lecture-03-research-design

Research Design

FALL 2026 · LECTURE 03

M303

Research Design

Choosing exploratory, descriptive, and causal approaches under uncertainty.

Where we are

FROM PROBLEM FORMULATION TO RESEARCH DESIGNToday we choose the research approach that fits the decision and the level of uncertainty.

01

Use your name tag

Help everyone learn names and build a more interactive classroom.

02

Keep phones, tablets, and laptops away

Unless they are requested for a class activity.

Learning Objectives

01Distinguish the three approaches

Explain the purpose and main characteristics of exploratory, descriptive, and causal research.

02Use uncertainty as a guide

Explain how uncertainty shapes the choice among the three research approaches.

Prior-class review

P · problemAMP · actionable management problemRO · research objectiveRQ · research questionH · hypothesis

Classify all 10 statements:1 P · 1 AMP · 3 RO · 3 RQ · 2 H

RQ

Which TV product attributes increase consumers’ quality perceptions the most?

H

Consumers care about screen quality the most.

P

Sales of Vizio Model XYZ 4K TV are low because consumers perceive its quality as inferior.

RO

Determine consumers’ additional willingness to pay for each upgraded product attribute.

AMP

Which attributes should be improved—screen, sound, color clarity, refresh rate, processor, or price?

RO

Determine which attributes consumers care about the most.

RQ

What is the cost of increasing quality for each attribute?

H

Consumers are willing to pay up to $100 more for a better refresh rate.

RO

Determine the cost of upgrading the attributes.

RQ

What additional willingness to pay do consumers have for each upgraded attribute?

Prior-class review — decision tree

COSGUNER RESEARCH INC.

Perfect information about the direct-mail conversion rate

The firm can determine whether the rate is 3% or 1% before the mailing decision.

Research fee · $100,000

YOUR DECISION

ANo direct mailing

BDirect mailing without research

CBuy research first

Prior-class review — decision-tree solution

Marketing
research?

YES → observe conversion rate

NO → decide now

3%60% probability

1%40% probability

Direct mail?

SendReturn · $1.0M

Do not sendReturn · $0.5M

Expected return with research
0.6($1.0M) + 0.4($0.5M) − $0.1M = $0.7M

Course access and analysis tools

CANVAS

Course files in one place

Lecture materials, instructions, announcements, and required data files will be posted on Canvas.

WEB TOOLS

Instructor-designed analysis tools

Hypothesis testing · Regression · Conjoint analysis

koraycosguner.github.io/tools.html

Examples and demonstrations will be presented in class.

This class

Problem
formulation

Preliminary stages of marketing research

Research
design

Choose an appropriate research approach

Data
collection

Secondary data
Primary data
Sampling

Data analysis &
recommendations

Preparation · Descriptive statistics and cross-tabulations · Hypothesis testing · Correlation and regression

How do we design and implement marketing research?

01

Choose the research approach

Exploratory, descriptive, causal—or a combination.

02

Choose data sources and methods

Secondary or standardized data; interviews, focus groups, surveys, or experiments.

03

Plan measurement and collection

Design instruments with the analysis in mind, choose a sample, collect the data, and code it.

04

Analyze and recommend

Interpret the results and translate them into managerial recommendations.

A research design is a detailed blueprint for conducting a research study. The elements of the design must fit together. It is not possible to conduct a research study without a design: the design may be preliminary, tentative, and subject to change, but some plan of action is necessary.

The design process involves many interrelated decisions. The most important is the choice of research approach because it determines how the information will be obtained. Typical questions include: Should we rely on secondary sources such as the Census? Is exploratory research with group discussions more appropriate than a survey? Would a mail, telephone, or personal interview fit the problem best?

Tactical research decisions follow once the research approach has been chosen. These include the specific measurements or questions, the structure and length of the questionnaire, and the procedure for choosing a sample. Time and budget also constrain these decisions, so estimated costs must be compared with the anticipated value before the study is implemented.

[Sources]

Research-design process and instructor talk track retained and edited from the instructor-provided source deck.

Three research approaches

EXPLORATORY

Discover possible explanations

Use when little is known about the problem and the goal is to uncover insights, drivers, or decision alternatives.

DESCRIPTIVE

Measure the current situation

Provide an accurate snapshot of customers, behaviors, or other aspects of the market environment.

CAUSAL

Test cause and effect

Determine whether changing one variable causes a change in another.

All marketing research approaches can be classified into these three broad categories.

Exploratory research is used when little information or insight is available. The general problem may be known—for example, declining sales with no clear explanation—but the possible causes are not. Exploratory research helps the researcher understand what may be happening and develop clearer objectives for more directed research.

Descriptive research describes the current situation by asking questions such as who, what, when, where, and how. The original lecture used historical examples from online car shopping: in 2000, 4% of car shoppers purchased online, 40% of new-car purchasers gathered information online, and 85% of BMW buyers used the internet before visiting a dealer. It also recorded a bridal-retail example in which customers tried on 12 dresses and 20% cried. These are legacy teaching examples, not current market statistics.

Causal research permits an inference that one variable causes another, but making that claim requires specific objectives, questions, hypotheses, and a rigorous design. Examples include testing whether a price increase reduces sales or whether one product design performs better than another for a target segment.

The appropriate approach partly depends on uncertainty. When ambiguity is high and little is known about the roots of a problem, begin with exploratory research. As the problem becomes clearer, descriptive or causal research may be appropriate depending on the research question. In practice, a study often combines all three approaches.

[Sources]

Definitions retained and clarified from the instructor-provided source deck.

