Research Problems, Objectives and Decision Frameworks

Introduction

Marketing research begins before a questionnaire is written or data are collected. It begins when a manager recognizes that a decision cannot be made confidently with the information currently available. A product team may know that sales are declining, but not whether the cause is price, distribution, product experience, competitive activity, or weak customer awareness. The task of the researcher is to convert this practical uncertainty into a researchable problem and then into evidence that supports a decision.

This chapter explains how product and marketing professionals move from a broad management concern to a precise research problem, research objectives, research questions, hypotheses, and decision criteria. It also distinguishes exploratory, descriptive, and causal research and shows how these approaches can be combined in a research brief.

Learning Objectives

By the end of this chapter, you should be able to:

1. Distinguish a management decision problem from a marketing research problem.
2. Diagnose and define research problems using symptoms, causes, and information gaps.
3. Write clear research objectives and research questions.
4. Develop testable hypotheses and identify variables.
5. Select exploratory, descriptive, or causal research appropriately.
6. Establish decision criteria and connect evidence to action.
7. Prepare a professional research brief.
8. Recognize ethical, practical, and technological issues in problem definition.

1. The Nature of a Marketing Research Problem

A marketing research problem is a clearly stated information problem that can be investigated systematically. It identifies what is unknown, why the information matters, and the population, market, product, or situation to which the inquiry relates.

A management problem asks what the organization should do. A research problem asks what must be learned in order to make that decision responsibly.

The distinction is important because managers often state symptoms rather than problems. “Sales are falling” is an observation. It does not identify the underlying cause or the information needed. A useful research problem should be specific enough to guide study design but open enough to permit discovery.

Symptoms, causes, and information gaps

A symptom is an observable condition, such as falling conversion, increased complaints, or declining retention. A possible cause is an explanation for that condition, such as confusing onboarding or a competitor’s lower price. An information gap is the uncertainty that prevents a decision-maker from evaluating those explanations.

Researchers should avoid assuming that the most visible explanation is the correct one. For example, a fall in sales may be attributed to price because competitors are cheaper. Research may instead reveal that customers do not understand the product’s value, or that the product is unavailable in important channels.

A useful diagnostic sequence is:

1. What decision must be made?
2. What is currently known?
3. What is uncertain or disputed?
4. What alternative explanations are plausible?
5. What evidence would distinguish among those explanations?
6. What action would follow from each possible finding?

2. Defining the Problem Systematically

Problem definition is an iterative process rather than a single meeting. It normally combines discussions with decision-makers, review of existing data, observation of customers and employees, and examination of relevant market or academic knowledge.

The problem-definition statement

A practical problem statement normally includes five elements:

• Context: the product, market, organization, and relevant time period.
• Decision: the choice that management must make.
• Symptoms: the observed performance or customer issue.
• Possible explanations: competing causes that should be investigated.
• Information need: the evidence required to compare alternatives.

For example:

A subscription-based fitness application has experienced a decline in three-month retention among new users during the last two quarters. Management must decide whether to redesign onboarding, introduce a lower-priced plan, or increase personal coaching. The study must determine the main reasons for early cancellation, identify differences among user segments, and estimate the relative importance of onboarding quality, perceived value, and coaching support.

This statement is more useful than “find out why customers leave” because it specifies the decision and the competing actions.

The “five whys” technique

The five-whys technique asks successive questions about an observed problem. It can uncover assumptions, although it should not be treated as proof of causation.

• Why are trial users not converting? Because many do not complete the setup process.
• Why do they not complete setup? Because the process requires several technical permissions.
• Why are those permissions required? Because the product team designed personalization around them.
• Why was that design chosen? Because personalization was expected to increase engagement.
• Why has the assumption not been tested? Because the team measured activation only after setup, not abandonment during setup.

The research problem may therefore concern the effect of setup friction on conversion, rather than simply the general appeal of the product.

