• Module 2 - Sources of Data

Table of Contents
1. Potential Data Sources
Categories of Data Sources
Uses of Databases in Predictive Analytics
More on Purchased Data

2. B2C Use Cases

Case 1: Rouge Motor Company
Case 2: See the World (STW)
Case 3: Giant Food Variety (GFV)

3. B2B Use Cases

Case 4: Data R Us
Case 5: White Hat Marketing

4. Determining the Business Question

Types of Consumer Data Used in Direct Marketing
Case 1: Rouge Motor Company
Case 2: See the World (STW)
Case 3: Giant Food Variety (GFV)
Case 4: Data R Us
Case 5: White Hat Marketing

Introduction
In order to successfully market using predictive analytics, it is crucial to obtain appropriate
data sources as inputs.
In this module, students will get an introduction to data used in targeted marketing
campaigns. Predicting consumer response, return on investment from campaigns,
and appropriate product or service offerings can only be done
with solid input data because when you put “garbage in,” you get “garbage out.”
Additionally, students will learn the importance of understanding the variety of data
sources available along with each source's intent, capabilities, relevance to your business,
cost, and stability—all key parts of deciding whether to make an investment in specific
external data sources.

Learning Objectives
• Explain to leadership the types of data sources available in the marketplace
• Effectively evaluate various data sources (internal vs. external) used in prospecting,
consumer marketing, customer segmentation, and lead qualification
• Frame up a data question to clearly articulate revenue goals and measure them for
success

1. Potential Data Sources

In any customer, client, or prospecting effort, capturing and organizing the best data
sources available for the goals of the effort are crucial. A predictive model can only do as
well as the inputs provided.
During this course, we are using five case studies as examples of industries and companies
that leverage data to optimize their value to customers or clients to ensure better
engagement and retention. In this module, we will investigate available data sources within
the context of each case study. These data sources will then be used to
build either a customer marketing database or a prospecting database.

Categories of Data Sources

There are a variety of data sources available to any business. There are four categories
which we will be using to assess the five case studies that appear throughout the
module. The categories are:

Internal data
These are data which are available in internal systems such as customer transactional data
(purchases), and customer contact data (email, phone).

Customer-provided data
These are data provided by the customer voluntarily, which could be at the point of sale,
online, by telephone, by mail, or by a survey.

Public data
Some data are publicly available via online sources, trade directories and resources such as
census bureaus or other government-related records.

Purchased data
There are many compilers of data who provide information for sale ranging from
demographics, to interests, to purchases by consumers.
Uses of Databases in Predictive Analytics
Customer Groups
Identify customer groups, from the most loyal, high-value customers to first-time
customers and occasional purchasers.

Customer Segments
Create detailed customer segments based on demographics, behaviors, or even
personal interests.

Lead Generation
Effective lead generation for better conversion rates.

Messages
Design personalized messages for both current and prospective customers.

Channel
Determine the best channel to reach customers or prospects.

Marketing Efficiency
Improve marketing efficiency by sending campaigns to those likely to respond.

Loyalty Programs
Build effective loyalty programs that provide the right incentives for repeat purchases.

The following video provides an overview of customer marketing databases and their importance in
organizing customer data.

More on Purchased Data

Purchased data is used quite widely by successful businesses. They realize that
supplementing their own internal data with data compiled by companies who specialize in
that arena can provide for better prospecting, better customer relationships and targeting,
and better lead generation. Below are some important considerations to keep in mind for
purchased data.

Start
Finding a data source
There are also a number of demographic compliers for B2C marketing. These data are often
used for overlay on consumer databases to append information such as age, income,
housing status, household composition, interests, and purchasing power. Demographic
providers include Experian, Acxiom, Epsilon, and many others.

Linking to database
There are also several providers of identity graphs. These providers match the online ID of
a customer or visitor to the company’s website to the offline ID. The first major identity
graph provider is LiveRamp, however there is a host of companies currently jumping into
the space. These data are used to get a sense for characteristics on who’s visiting one’s site.

