los Angeles County. The hot spot analysis of the non-normalized
data produced maps deemed more useful to the USCAA since they were indicative of where
alumni are concentrated along the coast and in the San Gabriel Valley, in general.
2. What segmentation lifestyle do alumni who open emails in Los Angeles County
belong to? Alumni who are located in hot spots generated by analyzing the non-normalized
data generally fell in lifestyle segmentations that contain the most affluent lifestyle profiles
(Esri 2012). Conversely, clusters of cold spots using non-normalized data placed alumni in
groups identified as working middle-class and with a greater amount of cultural and racial

82

diversity. Based on a review of the original data, the results from the non-normalized data
analysis more accurately profile the alumni population and where they open emails. And
although the non-normalized cold spots are less populated, they did identify and provide
insight into a group that the USCAA can now better target with customized messages in the
future.
4.8 Other Findings
The visualization of alumni response data displays significant clusters of alumni located on the
west side of Los Angeles and in the San Gabriel region regardless of email content or age group.
The 25-34 female population in this study extends over a broad area from Santa Monica to West
Hollywood, Mid-City, into downtown Los Angeles and ending in Pasadena. However, the male
25-34 population, even with a higher number of recipients compared to other age groups,
exhibits no significant clustering in Pasadena and surrounding cities. Even the cold spots within
this group were less defined than with the female population. Figure 24 depict the locations of
25-34 males who were sent an email and Figure 25 shows those who opened the email. The maps
show that this large group is less inclined to open emails from the alumni association. This is a
significant finding because this group represents a high number of alumni with email addresses.
It is far easier to maintain a relationship with an exiting customer than it is to try to attract new
ones as they grow older. Because this group is so large the USCAA should consider strategies
that this group may find more engaging. Recommendations for increasing email interest might
include targeted emails to the large areas of non-respondents and promoting, for example, career
services. A concerted effort should also be focused on effectively communicating with the non-
respondents. The number of non-respondents is a much larger group than the respondents and
presents an opportunity to increase email activity.

83

Emails that get sent to all alumni contain the same content, regardless of age group. The
lifestyle segmentation analysis conducted in this thesis shows that alumni are clearly at different
stages in their lives and have different expectations based on their socio and economic status. To
increase email participation, the USCAA should customize emails and could take design,
delivery, and content cues from the lifestyle segmentation profiles.

Figure 24 All Male 25-34 Football Email Recipients in Hotspots and in Pasadena

84

Figure 25 Male 25-34 Opened a Football Email in Hotspots and Pasadena

85

Another interesting find from this this study is that Long Beach was not identified as a hot or
cold spot regardless of age group or email content. This finding is important because the Long
Beach area has been the site of many alumni events and even boasts an alumni club. In addition,
the alumni population is well represented as seen in Figure 26. This means that the alumni events
and even the club might be better positioned in hot spot areas as opposed to Long Beach.

Figure 26 All Email Recipients in Long Beach

86

CHAPTER 5: DISCUSSION AND CONCLUSIONS
This first part of this chapter describes the benefits of visualizing alumni email open rates and its
contribution to future USC alumni email marketing campaigns. The second portion of this
chapter reviews the limitations of the data and analyses conducted, specifically the subject of
normalization as it applies to this study. The third section outlines future work necessary to map
and segment the complete email marketing cycle of alumni engagement: email campaign, open
rates, click-thru and finally conversion.
5.1 Study Impact
In interviews with alumni marketing staff there is very little understanding as to where alumni
reside in Los Angeles, let alone where email marketing campaigns reach the most alumni. In
fact, the general consensus amongst alumni staff is that alumni are dispersed fairly evenly
throughout the county. However, visualization from this study shows that alumni who open
emails are generally clustered near the coast and the San Gabriel Valley. Knowing where alumni
live means that alumni staff can be more strategic in their planning and communication when
developing email-marketing campaigns. Visualization of email interest allows USCAA staff to
see where alumni reside and where they interact with alumni email marketing campaigns. Prior
to this paper, the only analytics available to alumni staff was a report that displayed the open rate
for the email. This paper shows how an alumni association can use response rate data and
visualize the location of alumni interest and view changes in spatial clustering based on age
group and email content.
Location is an important variable for the alumni association and most organizations when
planning events. Business decision managers rely on data in which 75% of that data includes
some spatial element (Ozimec, Natter, and Reutterer 2010). A standard email marketing report