Stakeholder consultation and secondary evidence

Interviews with executives, product managers, sales staff, customer-service agents, and channel partners can reveal competing interpretations of the problem. Internal data—such as funnel analytics, churn records, complaint categories, campaign performance, and customer relationship management records—may identify where the issue occurs.

Secondary research can provide market size, competitor positioning, category trends, regulatory requirements, and prior findings. It also prevents the organization from commissioning primary research to answer a question that existing evidence can answer adequately.

However, internal and secondary data should be evaluated for relevance, accuracy, completeness, timeliness, and possible bias. A dashboard may show what happened without explaining why it happened. Customer complaints may reveal important problems but usually do not represent all customers.

3. From Problems to Objectives and Questions

Research objectives

A research objective states what the study will accomplish. Strong objectives are specific, measurable where appropriate, relevant to the decision, and feasible within available time and resources.

Weak objective: “Understand customers.”

Stronger objectives:

• Identify the principal reasons for cancellation among customers who joined within the last six months.
• Compare perceived value between retained and cancelled customers.
• Estimate the proportion of target customers who recognize the proposed brand position.
• Evaluate whether a simplified onboarding process increases activation under a controlled test.

Objectives should use action verbs such as identify, describe, compare, estimate, assess, evaluate, test, or explain. They should not promise more than the design can deliver. A cross-sectional survey can describe associations, but by itself it generally cannot establish that one factor caused another.

Research questions

Research questions convert objectives into answerable inquiries. They should specify the population, concepts, relationships, and time frame where relevant.

Examples:

• What are the most common reasons for cancellation among new subscribers?
• How do needs differ between occasional and frequent users?
• What benefits do target customers associate with the proposed product concept?
• Is exposure to the revised advertisement associated with higher brand recall?
• Does the simplified onboarding flow increase completion compared with the current flow?

A research question should not combine too many independent issues. “What do customers think about price, service, design, competitors, advertising, and distribution?” is too broad for one objective and may produce unfocused data.

Research questions and decision questions

A decision question concerns action: “Should the company launch?” A research question concerns evidence: “What proportion of the target segment indicates purchase intention at the proposed price, and what objections do non-buyers report?” Several research questions may be required to answer one decision question.

4. Hypotheses, Concepts, and Variables

A hypothesis is a provisional, empirically testable statement about a relationship or difference. It is not simply a prediction or a belief; it must identify concepts that can be observed or measured.

Examples:

• Customers who report greater perceived value will be more likely to renew.
• Exposure to the benefit-focused advertisement will produce higher aided brand recall than exposure to the feature-focused advertisement.
• The simplified checkout process will have a higher completion rate than the current process.

A hypothesis normally identifies:

• Independent variable: the factor expected to influence an outcome.
• Dependent variable: the outcome being explained or predicted.
• Unit of analysis: the customer, household, account, advertisement exposure, store, or other entity studied.
• Population: the broader group to which the findings may apply.

A null hypothesis represents no difference or relationship, while an alternative hypothesis represents the expected difference or relationship. Researchers may use formal statistical tests later, but the conceptual hypothesis should be clear before data collection.

Not every study requires formal hypotheses. Exploratory research often begins with broad questions because relevant variables have not yet been identified. Hypotheses become more useful when prior theory, qualitative findings, or existing evidence provides a basis for specifying expected relationships.

Operational definitions

Abstract concepts must be translated into observable indicators. “Customer loyalty” might be represented by renewal intention, repeat purchase, recommendation intention, or actual renewal behavior. These indicators are not identical, so the researcher must state which meaning is being used.

An operational definition should explain how a concept will be observed or measured. For example:

Activation is defined as completing account setup and performing at least one core product action within seven days of registration.

Clear operational definitions improve consistency and reduce disputes about interpretation.

5. Exploratory, Descriptive, and Causal Research

The three categories describe the primary purpose of a study. They are not always separate stages; a project may use more than one.