Operationalize
The marketer will have a service provider match their customer database to these
providers to gain additional information on their customers.

Activate to reach customers

Most serious marketers have a connection with Facebook, Google, or Amazon Advertising
to enable additional data for audience along with a means for connecting with customers
and prospects.

Account-based marketing
For B2B marketing, ABM without the right contact info is guesswork. This makes it
important to find a data compiler that can be used for a prospecting database. (Gartner
2015)

End

Activity
Which of the following are overall uses of databases
in predictive analytics?
Select all that apply, then click “Submit.”
Create detailed customer segments based on demographics, behaviors, or even personal
interests
Decrease marketing efficiency by sending campaigns to those not likely to respond
Build effective loyalty programs that provide the right incentives for repeat purchases

2. B2C Use Cases

In the previous section, we looked at the four categories of data sources available to
businesses: internal, customer-provided, public, and purchased. In this section, we will use
the four categories to assess our B2C case studies: Rouge Motor Company, See the World,
and Giant Food Variety.

Case 1: Rouge Motor Company

Scenario Refresher: An automotive manufacturer, Rouge Motor Company, has global

reach. Rouge has 4 separate product lines intended for consumer purchase (Basic Sedan,
Luxury Sedan, SUV, and Pickup Truck).

Internal data
• Rouge may have records of past consumer purchases from customers. These data
would include a customer identifier along with name, address, email, automobile
purchased with the price and category, and means of financing.
• Rouge may have records from dealers on auto servicing by customer.
• Rouge may also have records of consumer interactions from prospective customers,
or prospects. These data would include a prospect name, address, which
automobiles they considered with the price and category.

Customer-provided data
• Rouge may have captured data from consumer satisfaction surveys indicating
various points of information about the customer’s decision reasons.
• Rouge may also have a website where customers can share their experiences. For
expensive purchases such as an automobile, customers are more likely to share
their views with others.

Public data
 • Most countries, states, and provinces collect data on residents. This would include
household composition, some measure of home ownership, and attributes of
socioeconomic status. These measures can be important in correlating to vehicle
purchase.
• There are many data sources that provide availability of public transportation.
These can be important in determining whether a vehicle is needed.
• Many countries, states, and provinces keep data on consumer vehicle ownership.
These may not be available at a household level, but can be available in small-area
geographies.

Purchased data
• There are many compilers of data who provide information on consumer purchases
and searches throughout an industry category, such as automobiles. The availability
varies by country, state, and province. In the U.S., the primary provider is Polk, a
division of IHS Markit. Polk’s Automotive Data helps dealers, media and agencies
understand who’s in-market, where to find them, what they are looking for, when
they will buy and what will drive their purchase decision.
• In addition, providers of data can develop strategic agreements. The Nielsen
company is a global company that measures advertising effectiveness.
They’ve partnered with Polk Automotive to deliver campaign effectiveness.
(Nielsen, 2021)

• RECOMMENDED READING
• For more information on what companies can do in database marketing, CRM, one-
to-one marketing, and advanced analytics to stay competitive, read Legacy
Companies Need to Become More Data Driven—Fast.

Case 2: See the World (STW)

Scenario Refresher: A cruise line, See the World (STW), has global reach. STW travels to
several different parts of the world (Mediterranean, Caribbean, Australia/New Zealand,
and Alaska).

Internal data
• STW may have records of past consumer purchases from customers. These data
would include a customer identifier along with name, address, email, cruise
purchased along with the price, destination, and time of year.
 • STW may also have records of consumer interactions from prospective customers,
or prospects. These data would include a prospect name, address, which cruises
they considered along with the price, destination and time of year.

Customer-provided data
• STW may have captured data from consumer satisfaction surveys indicating various
points of information about the customer’s decision reasons.
• STW may also have a website where customers can share their experiences. For
expensive experiences such as a cruise, customers are more likely to share their
views with others.
• Cruise line Instagram accounts can show potential travelers the views they’ll
experience on their vacations, with the marketer capturing the views.