87

(Figure 27) with a 25% open rate, for example, may be interpreted in a way that makes it appear
that the rate applies evenly across all age groups and census tracts. However, when examining
the output from this thesis it is clear that location and the number of email opens can vary by age
group and email content. The visualization of alumni email interest can also be used for
developing events near populations that show higher rates of interest.

Figure 27 Current USCAA Email Marketing Report (Harris Connect 2014)

A byproduct of better visualization of alumni email interest through advanced targeting
means sending less emails. In some cases, USCAA alumni receive an email at least twice a
week. Being able to send emails which resonate most with the different groups is anticipated to
have a positive effect on response rates.
Other alumni associations can use the methods developed in this study to visualize their
own email response data with the goal of keeping alumni engaged. Most alumni associations are
fortunate to already have a relationship by default with alumni because the alumni attended the
university. Because of that relationship, alumni associations already have the benefit of access to

88

basic contact information that they can use as spatial inputs along with the results of their
marketing campaigns.
Alumni associations can enhance their contact data further by adding segmentation
profiles to alumni records to create tightly targeted email marketing. The profile segmentation
analysis should allow alumni association staff to be proactive by tailoring emails so that the latter
are more enticing to the user, since personalizing or customizing content has been shown to
increase click-thru rates by 62% (Asim, 2003).
5.2 Recommendations for Improving USCAA Email Marketing Lists
Using GIS to improve the email targeting lists created by alumni was the main purpose of this
thesis project. The following tables include suggestions for improving those lists. Table 13
summarizes the top three lifestyle segmentations of each group based on that group being in
either a hot or cold spot. The table can be used by USCAA staff to alter email content
accordingly. Table 14 consists of areas of opportunity based on the outcome of the hot spot and
dot density analyses. Again, USCAA staff can use this information to target underrepresented
groups and create relevant content for engaging groups that historically have not been interested
in the general messages being sent to them.

89

Table 13 Summary of Lifestyle Segmentation
Group Top Three Lifestyle
Segments
(Hot Spots)
Top Three Lifestyle
Segments
(Cold Spots)
Female 25-34 Trendsetters
Laptops and Lattes
Metro Renters
Urban Villages
Pleasantville
Las Casas
Male 25-34 Trendsetters
Laptops and Lattes
Metro Renters
Urban Villages
Pleasantville
Las Casas
Female 35-44 Urban Chic
Trendsetters
Laptops and Lattes
Urban Villages
Pleasantville
Pacific Heights
Male 35-44 Top Tier
Laptops and Lattes
Urban Chic
Pacific Heights
Pleasantville
Urban Villages
Female 45-54 Top Tier
Urban Chic
Pacific Heights
Pleasantville
Pacific Heights
Trendsetters
Male 45-54 Top Tier
Urban Chic
Laptops and Lattes
Trendsetters
International Marketplace
Urban Chic

90

Table 14 Recommendations for Improving Response Rates
Group Reason Recommendations
Target Area
Male 25-34 High alumni count,
yet very little interest
show in viewing
emails
Create email content
specific for this group
Pasadena and
surrounding cities
All Groups City has an alumni
club, yet low open
rates
Create Long Beach
specific content and
events
Long Beach
All Groups Significant number of
alumni in lifestyle
segmentations groups
identified as Hispanic
and Asian
Collaborate with
multicultural alumni
associations to create
community specific
content and
programming
Southern and eastern
sections of Los
Angeles County
Male 34-44, 45-54
and Female 45-54
High number of
individuals in these
groups are in the Top
Tier lifestyle segment
Create content and
events that appeal to
the alumni in the Top
Tier
Coastal areas of the
county and
Pasadena/La Canada
region