Exploratory research

Exploratory research investigates an incompletely understood problem and helps identify relevant concepts, language, explanations, or hypotheses. Common methods include secondary research, expert interviews, customer interviews, focus groups, observation, diary studies, and analysis of support conversations.

Its strengths are flexibility, depth, and sensitivity to unexpected issues. Its limitations include small or non-representative samples, interpretive complexity, and limited ability to estimate population prevalence or establish causality.

Exploratory research is appropriate when:

• the problem is poorly defined;
• managers disagree about possible causes;
• the product category or customer need is unfamiliar;
• the researcher needs customer language for a later survey; or
• new hypotheses must be generated.

Descriptive research

Descriptive research measures the characteristics, attitudes, behaviors, or incidence of a population or phenomenon. Surveys, structured observation, customer databases, panels, and market measurement systems are common sources.

Typical descriptive questions include:

• What percentage of customers use the feature weekly?
• Which segment has the highest awareness?
• How satisfied are customers with delivery?
• What proportion would consider the proposed product?

Descriptive research depends heavily on sampling, measurement quality, and question wording. It can identify patterns and associations but a one-time descriptive study usually cannot demonstrate that an observed factor caused an outcome.

Causal research

Causal research evaluates whether changing one factor produces a change in another. Experiments are the strongest practical design for causal inference because they manipulate an independent variable and compare outcomes under controlled conditions.

A product team might randomly assign users to the existing onboarding flow or a revised flow and compare activation. An advertising team might expose comparable audience groups to different creative executions and compare conversion or recall.

A causal claim is more credible when:

• the presumed cause occurs before the outcome;
• alternative explanations are controlled or balanced;
• the variables are measured reliably; and
• the relationship is replicated or supported by relevant evidence.

A/B testing is a common field experiment. It is most informative when assignment is genuinely random, groups are sufficiently large, exposure is correctly recorded, and the outcome is defined in advance. A difference in outcomes after a non-random campaign does not automatically prove that the advertisement caused the difference; audiences may differ in motivation, timing, or prior brand knowledge.

Combining the approaches

A staged design is often more useful than choosing one category in isolation. Exploratory interviews may identify barriers to adoption. A descriptive survey may estimate how common those barriers are. A controlled experiment may then test whether a redesigned message or onboarding flow improves adoption.

The sequence should reflect the decision, not habit. A team should not conduct focus groups merely because they are familiar, nor run a large survey before understanding the relevant customer vocabulary.

6. Decision Frameworks and Criteria

Research creates value only when findings can influence a decision. A decision framework links alternatives, evidence, thresholds, risks, and actions.

The alternative–evidence–action framework

For each realistic management alternative, specify:

1. the action under consideration;
2. the evidence needed to evaluate it;
3. the criterion for proceeding;
4. the risks and constraints; and
5. the action if the criterion is not met.

Criteria may be quantitative, qualitative, or mixed. A quantitative criterion might specify a minimum activation improvement, while a qualitative criterion might require that no serious safety, privacy, or accessibility concern remain unresolved.

Thresholds should be agreed before results are seen whenever possible. Pre-specified criteria reduce the temptation to change the standard after observing favorable or unfavorable findings. They do not eliminate managerial judgment; rather, they make that judgment more transparent.

Decision value and uncertainty

A research project should be proportionate to the consequences of the decision. More rigorous and expensive research is justified when the decision is costly, difficult to reverse, or ethically sensitive. A small exploratory study may be sufficient for improving wording in a low-risk campaign, whereas a major product launch may require representative measurement, usability evaluation, pilot testing, and legal review.

Decision-makers should distinguish statistical evidence from business significance. A very small difference may be statistically detectable in a large sample but commercially unimportant. Conversely, an important operational risk may deserve attention even when a small study cannot establish a precise estimate.

7. Preparing a Professional Research Brief

A research brief is a shared agreement between the commissioning organization and the research team. It prevents scope drift and makes assumptions visible.