Public data
• Most countries, states, and provinces collect data on residents. This would include
household composition, some measure of home ownership, and various attributes
of socioeconomic status. All these measures can be important in correlating to
cruise purchase.
• There are many data sources that provide availability of access to major airports or
departure points for cruises. These data give a sense for ease of getting to a cruise
departure.
• Many countries, states, and provinces keep data on travel outside the country. These
may not be available at a household level, but can be available in small-area
geographies.

Purchased data
• Many cruise lines partner with travel agents or partner brands to build robust
mailing lists that are made available for purchase. These are an efficient source of
previous cruise-goers and new travelers.
• Companies such as Deloitte have done studies on the cruise line industry along with
the passenger experience. (Deloitte, 2021)

Case 3: Giant Food Variety (GFV)

Scenario refresher: A grocery retailer, Giant Food Variety (GFV), has regional outlets within a mix of
urban and rural areas in a portion of the United States. GFV has 90 separate products called
“consumer packaged goods”, each of which consists of various items. Consumer packaged goods are
pre-packaged items sold to individual consumers (Examples: laundry detergent, carbonated
beverages, and frozen pizza).
Internal data
• Most grocery retailers offer programs to shoppers called a “loyalty program”. These
programs provide customers with discounts based on various factors. For all
shoppers that opt-in, the retailer will have a complete purchase history of all
transactions including products purchased, price, frequency of purchase, total
amount spent, and total number of times making a purchase

Customer-provided data
• Some retailers have websites where people can post, leave feedback, and share their
thoughts. GFV may be able to leverage to optimize product mix.
• GFV may offer coupons or other incentives in exchange for customer demographics.

Public data
• Most countries, states, and provinces collect data on residents. This would include
household composition, some measure of home ownership, and various attributes
of socioeconomic status. All these measures can be important in determining the
volume of food a household will buy along with purchasing power.
• There are many data sources that provide availability of public transportation.
These can be important in determining whether a person would take a big shopping
trip infrequently, or small shopping trips more frequently. It would also help enable
to mix of perishable items vs packaged items.

Purchased data
• There are also compilers of panel data that capture receipts for consumers that opt-
in. These data are scanned by consumers and transmitted. An example is
Numerator. These data can give a sense for share of wallet from a retailer’s
customers. (Numerator, 2021)

RECOMMENDED VIDEO
• To learn what happens when a marketer doesn't "do it right", watch the following
short video by Merkle.

Your Summer Dreams Begin (youtube.com)

RECOMMENDED READING
For more information on the importance of making decisions connected to revenue via the
effective use of data and analytics, read Is Your Marketing Strategy Based on the Right
Data?
 Activity
True or False: Many countries, states, and provinces
keep data on travel outside the country.
Choose the best answer, then click "Submit".
True
False

3. B2B Use Cases

In this section, we will use the four categories of data to assess our B2B case studies: Data R Us and
White Hat Marketing.

Case 4: Data R Us
Scenario: Suppliers of goods to retailers know what they’ve shipped to each retailer,
but they don’t know their market share nor the specific drivers impacting change in market
share. A marketing information supplier, Data R Us, has agreements with 800 different CPG
retailers throughout the U.S. Data R Us aggregates data for suppliers such as Coca-Cola,
Procter & Gamble, and Kraft-Heinz along with model-based explanations of why market
share has changed over time. Their service is sold via annual subscription.

Internal data
• Point-of-sale data for 800 retailers throughout the U.S.

Customer-provided data
• Some customers provide their shipment data; how much product has been shipped
to each retailer by time period and product type.

Public data
• Quarterly revenue by retailer.
Purchased data
• Bombora is a company that measures prospects’ digital journey across 5,000+
premium B2B websites, so you can know exactly what prospects are in the market
for, letting you guide them to your solution. (Bombora, 2024)

Case 5: White Hat Marketing

Scenario: Over the last decade, consumers have spent increasingly more time online, and marketers
have responded with digital advertising to drive website visits. Now, the best leads for
any business are their website visitors. Web visits are a strong data signal often paired with weak
converting digital display advertising. White Hat Marketing takes this strong signal and pairs it with
direct mail—a high-converting ad medium—to enable a higher response rates and average order
values for client customers.