5.3 Limitations of Data
5.3.1 Effects of Normalizing Data
After performing analysis with normalized data, it became clear that the data was
skewing the results to show hotspots in sparsely populated census tracts and cold spots where
with high population counts. This occurred because the normalization created a ratio of a
percentage of the total. A census tract with a ratio of 1:2 registered as a hotspot while a tract with
a ratio of 1:10 registered as a cold spot. Using the count of an area and comparing it with other
areas is not always accurate because counts and areas of census tracts vary (Dailey 2006). At its
core, marketing is about communicating with as many customers as possible to inform them
about a product or service (Lee 2013). Marketing is most cost-effective when a large group of
current or potential customers can be reached while keeping operational costs low.

91

If one had a uniformly spatially distributed dataset, then normalization might be a better
option than it turned out to be in this study. This is because the populations of each census tract
would be more consistent without the extremes of 100% open rates seen in some tracts with
fewer than five alumni living there.
While conducting this study, normalizing the data had the opposite effect intended, as
clusters of hot spots of normalized data were made up of census tracts with the smallest number
of email opens. For example, Figure 28 contains normalized data of alumni open rates of email
with corporate sponsor content, and five census tracts in the center of the map identified as hot
spots with a 99% confidence rate. Yet, when the points of the total alumni population are
overlaid on the census tract it, is apparent that the population for that tract is very low, and in this
case it ranges from one to three alumni in these tracts. These tracts register as hotspots because
although only one person may have lived in the census tract, that same person also happened to
receive and open the email with corporate sponsor content resulting, in a 100% open rate for that
tract.
Alternately, there were fewer cold spots created from normalized data because cold spots
were characterized as areas with higher alumni populations and relatively higher open rates
because the population was larger. Yet the end result is that the ratio of alumni who opened an
email over the total alumni population for each census tract does not help a marketing group
when viewing the map of normalized inputs. A marketing group is interested in knowing that
there is potential in these areas, rather than focus time and resources on tracts that are supposedly
hotspots yet only one or two alumni are living in those areas.
The benefit to using normalized data in an analysis is that the difference in highly
populated census tracts and lightly populated census tracts is evened out. In this study, the results

92

from the normalized data essentially showed that the tracts with higher alumni populations have
very few alumni opening emails when compared to the entire study area. Thus visualizing the
normalized data was still useful to USCAA staff because it shows that there is opportunity to
increase readership in the cold spots by connecting with the large non-respondent population.

Figure 28 Effect of Normalization

93

5.3.2 Accuracy of Alumni Addresses
The number of records in the 25-34 populations is greater than the older populations in this study
because email became a standard form of contact during the years that this group attended USC.
The majority of alumni in the older groups did not have access to email during their time at USC
so email addresses on file had to be captured after the alumni separated from the university.
The 25-34 year old population is the only group to consistently show hotspots in the
census tracts surrounding the campus. While it’s certainly possible that many of these alumni
continue to live near campus out of convenience or for graduate school, it does suggest that a
small portion of the home addresses may not be their current addresses. Conversely, addresses in
the younger group may be associated with the address of their parents if, for whatever reason,
addresses were not updated upon graduation. In short, the successful act of receiving and
opening an email does not necessarily mean that the recipient continues to live at the street
address on file.
5.4 Future Work
This paper addresses the first step in visualizing where USC alumni open and do not open emails
and where they are significantly clustered based on email content and age. The next step is to
begin better email targeting by making profile specific emails based on the lifestyle segmentation
of the alumni population and track whether custom content results in greater email open rates.
While an open rate is an important statistic, getting alumni to an event or to patronize a
corporate sponsor is the main goal. Further work should ultimately follow the full lifecycle of an
email campaign to discover if there is a direct relationship between those who actually open
emails with those who ultimately attend an event or make a purchase from a corporate sponsor.