A concise brief should include:

1. Background and context: What has happened, and why is the issue important now?
2. Management decision: What decision will the findings support?
3. Research problem: What information is missing?
4. Objectives: What must the study accomplish?
5. Research questions or hypotheses: What will be investigated?
6. Target population and unit of analysis: Who or what is being studied?
7. Geographic and time scope: Where and when does the study apply?
8. Known evidence: What internal and secondary data already exist?
9. Proposed methodology: Which design and methods are suitable, and why?
10. Key measures: How will important concepts be operationalized?
11. Decision criteria: What findings would support each action?
12. Deliverables: What reports, dashboards, prototypes, or presentations are required?
13. Timeline and budget: What resources and deadlines apply?
14. Responsibilities: Who approves instruments, recruitment, analysis, and decisions?
15. Ethics and governance: What privacy, consent, security, accessibility, or regulatory requirements apply?
16. Limitations and risks: What may prevent reliable conclusions?

A brief should state what is out of scope. For example, a study designed to diagnose early churn should not quietly become a full evaluation of brand strategy, pricing, advertising, and channel performance without additional resources and revised objectives.

8. Illustrative Case: The Declining Conversion of LearnLoop

This is a fictional illustrative case.

Background

LearnLoop, a mobile learning application, offers a seven-day free trial followed by a monthly subscription. Trial registrations have increased, but conversion to paid membership has declined from 12% to 8% over two quarters. The product team believes that the price is too high. The marketing team suspects that advertising attracts users with expectations the product does not meet.

Problem

The company must decide whether to change price, redesign the trial experience, revise advertising claims, or pursue a combination of actions. Existing analytics show where users stop engaging but do not explain their reasons.

Analysis and research objectives

The team defines three objectives:

1. Identify the main reasons trial users fail to convert.
2. Compare expectations and perceived value among converters and non-converters.
3. Test whether a revised expectation-setting message improves completion of a core learning activity and paid conversion.

The first objective is exploratory. The second is descriptive and comparative. The third is causal.

Approach

The research team first conducts interviews with recent converters and non-converters, reviews support contacts, and observes five users completing onboarding. These activities reveal two unexpected issues: some users believe the trial includes live tutoring, and others cannot find content at the difficulty level promised in the advertisement.

A survey of recent trial users then measures advertising exposure, expected benefits, perceived value, onboarding clarity, and conversion intention. Finally, eligible new users are randomly assigned to either the existing onboarding message or a revised message that clearly explains the included learning format and content levels. The primary outcome is paid conversion within the defined follow-up period; completion of the first learning activity is a secondary outcome.

Decision criteria

The team agrees that the revised message will be recommended for wider rollout if it produces a meaningful improvement in conversion without increasing complaints or misleading impressions. A price change will not be recommended solely because non-converters mention price; the team will also examine whether perceived value and expectation mismatch explain the response.

Outcome

Suppose the research finds that many non-converters expected live tutoring, while the revised message reduces that misunderstanding and improves first-activity completion. The company decides to revise advertising and onboarding before testing a price change. It also adds clearer content-level information to the trial page.

Learning points

The visible symptom was lower conversion, but the research uncovered an expectation gap. The staged design prevented the team from treating price as the only explanation. It also connected each method to a distinct information need and each finding to a pre-agreed action.

9. Quality, Ethics, and Governance

Problem definition has ethical consequences because it determines whose experiences are measured and which actions may follow. Researchers should avoid framing a study to justify a preferred decision. Questions should not be designed to manipulate respondents or conceal material information.

Important safeguards include:

• obtain informed consent where required;
• collect only data necessary for the stated purpose;
• explain how personal information will be used and retained;
• protect confidential and sensitive data;
• avoid unnecessary collection of sensitive characteristics;
• provide accessible participation options;
• distinguish research from sales or marketing communication;
• disclose important limitations and conflicts of interest; and
• report unfavorable findings honestly.