Internal data
• History of conversion rates by industry based on actions on website; identity graph
matching device ID to physical mailing address.

Customer-provided data
• Clients provide their first-party data from website visits including time on the
domain, which URLs were visited, number of clicks, source of entry to site (direct vs.
search).

Public data
• Total revenue for each prospective publicly traded client.

Purchased data
• Firmographic segmentation to group companies by filters; Coresignal is a good
example. (Coresignal, 2024)

RECOMMENDED READING
• Read about the 3 Reasons to Build an Effective B2B Prospecting Database to learn
how to construct an effective database.
 RECOMMENDED VIDEO
• In this video, Colin Carrasquillo from Experimarketing provides insights into a
major provider of automotive ownership data. Watching this video will help you to
understand the importance of utilizing data to inform automotive advertising
decisions

Jira Exit Interview | 1.2 (youtube.com)

4. Determining the Business Question

Before spending the time, effort, and money on acquiring, organizing, and ingesting data, it
is crucial to understand how to evaluate the data sources. There are some key questions to
ask about the “why” behind the data. It is also important to look beyond the immediate
need and think about broader applications into the future.
Once sources of data are in place, most organizations build entire intelligence systems
around them. Taking the time up front to ensure that a complete evaluation is done can set
the tone for marketing for years into the future. Evaluation needs to be done across three
dimensions.

Accuracy

First, it's important to evaluate the quality of the data. Is it accurate?

Coverage
Next, it's important to consider coverage. Does it have high coverage? That is, does it capture a high
percentage of the target audience?

Predictive Power
Finally, it's important to assess the data's usefulness in predicting consumer behavior. Can it be used
to segment the audience appropriately and predict how much clients or customers will spend?

Types of Consumer Data Used in Direct Marketing

Types of consumer marketing data used in direct marketing involve demographic,
psychographic, and behavioral information. This data helps to identify target audiences for
marketing campaigns, forming the foundation for a data-driven marketing strategy.
Demographic Data and Its Impact
Accurate information about consumers is created by demographic data, including census data,
home ownership, and potentially thousands of external sources. This data is crucial as a
foundation for a dataset before more advanced tactics can begin.
Essentially, before you attempt to understand how a consumer will behave or how they will feel
about something, you first need to understand the basics of who they are. This can be as simple
as basic attributes like marital and family status, age, gender, and income, but can expand to
include hundreds of variables. Some examples include how long they have lived in their home,
who else shares their home, and whether they have pets. (Meeks 2024)

Behavioral Data Explains What a Consumer Does

Now that a solid foundation of demographic data is in place, you can connect more temporary
actions or behaviors to a persistent consumer record. Behavioral data is really about what is
happening now or has happened in the recent past. What types of purchases is a consumer
making? Do they engage with any specific activities or display any interests, such as gaming?
This helps marketers paint an even more detailed picture of who their important prospects are.
(Meeks 2024)

Psychographic Data to Understand Consumer Motivations

Understanding consumer motivations is crucial for marketing success. Psychographic data
includes a consumer’s values, opinions, and motivations. Essentially, while behavioral data
may tell us what a person is doing, psychographic data helps uncover why they are making
the choices they do. This data allows marketers to look forward, uncovering predictive
insights. (Meeks 2024)

RECOMMENDED VIDEO
While customer demographics provide you with an idea of what products and services
your customers may be interested in, psychographics include things like personality,
values, attitudes, interests, and lifestyles. In this video, the team from Elegant Themes
explains 3 key steps for using psychographics to improve marketing efforts. Watching this
video will help you to better understand the use of motivations and desires in marketing
success.
Using Psychographics to Improve Marketing Techniques (youtube.com)

Case 1: Rouge Motor Company

Scenario: Rouge Motor Company wants to engage in a marketing campaign for immediate
sale; it could be the only sale to a consumer in a 5-year period. As a CRO, it is your
responsibility to determine the question that you want your analytics team to answer. As
we discussed in the previous video, it is essential to ask the right question. As a CRO, you
will need to ensure the data exists; the analyst is responsible for answering the question.
In the carousel below, we'll explore three potential questions for Rouge Motor Company
and consider what sources of data are required to answer these questions.