94

Further work will also incorporate the actual click through data already collected by
email marketing systems. The open rate data is related to a single alumni opening an email, but
does not indicate what, if anything, that person clicked on. Further segmentation on click-thru
data can be visualized and analyzed at a closer level to discover exactly what certain groups are
interested in and whether an individual click led to attendance at an event or patronization of a
corporate sponsor.
While this thesis identifies where there are high and low points in the data with regard to
alumni who open emails, it does not answer the questions as to why they open emails. Further
work should attempt to identify the motivation for opening alumni emails.
This thesis also showed that alumni who open emails from USCAA, regardless of age,
are clustered in the same areas of the county. The question of whether the clustering is due to
normal population centers or because of the socio and economic status of alumni should be
addressed in future work.

95

REFERENCES

Andreasen, Alan R. 2012. "Rethinking the Relationship between Social/Nonprofit Marketing and
Commercial Marketing." Journal of Public Policy & Marketing 31 (1): 36-41.
doi:10.1509/jppm.09.035.

Bulger, David W. 1999. "The Evolution of Relationship Marketing: Reaching an Audience of One."
Direct Marketing 61 (12): 54-59.

Clotfelter, C. T. 2003. Alumni Giving to Elite Private Colleges and Universities. Economics of Education
Review 22 (2): 109-120. doi:10.1016/S0272-7757(02)00028-6.

Coelho, Pedro S. and Jörg Henseler. 2012. "Creating Customer Loyalty through Service Customization."
European Journal of Marketing 46 (3): 331-356. doi:10.1108/03090561211202503.

Constant Contact. "The Average Open, Click-through, and Bounce Rates of Other Constant Contact
Customers by Industry." (accessed 01/25, 2015)

Dailey, George. 2006. "Normalizing Census Data using ArcMap." ArcUser 9 (1): 52.

Dibb, Sally. 2001. "New Millennium, New Segments: Moving Towards the Segment of One?" Journal of
Strategic Marketing 9 (3): 193-213. doi:10.1080/096525401301420134.

eMarketer. Email Benchmarks: Key Metrics and Trends for 2013. eMarketer,
2http://www.emarketer.com/corporate/digital-marketing-topics.aspx#/results/1265. (accessed 05 15,
2015).

Esri. 2015. Optimized Hot Spot Analysis. https://desktop.arcgis.com/en/desktop/latest/tools/spatial-
statistics-toolbox/optimized-hot-spot-analysis.htm (accessed 02 21, 2015).

———. 2015 Optimized Hot Spot Analysis (Spatial Statistics).
http://resources.arcgis.com/en/help/main/10.2/index.html#//005p00000058000000.

———. 2014 Tapestry Segmentation Reference Guide.
http://www.esri.com/library/whitepapers/pdfs/community-tapestry.pdf (accessed 04 03, 2015)

Fisher, Lauren T. "Email Marketing Benchmarks for 2014 - how do You Stack Up?" eMarketer,
http://crisptrends.com/ArquivosParaDownloads/emarketerwebinaremailmarketingbenchmarks2014-
141120135440-conversion-gate02.pdf (accessed 05 02, 2015).

German, Steven Dale. 1997. Nonprofit Relationship Marketing: The Role of Identification. Ph.D., Texas
Tech University.

Ghita, Cornel. 2014. A Decision Support System for Business Location Based on Open GIS Technology
and Data. Managing Global Transitions 12 (2): 101-120.

Greene, Henry and Shanti Greene. 2008. "Enhancing Segmentation Systems." Journal of Targeting,
Measurement & Analysis for Marketing 16 (4): 298-311. doi:10.1057/jt.2008.22.

96

Han-wen Chang, Yu-chinTai. 2010. Context-Aware Taxi Demand Hotspots Prediction. International
Journal of Business Intelligence and Data Mining 5 (1): 3.

Hasouneh, Abdel Baset I. and Marzouq Ayed Alqeed. 2010. Measuring the Effectiveness of E-Mail
Direct Marketing in Building Customer Relationship. International Journal of Marketing Studies 2
(1): 48-64.