When using customer data, product analytics, or advertising platforms, teams should check applicable privacy and consumer-protection requirements. Consent, lawful basis, retention, access control, and data minimization should be considered at the brief stage rather than after collection.

The role of AI and automated analytics

AI tools can help summarize interviews, classify support conversations, generate alternative hypotheses, or identify unusual patterns in large datasets. They can also reproduce bias, misclassify context, expose confidential information, or create apparently convincing but unsupported interpretations. Human researchers must verify outputs against source evidence, protect confidential data, document important prompts or transformations, and ensure that automated classifications do not become consequential decisions without review.

AI-generated research questions should be treated as candidate inputs, not as evidence. The responsibility for validity, consent, fairness, and interpretation remains with the research team and commissioning organization.

10. Common Errors and Best Practices

Common errors

• Defining the problem as a symptom rather than an information gap.
• Assuming one cause before investigating alternatives.
• Writing objectives so broad that no study can complete them.
• Asking questions that cannot be answered by the proposed method.
• Treating stated purchase intention as equivalent to actual behavior.
• Claiming causation from a simple correlation or one-time survey.
• Selecting a method because it is fashionable rather than appropriate.
• Omitting decision criteria until after results are available.
• Ignoring non-users, dissatisfied customers, or excluded segments.
• Reporting findings without explaining uncertainty and limitations.

Best practices

Begin with the decision, not the preferred method. Separate symptoms, causes, and information gaps. Write objectives using precise verbs. Map every objective to one or more research questions and every question to a method or data source. Define the target population and unit of analysis. Pre-specify important outcomes and decision thresholds. Pilot instruments and revise ambiguous wording. Keep the research brief current when scope changes. Finally, present findings in a way that makes the connection between evidence, uncertainty, and action explicit.

Chapter Summary

Marketing research supports decisions by converting uncertainty into structured evidence. A management problem asks what the organization should do, whereas a research problem specifies what must be learned to make that decision. Effective problem definition distinguishes symptoms from possible causes and identifies the information gap.

Research objectives state what a study will accomplish; research questions make those objectives answerable; hypotheses express testable expectations about variables and relationships. Exploratory research clarifies poorly understood issues, descriptive research measures characteristics and patterns, and causal research evaluates the effects of interventions. These approaches may be combined in a staged design.

A decision framework links alternatives to evidence, criteria, risks, and actions. A research brief records the context, decision, objectives, population, methods, measures, deliverables, resources, and governance requirements. Strong practice is systematic, proportionate, transparent, ethically responsible, and attentive to the difference between association and causation.

Key Terms

Management decision problem: The action or choice facing an organization.

Marketing research problem: A defined information gap that can be investigated systematically.

Symptom: An observable condition indicating that a problem may exist.

Information gap: Relevant knowledge that is missing or uncertain.

Research objective: A precise statement of what a study intends to accomplish.

Research question: An answerable inquiry derived from a research objective.

Hypothesis: A provisional, testable statement about a relationship or difference.

Independent variable: A factor proposed to influence an outcome.

Dependent variable: The outcome being explained or measured.

Unit of analysis: The entity about which data and conclusions are made.

Operational definition: A specification of how an abstract concept will be observed or measured.

Exploratory research: Research used to clarify problems and generate concepts or hypotheses.

Descriptive research: Research that measures characteristics, behaviors, attitudes, or incidence.

Causal research: Research designed to assess whether an intervention produces an outcome.

Experiment: A design involving manipulation and comparison under controlled conditions.

Decision criterion: A pre-specified standard used to guide action based on findings.

Research brief: A document defining the purpose, scope, design, resources, and governance of a study.

Secondary data: Existing information collected for an earlier or different purpose.

Target population: The broader group to which a study seeks to apply its findings.

Scope drift: Uncontrolled expansion of a project beyond its agreed purpose and resources.