One study can contain all three approaches

LEGACY INDUSTRY STUDY · 2012

Smart TV sales did not necessarily mean smart-feature use

Many households did not use their TVs’ smart functions.
Users primarily accessed streaming services such as Netflix and Hulu.
In the original study, 25% of households owned a smart TV.

Smart televisions shown in the original lecture

EXPLORATORY

How do consumers use smart-TV technology?

DESCRIPTIVE

What percentage of households own a smart TV?

CAUSAL

What factors increase smart-TV use?

What characterizes exploratory research?

WHAT?

01

Little prior knowledge
The problem is recognized, but its causes are unclear.

02

Search for underlying drivers
Ask what could be causing the problem.

03

Flexible, mostly qualitative methods
Literature reviews, interviews, and case studies.

04

Hypotheses are tentative
They may be vague or may not yet exist.

Exploratory research begins with open questions

LUXURY CLOTHING MARKET

Managers notice a growing group of teenage customers.

They know only that the customers are teenagers.

Who are these customers?What do they want?

Exploratory findings reveal possible drivers

01

Teenage customers do not necessarily belong to the highest-income group.

02

They save money or ask their parents to buy products.

03

They spend mostly on accessories and shoes.

04

Their preferences can differ from their parents’ preferences.

A coordinated handbag, watch, sunglasses, and shoes illustrating the accessory categories named in the research findings

What characterizes descriptive research?

WHAT?
WHO · WHERE
WHEN · HOW

01

Describe the existing situation
Build an accurate snapshot rather than prove causality.

02

Start with tentative hypotheses
The questions are more specific than in exploratory research.

03

Use structured, quantitative methods
Surveys, experiments, and statistical analysis.

04

Report measurable patterns
Example: 50% of people in Indianapolis drink beer once a week.

What can descriptive research tell us?

DESCRIBE GROUPS

Who differs from whom?

How do the demographic characteristics of online buyers differ from those of offline buyers?

ESTIMATE PROPORTIONS

How common is a behavior?

What percentage of customers respond to a sales promotion?

COMPARE BEHAVIORS

Which option is used most?

Which shopping channels are used most often by retired consumers?

Descriptive research makes patterns visible

PER-CAPITA BEER CONSUMPTION

A trend can be described without claiming what caused it.

The chart compares Alberta with Canada over time. It reveals levels and patterns—not causal mechanisms.

Cross-buying can look profitable at first

DESCRIPTIVE RESEARCH QUESTION

Customers who buy across multiple product categories are often assumed to be more profitable.

Is that always true?

How does cross-buying relate to customer profitability?

PAYMENTSCredit card

HOMEMortgage

FUTURERetirement

One customer · multiple categories · different service costs

Cross-buying has a more complicated relationship with profit

01

Across all customers, cross-buying shows a positive relationship with profits.

02

For some customers, cross-buying can be negatively related to profits.

03

Frequent cross-buyers may also complain more and demand more service, creating excessive losses.

HBR · 2012

What characterizes causal research?

PURPOSE

Prove that one variable causes another.

WHY? HOW?

01

The causal factor precedes the outcome.

02

The researcher can manipulate the causal factor.

03

Other variables are fixed or controlled.

04

Hypotheses are specific and the design is carefully structured.

Experiments are the only design that allows researchers to argue causality.

What questions can causal research address?

BEER COMPANY

Will the new TV commercial lead to stronger consumer preference for our brand?

CHARITY

Will mailing an attractive, more expensive brochure increase contributions from previous donors?

LIFE INSURER

Will using computers for client management increase agents’ sales?

A causal question: do return policies change purchasing?

ONLINE RETAIL

Economic research suggested that retailers should toughen online return-shipping policies.

Customers would pay return shipping when the retailer determined that the customer was at fault.

Does a shift in return policy affect customers’ purchase decisions?

ORDER

→

Customer pays return shippingRetailer offers free returns

PURCHASE BEHAVIOR?

A field experiment isolates the policy effect

Researchers conducted a field experiment with three groups to estimate how return policies affected consumers’ online purchase spending.

01Pay for returns

Customers paid return shipping.

↓ 75%–100%spending decreased

02Free returns

Customers were offered free return shipping.

↑ 158%–457%spending increased

03Control

No return policy was offered.

Comparison group

Reported ranges are retained from the original lecture.

Uncertainty guides the research approach

HIGH AMBIGUITYLOW AMBIGUITY

EXPLORATORYAmbiguous problem

Discover plausible explanations.

DESCRIPTIVEProblem recognized

Measure the situation and patterns.

CAUSALProblem clearly defined

Test a specific cause-and-effect claim.

The detective funnel combines the approaches

Many studies use a combination of all three approaches.

EXPLORE

Generate a broad set of possible explanations.

DESCRIBE

Measure which patterns are present and how common they are.

TEST

Narrow the plausible causes with causal evidence.

POSSIBLE EXPLANATIONS

MEASURED PATTERNS

LIKELY CAUSE

In-class exercise 1

WORK WITH YOUR GROUP

Choose a marketing research problem or opportunity at a firm that interests you.

State the problem or opportunity and the actionable management decision.State research objectives and research questions.State your hypotheses.

In-class exercise 2

BUILD ON EXERCISE 1

Formulate research questions or hypotheses that can be addressed with each research approach.

At least one exploratory questionAt least one descriptive questionAt least one causal question or hypothesis

To sum up

01

Exploratory

Discover what might be happening.

02

Descriptive

Measure what is happening.

03

Causal

Test why it is happening.

Choose based on uncertainty and the research question.Many studies move through more than one approach.

Next class

COMING NEXT

Sources of
Secondary Data

01Problem formulation

02Research design

03Data collection

04Data analysis & recommendations

How can existing information answer a new marketing question?