Holt, James B., C. P. Lo, and Thomas W. Hodler. 2004. Dasymetric Estimation of Population Density and
Areal Interpolation of Census Data. Cartography and Geographic Information Science 31 (2): 103-
121. doi:10.1559/1523040041649407.

Jardine, Daniel D. 2003. "Using GIS in Alumni Giving and Institutional Advancement." New Directions
for Institutional Research 2003 (120): 77-89. doi:10.1002/ir.94.

Jarratt, Denise and Ramzi Fayed. 2012. Services Segmentation and Emergent Customer Behavior: A Case
Study. Services Marketing Quarterly 33 (1): 49-67. doi:10.1080/15332969.2012.633438.

Kiron, David. 2012. Why Detailed Data is as Important as Big Data. MIT Sloan Management Review 53
(4): 1-3.

Laroche, Michel, Mohammad Reza Habibi, and Marie-Odile Richard. 2013. To be Or Not to be in Social
Media: How Brand Loyalty is Affected by Social Media? International Journal of Information
Management 33 (1): 76-82.

Lee, Deborah1,.2013. What is Marketing? Vol. 9. doi:10.1080/15228959.2013.785900.

Lien, Che-Hui. 2005. Revisiting the Problem of Market Segmentation: A New Approach using Self-
Organizing Maps, a Data Mining Technique, in Database Marketing. Carleton University (Canada).

Los Angeles County. 2014.GIS Data Portal. 2014, http://egis3.lacounty.gov/dataportal/ (accessed 11 20,
2014).

Magliozzi, Thomas L., Paul D. Berger, and Kevin J. Clancy. 2003. The Ecological Fallacy: Some
Fundamental Research Misconceptions Corrected. Journal of Advertising Research 43 (4): 370-380.

McDearmon, J. Travis. 2013. Hail to Thee, our Alma Mater: Alumni Role Identity and the Relationship to
Institutional Support Behaviors. Research in Higher Education 54 (3): 283-302.
doi:10.1007/s11162-012-9271-6.

Newman, Melissa D. and Joseph M. Petrosko. 2011. Predictors of Alumni Association Membership.
Research in Higher Education 52 (7): 738-759.

Ozimec, Ana-Marija, Martin Natter, and Thomas Reutterer. 2010. Geographical Information Systems–
Based Marketing Decisions: Effects of Alternative Visualizations on Decision Quality. Journal of
Marketing 74 (6): 94-110.

Performance Enhancement Group. 2013. Alumni Attitude Study. Alumni Attitude Study.
http://alumnigroups.usc.edu/arc/research/PEGSurvey2013Mini-Deck.pdf (accessed 04 25, 2014).

97

Robbins, Daniel W. 2010. The National Football League (NFL) Combine: Does Normalized Data Better
Predict Performance In The NFL Draft? Journal of Strength and Conditioning Research 24 (11):
2888-99.

Rodrigue, Arlette Barbara Rodriguez. 2012.Comparison of Alumni Donors and Alumni Nondonors on
Selected Demographic, Educational and Involvement Factors.

S.M. Musyoka, S.M. Mutyauvyu, J.B.K. Kiema, F.N. Karanja, and D.N. Siriba. 2007. Market
Segmentation using Geographic Information Systems (GIS)."Marketing Intelligence & Plan 25 (6):
632-642.

Singleton, AD. 2004. "A State of the Art Review of Geodemographics and their Applicability to the
Higher Education Market."

U.S. Census Bureau. 2014. "State & County QuickFacts." U.S. Census Bureau: State and County
QuickFacts. http://quickfacts.census.gov/qfd/states/06/06037.html (accessed 12 10, 2014)

USC Alumni Association. 2015. "Football Weekenders."
http://alumnigroups.usc.edu/email/events/weekenders053914.html (accessed 05 01, 2015).

———. 2015. "USC Around Town."
http://alumnigroups.usc.edu/news/aroundtown/around_town07_15.html (accessed 05 01, 2015).