Determine the question we want to answer

• Maximize overall sales of inventory on the lot?
• Optimize offer based on longer-term value, even if inventory’s not ready?
• Maximize profit vs. overall sales?
Next, it's important to determine what data would help answer each question.

Maximize overall sales of the inventory on the lot?

To answer this question, we could do the following:
• Collect which vehicles are available and in which quantity.
• Find sources of data showing either prior purchase of those vehicles or similar vehicles.
• Capture travel distance to the nearest dealership for each prospect that meets the above
profile by matching prior purchase vs. availability at the specific dealership.

Optimize the offer based on long-term value?

To answer this question, we could do the following:
• Use survey data to determine interests and wish lists, called psychographics
• Capture demographics on switching to a new vehicle type vs. buying one similar to a
previous purchase
• Find an identity provider to determine who browsed which vehicle configuration on the
website

Maximize profit vs. overall sales?

To answer this question, we could do the following:
• Capture the profit margin for each vehicle
• Determine segments where the population is most skewed toward profitable vehicles
• Capture the geographic penetration of those segments and overlay them with demographics

Case 2: See the World (STW)

Scenario: STW wants to do a marketing campaign for the upcoming season. They also want
to set the right impression for the following seasons. In the carousel below, we'll explore
three potential questions for achieving these outcomes.
Determine the question we want to answer
• Prospect new potential customers to get them in the door?
• Upsell previous customers on a more expensive option?
• Ensure that the ships are as close to full as possible?
Next, it's important to determine what data would help answer each question.

Prospect new potential customers to get them in the door?

To answer this question, we could do the following:
• Find an identity provider to determine who browsed the website
• Overlay existing customers with demographic data such as income, net worth, age, and
household composition and then target areas with similar demographic characteristics

Upsell previous customers on a more expensive option?

We could do the following:
• Use survey data to determine interests and wish lists (psychographics)
• Look at purchase history for potential trends and predictors
• Overlay existing customers with demographic data such as income and net worth and target
those with higher values

Ensure that the ships are as close to full as possible?

We could do the following:
• Focus on ship- or destination-specific data
• Capture geography of residence for customers and prospects
• Capture previous years’ capacity

Case 3: Giant Food Variety (GFV)

Scenario: GFV’s best customers shop weekly and buy a large cross-section of their 90
available products; as a result, the company wants to market for continual interaction.
In the carousel below, we will explore three potential questions that could identify where
to best focus their efforts.

Determine the question we want to answer

Should they try to:
• Maximize sales from existing customers?
• Maximize the number of products purchased by existing customers?
• Lure new customers from competing retailers?
Next, it's important to determine what data would help answer each question.
Maximize sales from existing customers?
To answer this question, we could do the following:
• Capture complete purchase history from the previous 12 months, accounting for seasonality
• Focus on data such as purchase amount and frequency, currency amount of purchase, and
how recently the last purchase was made
• Capture data on the products bought on promotion
• Capture the distance from home to store relative to product mix

Maximize the number of products purchased by existing

customers?
We could do the following:
• Capture complete purchase history from the previous 12 months, accounting for seasonality
• Focus on data such as the units purchased and the overall shopping basket size relative to
the number of items purchased
• Capture data on the units bought on promotion
• Capture the distance from home to store relative to the number of items purchased

Lure new customers from competing retailers?

We could do the following:
• Purchase data that provides categorized market share estimates across various local
retailers
• Capture the trade area demographics for GFV stores vs. competitors
• Purchase data on retail promotions by competitors

Case 4: Data R Us
Scenario: Data R Us wants to upsell their clients, CPG suppliers, from their basic offering
that solely provides data on market share to the premium package that uses predictive
modeling to identify the primary drivers impacting change in market share.
In the carousel below, we'll explore two potential questions that could help achieve this
goal. Which one would you prioritize?