Review Questions

Short Answer Questions

1. What is the difference between a management decision problem and a marketing research problem?
2. Why should researchers distinguish symptoms from possible causes?
3. State two characteristics of a strong research objective.
4. When is exploratory research especially appropriate?
5. What is the purpose of a decision criterion?

Long Answer Questions

1. Explain the process of converting a broad business concern into a research problem, objectives, and research questions.
2. Compare exploratory, descriptive, and causal research in terms of purpose, methods, strengths, and limitations.
3. Explain the role of hypotheses, variables, units of analysis, and operational definitions in marketing research.
4. Describe the essential contents of a professional research brief and explain why each is important.
5. Discuss how ethical and privacy considerations should influence problem definition and research design.

Analytical / Application Questions

1. A retailer observes declining sales after a competitor launches a discount campaign. Develop a research problem, three research objectives, and one suitable method for each objective.
2. A product team claims that users who watch more tutorial videos retain longer. Explain why this observation does not by itself establish causation, and propose a stronger design.
3. A company is considering a new subscription tier. Construct a decision framework showing two alternatives, the evidence required, decision criteria, and possible actions.

Multiple Choice Questions with Answers

Question 1

Which statement best describes a marketing research problem?

A. A final decision that management has already made

B. An information gap that can be investigated systematically

C. A company’s annual sales target

D. A promotional message for a product

Correct Answer: B. An information gap that can be investigated systematically

Question 2

Which research objective is written most precisely?

A. Understand customers better

B. Study the market in detail

C. Identify the main reasons for cancellation among customers who joined in the last six months

D. Find useful information about the product

Correct Answer: C. Identify the main reasons for cancellation among customers who joined in the last six months

Question 3

Which type of research is primarily used to generate concepts and hypotheses when a problem is poorly understood?

A. Exploratory research

B. Descriptive research

C. Causal research

D. Audit research

Correct Answer: A. Exploratory research

Question 4

Which method provides the strongest practical basis for assessing whether a product change causes an outcome?

A. A one-time customer interview

B. A non-random opinion poll

C. A controlled experiment with comparison groups

D. A review of competitor advertisements

Correct Answer: C. A controlled experiment with comparison groups

Question 5

In the hypothesis ‘Customers reporting greater perceived value are more likely to renew,’ what is the dependent variable?

A. Perceived value

B. Customer age

C. Renewal behavior or intention

D. Research method

Correct Answer: C. Renewal behavior or intention

Question 6

What is the main purpose of an operational definition?

A. To select the research budget

B. To specify how an abstract concept will be observed or measured

C. To guarantee that a hypothesis is true

D. To replace the research question

Correct Answer: B. To specify how an abstract concept will be observed or measured

Question 7

Which item should normally be included in a research brief?

A. Only the questionnaire wording

B. The decision, objectives, target population, methods, and decision criteria

C. Only the researcher’s preferred statistical software

D. A guarantee of a positive business outcome

Correct Answer: B. The decision, objectives, target population, methods, and decision criteria

Question 8

A survey finds that customers who use a feature more often also renew more often. What is the most appropriate conclusion from this finding alone?

A. Feature use definitely causes renewal

B. Renewal definitely causes feature use

C. The variables are associated, but causation has not been established

D. Neither variable can be measured

Correct Answer: C. The variables are associated, but causation has not been established

Question 9

Which decision criterion is most useful?

A. Use the results if they look interesting

B. Proceed if the revised onboarding improves activation without creating unacceptable support costs

C. Proceed if every respondent likes the product

D. Change the criterion after seeing the results

Correct Answer: B. Proceed if the revised onboarding improves activation without creating unacceptable support costs

Question 10

Which sequence best represents a staged research design for a poorly understood adoption problem?

A. Causal experiment, then define the problem, then interview users

B. Exploratory research, descriptive measurement, then causal testing where appropriate

C. Advertising launch, then sampling, then formulate the decision

D. Decision first, then select findings that support it

Correct Answer: B. Exploratory research, descriptive measurement, then causal testing where appropriate