Wang, Xinhao and David P. Varady. 2005. Using Hot-Spot Analysis to Study the Clustering of Section 8
Housing Voucher Families. Housing Studies 20 (1): 29-48.

98

APPENDIX A: Email Content Examples

99

100

101

APPENDIX B: Alumni Staff Survey

102

103

APPENDIX C: Dot Density of Entire Email Populations in Each Category

Figure 29 - Dot density map of all 25-34 females who received football email

104

Figure 30 - Dot density map of all 25-34 males who received football email

105

Figure 31 - Dot density map of all 35-44 females who received football email

106

Figure 32 - Dot density map of all 35-44 males who received football email

107

Figure 33 - Dot density map of all 45-54 females who received football email

108

Figure 34 - Dot density map of all 45-54 males who received football email

109

Figure 35 - Dot density map of all 25-34 females who received corporate sponsor email

110

Figure 36 - Dot density map of all 25-34 males who received corporate sponsor email

111

Figure 37 - Dot density map of all 35-44 females who received corporate sponsor email

112

Figure 38 - Dot density map of all 35-44 males who received corporate sponsor email

113

Figure 39 - Dot density map of all 45-54 females who received corporate sponsor email

114

Figure 40 - Dot density map of all 45-54 males who received corporate sponsor email

115

Figure 41 - Dot density map of all 25-34 females who received Around Town email

116

Figure 42 - Dot density map of all 25-34 males who received Around Town email

117

Figure 43 - Dot density map of all 35-44 females who received Around Town email

118

Figure 44 - Dot density map of all 35-44 males who received Around Town email

119

Figure 45 - Dot density map of all 45-54 females who received Around Town email

120

Figure 46 - Dot density map of all 45-54 males who received Around Town email

121

APPENDIX D: Esri Tapestry Segmentation of Alumni Who Opened an Email Using Non-
normalized data

Table 15 - Segmentation of 45-54 males who received football email
Segment Name Count Percent
Top Tier 264 38%
Urban Chic 132 19%
Laptops and Lattes 69 10%
City Lights 42 6%
Pacific Heights 32 4%
Trendsetters 31 4%
Metro Renters 29 4%
Enterprising Professionals 17 2%
Pleasantville 14 2%
Downtown Melting Pot 14 2%
Exurbanites 11 1%
International Marketplace 11 1%

122

Table 16 - Segmentation of 25-34 females who received corporate sponsor email
Segment Name Count Percent
Trendsetters 445 18%
Laptops and Lattes 332 14%
Metro Renters 292 12%
Top Tier 268 11%
Urban Chic 230 9%
International Marketplace 174 7%
Dorms to Diplomas 107 4%
City Lights 76 3%
Downtown Melting Pot 72 3%
Enterprising Professionals 48 2%
Pacific Heights 40 1%
NeWest Residents 29 1%

123

Table 17 - Segmentation of 25-34 males who received corporate sponsor email
Segment Name Count Percent
Trendsetters 413 20%
Laptops and Lattes 330 16%
Metro Renters 273 13%
Dorms to Diplomas 185 9%
Urban Chic 162 7%
International Marketplace 159 7%
Top Tier 122 5%
NeWest Residents 48 2%
Las Casas 35 1%
City Lights 34 1%
Young and Restless 34 1%
Fresh Ambitions 32 1%
College Towns 32 1%
Downtown Melting Pot 28 1%

124

Table 18 - Segmentation of 35-44 females who received corporate sponsor email
Segment Name Count Percent
Urban Chic 243 15%
Trendsetters 235 15%
Top Tier 216 13%
Laptops and Lattes 192 12%
Metro Renters 113 7%
City Lights 92 5%
Pacific Heights 91 5%
International Marketplace 51 3%
Exurbanites 44 2%
Urban Villages 40 2%
Pleasantville 39 2%
Enterprising Professionals 38 2%
Downtown Melting Pot 37 2%
Las Casas 20 1%