Determine the question we want to answer

• How can we identify the primary drivers impacting market share?
• What value would each client put on this information?
Next, it's important to determine what data would help answer each question.
Identify the primary drivers impacting market share?
To answer this question, we could do the following:
• Establish consistency across the data provided by each retailer; for instance, does frozen
pizza have the same definition in the data from retailer X as from retailer Y?
• Ensure that current price data is provided by each retailer in a consistent, timely fashion.
• Obtain data on promotional efforts for our target client and their competitors.

Determine what value each client would put on this information?

To answer this question, we could do the following:
• Obtain data on return on ad spend (ROAS)
• Calculate the potential improvement in ROAS based on information from all supplier data
leveraging increased revenue to clients who have upgraded to the premium product
• Capture demographics consistent with retailer trade areas

Case 5: White Hat Marketing

Scenario: White Hat Marketing wants to prospect for new customers that would find
benefit in retargeting via direct mail.
In the carousel below, we'll explore two potential questions, one leading to the other, that
might help address this goal and the kinds of data sources required.

Determine the question we want to answer

• How can we identify potential clients that would benefit most from our retargeting offering?
• How can we get to the right decision-maker for target clients?
Next, it's important to determine what data would help answer each question.

Identify potential clients that would benefit most from our

retargeting offering?
To answer this question, we could do the following:
• Look for sources of data that provides companies operating in our most profitable sectors.
• Find out which industries benefit most from catalogs or full-color brochures.
• Obtain data on which companies have the most time spent by web visitors on a company’s
website researching potential products.
Determine how to get to the right decision maker for target
clients?
To answer this question, we could do the following:
• Obtain data from recent conferences or shows
• Secure data from companies that track who the executives are for marketing across various
potential clients
• Network

REMEMBER
In each case you must evaluate each data source within the dimensions of accuracy, high
coverage, and predictive power toward the question being asked.

The effective marketer or salesperson will need to work closely with data architects to
build a data model that will house the data in such a way to allow marketers and
salespeople to meet all the goals of the customer marketing or prospecting database. Data
modeling is a process used to define and analyze data requirements needed to support the
business processes within the scope of corresponding information systems in
organizations. (IBM, 2024)
Above all, the purpose of collecting data is not only to increase sales to customers, but to
establish relevancy so that the customer feels good about the purchase and comes back.
After each effort, additional data needs to be added and sources re-evaluated.
Finally, there can always be errors or missing data within any data sources. The smart
business leader uses domain expertise to determine rules for whether data are “good” or
need to be thrown out.

RECOMMENDED READING
For more information on ways to determine the quality of your data, read How to Measure
Data Quality – 7 Metrics to Assess the Quality of Your Data.

Glossary
For your convenience, here is a list of the key terms presented in this module.
Behavioral Data
Behavioral data is based on what is happening now, or in the recent past. This data tells us
what a person is doing.

Customer-Provided Data
These are data provided by the customer voluntarily, which could be at the point of sale,
online, by telephone, by mail, or by a survey.

Data Model
Data models are visual representations of an enterprise's data elements and the
connections between them.

Data Modeling
Data modeling is a process used to define and analyze data requirements needed to
support the business processes within the scope of corresponding information systems in
organizations.

Demographic Data
This type of data includes census data, home ownership, and potentially thousands of
external sources.

Internal Data
These are data available in internal systems such as customer transactional data
(purchases) and customer contact data (email, phone number).

Psychographic Data
Psychographic data includes data based on a consumer’s values, opinions, and motivations.
This data helps uncover why they are making the choices they do.

Public Data
Some data are publicly available via online sources, trade directories, and resources such
as census bureaus or other government-related records.

Purchased Data
These are data that companies compile in order to sell the related information, which
can range from demographics to consumer interests and purchases.

Knowledge Check
Module 2
The following series of questions will test your understanding of the contents of
this module.
Please read each question carefully before selecting your answer.