125

Table 19 - Segmentation of 35-44 males who received corporate sponsor email
Segment Name Count Percent
Urban Chic 274 16%
Top Tier 264 16%
Laptops and Lattes 255 15%
Trendsetters 240 14%
Metro Renters 151 9%
City Lights 93 5%
Pacific Heights 68 4%
Enterprising Professionals 45 2%
Exurbanites 40 2%
Pleasantville 38 2%
International Marketplace 27 1%
Downtown Melting Pot 26 1%

126

Table 20 - Segmentation of 45-54 females who received corporate sponsor email
Segment Name Count Percent
Top Tier 184 29%
Urban Chic 130 20%
Laptops and Lattes 56 8%
Pacific Heights 53 8%
City Lights 39 6%
Exurbanites 37 5%
Trendsetters 30 4%
Pleasantville 26 4%
Metro Renters 17 2%
Enterprising professionals 16 2%
Savvy Suburbanites 9 1%

127

Table 21 - Segmentation of 45-54 males who received corporate sponsor email
Segment Name Count Percent
Top Tier 289 39%
Urban Chic 133 18%
Laptops and Lattes 65 8%
City Lights 39 5%
Metro Renters 33 4%
Pacific Heights 31 4%
Trendsetters 31 4%
Enterprising Professionals 20 2%
Pleasantville 15 2%
Downtown Melting Pot 15 2%
Exurbanites 15 2%
Savvy Suburbanites 9 1%
International Marketplace 8 1%

128

Table 22 - Segmentation of 25-34 females who received Around Town email
Segment Name Count Percent
Trendsetters 358 17%
Laptops and Lattes 285 14%
Metro Renters 242 11%
Top Tier 236 11%
Urban Chic 198 9%
International Marketplace 140 6%
Dorms to Diplomas 97 4%
Downtown Melting Pot 76 3%
City Lights 65 3%
Pacific Heights 55 2%
Enterprising Professionals 34 1%
NeWest Residents 25 1%
Las Casas 23 1%

129

Table 23 - Segmentation of 25-34 males who received Around Town email
Segment Name Count Percent
Trendsetters 350 20%
Laptops and Lattes 276 16%
Metro Renters 235 13%
Dorms to Diplomas 153 9%
International Marketplace 141 8%
Urban Chic 114 6%
Top Tier 98 5%
NeWest Residents 35 2%
College Towns 31 1%
Las Casas 30 1%
Fresh Ambitions 27 1%
City Lights 26 1%
Young and Restless 25 1%
Downtown Melting Pot 23 1%

130

Table 24 - Segmentation of 35-44 females who received Around Town email
Segment Name Count Percent
Urban Chic 196 18%
Trendsetters 177 16%
Laptops and Lattes 170 15%
Top Tier 157 14%
Metro Renters 88 8%
City Lights 50 4%
Pacific Heights 41 3%
Enterprising Professionals 31 2%
International Marketplace 26 2%
Exurbanites 25 2%
Pleasantville 24 2%
Downtown Melting Pot 19 1%
Urban Villages 18 1%
Savvy Suburbanites 13 1%

131

Table 25 - Segmentation of 35-44 males who received Around Town email
Segment Name Count Percent
Urban Chic 249 19%
Laptops and Lattes 223 18%
Trendsetters 185 14%
Top Tier 169 13%
Metro Renters 133 10%
City Lights 68 5%
Pacific Heights 37 2%
Enterprising Professionals 35 2%
Downtown Melting Pot 30 2%
Pleasantville 27 2%
International Marketplace 25 1%

132

Table 26 - Segmentation of 45-54 females who received Around Town email
Segment Name Count Percent
Top Tier 159 27%
Urban Chic 115 19%
Pacific Heights 71 12%
Laptops and Lattes 50 8%
City Lights 34 5%
Exurbanites 31 5%
Pleasantville 30 5%
Trendsetters 25 4%
Enterprising Professionals 14 2%
Metro Renters 12 2%
Savvy Suburbanites 9 1%
Urban Villages 9 1%

133

Table 27 - Segmentation of 45-54 males who received Around Town email
Segment Name Count Percent
Top Tier 312 34%
Urban Chic 169 18%
Laptops and Lattes 90 10%
Pacific Heights 69 7%
Trendsetters 54 6%
Exurbanites 39 4%
City Lights 35 3%
Metro Renters 34 3%
Pleasantville 28 3%
Enterprising Professionals 19 2%
Downtown Melting Pot 12 1%

Abstract (if available)

Abstract Permission-based email marketing is an inexpensive and effective method for companies to develop an ongoing conversation with customers and to acquire new ones. Almost 80% of all adult U.S. consumers use email each day. While the process of sending emails to customers is relatively simple, interpreting email response rates to develop relevant, customized email content is challenging. Companies who are able to offer a personalized email experience to their customers report improved response rates and engaged customers. Email response rate metrics in their current form tend to be one-dimensional and do not provide the location of who opens an email or to what socio-economic group they belong. Existing studies focus on email response rate data but do not include visualization of email recipients nor do they assign them to a lifestyle segmentation profile. The purpose of this thesis project was to utilize hot-spot and dot density analysis by census tract to identify spatial clusters of USC alumni email interaction in Los Angeles County, and to conduct market segmentation of alumni who open and do not open emails. The outcome of this thesis is to provide maps that allow marketing and event managers at the USC Alumni Association to visualize alumni email recipients that respond to specific categories of email in order to improve email marketing campaigns and to better position resources for event planning.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Integrating spatial visualization to improve public health understanding and communication

PDF

Developing, maintaining, and employing crowd-sourced geospatial data in support of helicopter landing zone surveys for disaster response operations

PDF

Cartography for visualizing anthropogenic threats: a semiotic approach to communicating threat information in 3-D spatial models

PDF

Cartographic approaches to the visual exploration of violent crime patterns in space and time: a user performance based comparison of methods

PDF

Social media canvassing using Twitter and Web GIS to aid in solving crime

PDF

GIS-aided production of equipment locator maps for Metro rail maintenance and support

PDF

Semi-automated visualization of spatial information in unstructured text

PDF

GIS data curation and Web map application for La Brea Tar Pits fossil occurrences in Los Angeles, California

PDF

Wake County District Overlay: an online electoral data visualization application

PDF

The spatial effect of AB 109 (Public Safety Realignment) on crime rates in San Diego County

PDF

Visualizing historic space through the integration of geographic information science in secondary school curriculums: a comparison of static versus dynamic methods

PDF

San Diego Wildfire Hazards Information Center mashup

PDF

Guiding business oriented volunteered geographic information through geotigger services: a case study of CrossFit affiliates

PDF

A user study of GIS infused genealogy with dynamic thematic representation and spatiotemporal control

PDF

A geospatial analysis of income level, food deserts and urban agriculture hot spots

PDF

Peoples of Washington historical geographic information system: geocoding culture using archival standards

PDF

Precipitation triggered landslide risk assessment and relative risk modeling using cached and real-time data

PDF

Development of a mobile GIS high-water mark data collection application for the Mississippi River Basin

PDF

A spatial analysis of violent crime cold spots: testing the capable guardian component of the routine activity theory

PDF

Creating a flood vulnerability index for Houston, Texas

Asset Metadata

Creator Gonzalez, Jason (author)

Core Title Visualizing email response data to improve marketing campaigns

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 10/21/2017

Defense Date 09/01/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag alumni,email,hot spot analysis,market segmentation,OAI-PMH Harvest,response rates

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Swift, Jennifer (committee chair), Chiang, Yao-Yi (committee member), Oda, Katsuhiko (committee member)

Creator Email jasongon@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-192290

Unique identifier UC11278524

Identifier etd-GonzalezJa-3987.pdf (filename),usctheses-c40-192290 (legacy record id)

Legacy Identifier etd-GonzalezJa-3987.pdf

Dmrecord 192290

Document Type Thesis

Format application/pdf (imt)

Rights Gonzalez, Jason

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA