Foundations of Customer and Digital Analytics

Introduction and Context

Customer and digital analytics are central to contemporary business decision making. Organizations now collect information from transactions, websites, mobile applications, social platforms, customer-service interactions, loyalty programs, connected devices, and market research. The availability of data, however, does not automatically create insight. Data becomes strategically valuable only when it is interpreted in relation to a business question and used to support an appropriate decision.

Customer analytics examines customers, prospects, households, and customer relationships. Digital analytics focuses on behaviour and performance in digitally mediated environments, such as websites, applications, search engines, digital advertising, email, social media, and digital commerce. The two fields overlap substantially. A customer may browse a website, receive an email, contact a service centre, purchase in a store, and later post a review. Understanding the relationship requires integration across channels rather than isolated measurement of each interaction.

The central question is not simply “What happened?” It is “What evidence can help the organization create value for customers and stakeholders, and what action should follow?” This question connects analytics with market orientation, customer centricity, digital transformation, and the customer lifecycle.

Learning Objectives

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

1. Explain the evolution, scope, and terminology of customer and digital analytics.
2. Relate market orientation, customer centricity, digital transformation, and the customer lifecycle to evidence-based decision making.
3. Distinguish descriptive, diagnostic, predictive, causal, and prescriptive analytics.
4. Select suitable analytical questions, data, and metrics for common business situations.
5. Recognize limitations, ethical risks, governance requirements, and professional responsibilities in customer analytics.

1. The Evolution of Customer and Digital Analytics

From market research to data-intensive decision making

Early marketing decisions relied heavily on managerial experience, sales reports, surveys, focus groups, and periodic market research. These sources remain useful, but they often provided relatively infrequent and aggregated evidence. A manager might learn that a product’s quarterly sales had declined, but not be able to determine which customer journeys, messages, or service failures contributed to the decline.

The development of customer databases and customer relationship management systems enabled organizations to store customer identifiers, transactions, contacts, preferences, and service histories. This encouraged database marketing: using customer-level information to segment audiences, personalize communication, and evaluate relationship value.

The growth of the internet introduced clickstream data, search behaviour, online conversion data, digital advertising impressions, and email engagement measures. Mobile devices added location, application events, and continuous interaction data. Social platforms generated public and semi-public expressions of opinion, although these data require careful interpretation because visibility does not necessarily equal representativeness.

Cloud computing, application programming interfaces, data warehouses, customer data platforms, machine learning, and real-time processing have expanded the scale and speed of analysis. Organizations can now combine data from many touchpoints and produce automated recommendations. At the same time, the risks of surveillance, discrimination, poor data quality, cyberattack, and unjustified automated decisions have also increased.

A shift in managerial logic

The evolution can be summarized as a movement:

This evolution does not mean that older methods are obsolete. A well-designed survey may reveal motivations that clickstream data cannot observe. Qualitative interviews may explain frustration that a dashboard merely detects. Strong analytics combines behavioural, attitudinal, operational, and contextual evidence.

2. Scope and Terminology

Data, information, insight, and decision

These terms should be distinguished:

• Data are recorded observations, such as an order amount, page view, survey response, or service-contact timestamp.
• Information is organized data that answers a basic question, such as monthly revenue by region.
• Insight is a meaningful interpretation that explains a situation or identifies an opportunity, such as evidence that mobile checkout friction is concentrated among first-time users.
• Decision is a selected course of action, such as simplifying the mobile payment process and testing its effect.

Analytics is the disciplined process of transforming data into evidence that supports decisions. A dashboard is not itself an analytical conclusion; it is an instrument for presenting selected measures.

Customer analytics

Customer analytics includes the measurement and analysis of:

• acquisition and prospecting;
• customer segmentation and targeting;
• purchasing and usage behaviour;
• customer profitability and lifetime value;
• satisfaction, loyalty, and advocacy;
• churn, retention, and reactivation;
• customer service and experience;
• cross-selling and up-selling;
• campaign response; and
• fairness and accessibility across customer groups.

A customer is not merely a transaction. Customer analysis considers needs, expectations, value exchange, relationships, context, and the consequences of organizational actions.

Digital analytics

Digital analytics is the collection, measurement, analysis, and interpretation of digital interaction data. Common sources include:

• website and application event logs;
• search queries and search advertising;
• email delivery and engagement records;
• digital advertising exposure and response;
• social media interactions;
• online reviews and user-generated content;
• digital commerce and payment systems; and
• connected-product or Internet of Things data.

Digital analytics commonly measures traffic, engagement, conversion, retention, acquisition cost, campaign response, and user flows. However, a metric is useful only when its definition, denominator, time period, attribution rule, and business purpose are clear.

Related terms

Customer relationship management (CRM) refers to the processes and technologies used to manage customer relationships and interactions. CRM data may support analytics, but CRM and analytics are not identical.

Customer experience (CX) is the customer’s overall perception of interactions with an organization across touchpoints and over time.

Customer journey is the sequence of stages and interactions through which a customer moves in relation to a need, purchase, use, or relationship.

Touchpoint is a specific interaction between a customer and an organization, such as an advertisement, website visit, purchase, delivery, or support call.

Conversion is a defined desired action, such as registration, purchase, application completion, or subscription.

Engagement refers to interaction or involvement, but it must be operationally defined. A video view, comment, repeat visit, and completed product configuration are not equivalent forms of engagement.

Segmentation divides a market or customer base into groups that are meaningfully similar for a specified purpose. Segments may be demographic, geographic, behavioural, psychographic, value-based, or needs-based.

Personalization adapts content, offers, interfaces, or service based on customer information. Personalization should be relevant, proportionate, transparent, and respectful of customer choice.

Attribution assigns credit for an outcome to marketing contacts or other influences. Simple last-touch attribution is easy to implement but can misrepresent journeys involving multiple interactions.

Cohort means a group sharing a common starting event or characteristic, such as customers acquired in a particular month. Cohort analysis helps distinguish customer age effects from calendar-period effects.

Key performance indicator (KPI) is a measure linked to an important objective. Not every available metric is a KPI.

3. Market Orientation and Customer Centricity

Market orientation

Market orientation is an organizational philosophy emphasizing the generation of market intelligence, dissemination of that intelligence across functions, and coordinated responsiveness. It therefore involves more than asking the marketing department to study customers. Product, operations, finance, sales, technology, and service teams must share relevant knowledge and act coherently.

A market-oriented organization asks:

1. What needs, alternatives, constraints, and expectations exist in the market?
2. How is this knowledge communicated to the people who can act on it?
3. What coordinated response creates sustainable value while supporting organizational objectives?

Analytics supports market orientation by making evidence more timely, comparable, and actionable. It does not replace interpretation of competitors, regulation, culture, or unobserved needs.

Customer centricity

Customer centricity places customer value and long-term relationship quality at the centre of organizational decisions. It does not mean satisfying every request regardless of cost, nor does it mean maximizing a single customer metric. A customer-centric organization balances customer needs with feasibility, profitability, employee capability, social responsibility, and legal obligations.

Customer-centric analytics should answer questions such as:

• Which customer problems are most important and solvable?
• Which customers are underserved or exposed to unnecessary friction?
• Does a proposed intervention improve outcomes for customers as well as the firm?
• Are apparent differences between segments caused by real needs or by biased measurement?

A common failure is metric substitution: optimizing an easily measured activity rather than the underlying customer objective. For example, increasing email clicks may reduce trust if the messages are excessive or misleading. A broader scorecard should connect activity metrics to customer and business outcomes.

4. Digital Transformation and Analytics Capability

Digital transformation is the organizational redesign of processes, capabilities, technologies, and sometimes business models enabled by digital technologies. It is not equivalent to purchasing software or moving documents into a cloud system.

Customer analytics contributes to transformation in four ways:

1. Visibility: digital traces make customer journeys and operational bottlenecks more observable.
2. Coordination: shared data and definitions connect marketing, sales, service, product, and operations.
3. Experimentation: digital channels make it easier to test variations and learn from outcomes.
4. Adaptation: predictive and prescriptive systems can support timely responses.

Transformation also creates implementation challenges. Data may be held in incompatible systems, departments may use conflicting definitions, incentives may encourage local optimization, and employees may distrust automated recommendations. Consequently, analytics capability includes people, processes, governance, technology, data quality, and organizational culture.

A useful maturity progression is:

Maturity is not a race toward maximum automation. A highly automated process with poor data or weak governance can be less capable than a carefully managed analytical process using simpler tools.

5. The Customer Lifecycle as an Analytical Framework

The customer lifecycle describes the stages through which a relationship may develop. A common framework includes awareness, consideration, acquisition, onboarding, usage, retention, expansion, advocacy, and possible churn or reactivation. The exact stages vary by industry and business model.

Lifecycle analysis prevents an organization from evaluating every customer with the same objective. For example:

Lifecycle stages are analytical constructs, not universal facts. A customer may move backward, skip stages, or occupy several relationships simultaneously. The framework should therefore be tested against actual behaviour and customer research.

Selected lifecycle metrics

A basic conversion rate is:

The denominator must be defined carefully. If 800 eligible visitors produce 32 purchases, the conversion rate is 32 ÷ 800 × 100% = 4%. Counting all website visitors instead of eligible visitors may produce a different and less meaningful measure.

A simple retention rate for a period is:

This formulation assumes that the customer population and start/end definitions are consistent. If there are mergers, migrations, paused accounts, or different contractual rules, the calculation requires adjustment.

A simple churn rate is often defined as:

Retention and churn are not always exact complements because organizations may use different populations, timing rules, and treatment of new customers.

A simplified customer lifetime value estimate for a contractual business may be expressed as:

Here, CLV is a simplified contribution-based estimate; average revenue and churn must use the same period, and the churn rate must be greater than zero. The estimate assumes relatively stable revenue, margin, and churn, and it ignores discounting, expansion, service costs, and changing behaviour. More advanced models use contribution margins, retention curves, discount rates, and customer-specific forecasts.

6. Five Forms of Analytics

The forms of analytics are complementary rather than strictly separate. A responsible analysis often progresses from description to diagnosis, prediction, causal learning, and action.

Descriptive analytics: What happened?

Descriptive analytics summarizes historical or current data. Examples include monthly revenue, active users, conversion by device, average response time, and retention by cohort. It uses aggregation, visualization, frequency distributions, and dashboards.

Descriptive analysis is foundational but limited. A decline in conversion is a fact; it does not establish why the decline occurred or what intervention will improve it.

Diagnostic analytics: Why did it happen?

Diagnostic analytics investigates patterns and possible explanations. Analysts may compare segments, examine funnels, drill into time periods, inspect error logs, link service contacts to outcomes, or analyze qualitative feedback.

Diagnostic evidence remains vulnerable to confounding. If mobile conversion is lower than desktop conversion, the difference might reflect screen design, customer intent, payment options, traffic source, product mix, or unobserved customer characteristics. Association is not proof of cause.

Predictive analytics: What is likely to happen?

Predictive analytics estimates future or unobserved outcomes using historical data and statistical or machine-learning models. Applications include churn risk, demand forecasting, propensity to respond, fraud detection, and estimated delivery time.

A predictive model may be evaluated using measures such as accuracy, precision, recall, calibration, mean absolute error, or area under a receiver operating characteristic curve, depending on the task. Evaluation must reflect the practical cost of false positives and false negatives. A high overall accuracy can be misleading when the event of interest is rare.

A prediction is not a causal explanation. A churn score identifies customers who resemble previously churned customers; it does not necessarily reveal which intervention will retain them.

Causal analytics: What caused what?

Causal analytics estimates the effect of an intervention or exposure. Examples include whether a discount increased incremental purchases, whether a redesigned checkout caused higher completion, or whether a service intervention reduced churn.

The strongest practical design is often a randomized controlled experiment. Eligible units are randomly assigned to treatment and control groups, and outcomes are compared under a defined protocol. Randomization helps balance observed and unobserved characteristics in expectation. Observational data may also support causal analysis through matching, regression adjustment, instrumental variables, difference-in-differences, or other methods, but each requires assumptions that must be examined.

A key distinction is between observed outcome differences and incremental effects. If a treated group purchases more than a control group, the incremental effect is commonly estimated as:

The units of the outcome must match. If the outcome is revenue per customer, the effect is revenue per customer. The estimate is interpretable as causal only when the design and assumptions justify that interpretation, including appropriate assignment, measurement, compliance, and absence of material interference between units.

Prescriptive analytics: What should be done?

Prescriptive analytics recommends actions by combining predictions, objectives, constraints, and possible interventions. Examples include selecting an offer, allocating a service team, deciding which customers to contact, or choosing an advertising budget.

Prescriptive systems should make assumptions and constraints visible. A recommendation that maximizes short-term revenue might increase complaints, regulatory risk, or long-term churn. A responsible objective function therefore considers multiple outcomes, such as profit, customer benefit, fairness, service capacity, and risk.

7. A Decision-Centred Analytics Process

A professional analytics project begins with a decision, not a dataset.

Step 1: Frame the decision

Specify who will decide, what action is available, when it must be taken, and what outcome matters. “Understand customers” is too broad; “Decide whether to simplify the mobile checkout for first-time users” is more actionable.

Step 2: Define constructs and metrics

Clarify terms such as active user, conversion, retained customer, complaint, and revenue. Define the population, unit of analysis, time window, inclusion rules, and denominator.

Step 3: Assess data quality and provenance

Check completeness, accuracy, consistency, timeliness, duplication, missingness, identity resolution, and possible selection bias. Document where the data came from and how it was transformed.

Step 4: Match method to question

Use descriptive analysis for status reporting, diagnostic analysis for investigation, predictive models for forecasting or prioritization, causal designs for intervention effects, and prescriptive methods for constrained choices.

Step 5: Analyze and validate

Check assumptions, compare alternative explanations, test robustness, and separate development data from evaluation data when building predictive models. In experiments, monitor assignment integrity and pre-specified outcomes.

Step 6: Communicate evidence and uncertainty

A useful report states what is known, what is uncertain, what assumptions were made, and what action is recommended. Visualizations should clarify rather than exaggerate differences.

Step 7: Implement, monitor, and learn

An analytical recommendation is not the end of the process. Monitor outcomes, unintended effects, customer feedback, model drift, fairness indicators, and whether the action remains appropriate as conditions change.

8. Illustrative Case: Improving a Mobile Subscription Journey

Background

This fictional case concerns Northstar Mobile, a subscription-based telecommunications provider. The company observed that website visitors using mobile devices completed subscriptions less frequently than desktop visitors.

Problem

A dashboard showed the following monthly figures:

Desktop conversion was 600 ÷ 10,000 × 100% = 6%. Mobile conversion was 800 ÷ 20,000 × 100% = 4%. The descriptive result indicated a gap but did not establish its cause.

Analysis

The analytics team examined the mobile funnel and found that many users abandoned the payment step. Customer-service transcripts suggested confusion about identity verification, while usability testing identified a form that was difficult to complete on small screens. The team also checked traffic sources and found that mobile visitors included a higher proportion of first-time prospects, which could partly explain the difference.

The team therefore avoided concluding that “mobile users are less valuable.” Instead, it formed two hypotheses: payment-form friction reduced completion, and the mobile audience had a different customer mix.

Approach and solution

Northstar designed a randomized experiment among eligible mobile checkout users. The control group saw the existing form. The treatment group saw a simplified form with clearer instructions, fewer simultaneous fields, and an alternative verification sequence. The primary outcome was completed subscription within the session; secondary outcomes included verification failure, customer-service contact, and early cancellation.

The organization pre-defined privacy protections, retained the existing consent requirements, and ensured that the alternative process met security and regulatory standards. It also monitored whether the treatment affected different customer groups unevenly.

Suppose the control group contained 15,000 users and 600 completed subscriptions, while the treatment group contained 15,000 users and 690 completed subscriptions. Control conversion was 4.0%; treatment conversion was 4.6%. The estimated absolute treatment effect was 0.6 percentage points, or 0.006 as a proportion. The relative lift was 0.6 ÷ 4.0 × 100% = 15%.

These results would support the conclusion that the redesigned process improved completion under the experimental conditions, subject to statistical uncertainty and implementation checks. They would not prove that every mobile customer prefers the new design, nor would they establish long-term profitability without examining downstream outcomes.

Outcome and learning points

Northstar implemented the redesign gradually, monitored early cancellation and service contacts, and continued testing. The case demonstrates several principles:

• Descriptive analytics identified the performance gap.
• Diagnostic analysis generated competing explanations.
• Customer research added evidence unavailable in clickstream data.
• Causal experimentation estimated incremental impact.
• Prescriptive decision making required consideration of security, accessibility, capacity, and long-term value.
• A conversion increase alone was insufficient evidence of overall success.

9. Challenges, Ethics, and Governance

Data quality and integration

Customer records may contain duplicate identities, missing values, inconsistent timestamps, bot traffic, tracking failures, and incompatible definitions. A single customer may appear under multiple identifiers. Identity resolution can improve analysis but also creates privacy and security risks.

Bias and representativeness

Digital behavioural data often represent only people who have access to a channel, consent to tracking, remain observable, and interact in measurable ways. Non-users and silent users may be systematically absent. Models trained on historical decisions may reproduce discrimination present in those decisions.

Analysts should examine performance across relevant groups, test for disparate error rates where appropriate, avoid unnecessary sensitive attributes, and involve domain experts and affected stakeholders. Fairness is context-dependent and cannot be solved by one universal metric.

Privacy and consent

Organizations should collect and use personal data lawfully, fairly, transparently, and for specified purposes. Depending on jurisdiction, obligations may include informed consent, purpose limitation, data minimization, access and correction rights, deletion rights, security safeguards, retention limits, and restrictions on automated decision making. Legal requirements vary, so analysts must consult applicable law and organizational policy.

Privacy-preserving practice includes collecting only necessary data, restricting access, pseudonymizing or anonymizing where appropriate, encrypting sensitive information, documenting processing, and explaining relevant personalization or automated decisions. Anonymization is not automatically permanent or risk-free, especially when datasets can be combined.

Manipulation and dark patterns

Analytics can be used to optimize interfaces in ways that pressure, confuse, or exploit customers. Examples include hidden costs, difficult cancellation, preselected consent, misleading countdowns, and asymmetrical choices. A higher conversion rate achieved through deception is not a customer-centric success.

Artificial intelligence and automated decisions

Machine-learning systems can improve recommendation, forecasting, text analysis, and service routing. They can also amplify biased data, produce unstable outputs, expose confidential information, or create false confidence. Governance should include human oversight for consequential decisions, documentation of model purpose and limitations, monitoring for drift, security controls, escalation paths, and the ability to correct harmful outcomes.

Governance best practices

A mature governance program commonly includes:

• a data dictionary and metric ownership;
• documented data lineage and quality checks;
• role-based access and retention rules;
• privacy and security impact assessments;
• experiment and model review procedures;
• reproducible analysis and version control;
• bias, fairness, and accessibility assessment;
• clear accountability for decisions; and
• ongoing monitoring after deployment.

10. Contemporary Developments and Professional Relevance

Customer and digital analytics are increasingly shaped by privacy-preserving measurement, first-party data strategies, real-time decision systems, conversational interfaces, synthetic or simulated data, customer data platforms, and generative AI. These developments may improve speed and personalization, but they do not remove the need for sound research design and human judgement.

For employability, business analytics graduates should be able to translate a business problem into an analytical question, define a metric precisely, query and combine data responsibly, interpret uncertainty, communicate with nontechnical stakeholders, and challenge misleading conclusions. Technical competence is valuable, but professional credibility depends equally on ethics, curiosity, documentation, and the ability to connect analysis to a decision.

A strong analyst asks five practical questions:

1. What decision will this analysis support?
2. Who is represented in the data, and who is missing?
3. What alternative explanations exist?
4. What evidence would justify claiming an effect rather than an association?
5. What could go wrong if the recommendation is implemented?

Chapter Summary

Customer and digital analytics transform customer, market, and interaction data into evidence for decisions. Customer analytics covers relationships, value, behaviour, experience, retention, and service, while digital analytics focuses particularly on digitally mediated interactions. Their evolution has moved from periodic, aggregated reporting toward integrated, cross-channel, experimental, predictive, and increasingly automated decision support.

Market orientation requires intelligence generation, dissemination, and coordinated response. Customer centricity extends this logic by asking whether decisions create meaningful and fair customer value rather than merely improving isolated metrics. Digital transformation depends on analytics capability, but also on data governance, processes, people, and organizational culture.

The customer lifecycle provides a framework for selecting appropriate questions and metrics from awareness through advocacy, churn, and reactivation. Descriptive analytics explains what happened; diagnostic analytics investigates why; predictive analytics estimates what is likely; causal analytics evaluates effects; and prescriptive analytics recommends actions under objectives and constraints. These forms should be used together, with explicit attention to data quality, uncertainty, privacy, bias, security, and accountability.

The most reliable analytics is decision-centred. It defines the decision, population, measures, assumptions, and evidence requirements before selecting tools. It combines behavioural data with customer research, uses experimentation where causal claims matter, and monitors outcomes after implementation.

Key Terms

Data: Recorded observations or measurements.

Information: Organized data presented in a meaningful context.

Insight: An evidence-based interpretation that informs understanding or action.

Customer analytics: Analysis of customer characteristics, behaviour, value, experience, and relationships.

Digital analytics: Measurement and analysis of digitally mediated interactions and outcomes.

Market orientation: Organizational commitment to generating, sharing, and responding to market intelligence.

Customer centricity: Decision making that places customer value and relationship quality at the centre while balancing organizational and societal responsibilities.

Digital transformation: Organizational change in processes, capabilities, technologies, and business models enabled by digital technologies.

Customer lifecycle: Stages through which a customer relationship develops, such as acquisition, usage, retention, and advocacy.

Touchpoint: A specific interaction between a customer and an organization.

Conversion: A defined desired action.

Segmentation: Division of a customer or market population into meaningful groups.

Personalization: Adaptation of content, offers, or experiences using customer information.

Cohort: A group sharing a defined starting event or characteristic.

Attribution: Assignment of credit for an outcome to contacts or influences.

Descriptive analytics: Analysis of what happened.

Diagnostic analytics: Analysis of why something happened.

Predictive analytics: Estimation of likely future or unobserved outcomes.

Causal analytics: Analysis of whether an intervention or exposure produced an effect.

Prescriptive analytics: Recommendation of actions given objectives, predictions, and constraints.

Review Questions

5 Short Answer Questions

1. Distinguish data, information, insight, and decision.
2. What are the three central activities of market orientation?
3. Why is a dashboard not equivalent to an analytical conclusion?
4. Define a customer lifecycle and name four possible stages.
5. Why does a prediction not necessarily establish causation?

5 Long Answer Questions

1. Explain how customer and digital analytics have evolved from traditional market research and periodic reporting.
2. Compare market orientation and customer centricity. How can analytics support both?
3. Explain the five forms of analytics and show how they can be combined in a business project.
4. Discuss the role of digital transformation in developing organizational analytics capability.
5. Evaluate the major privacy, bias, governance, and accountability issues in customer analytics.

3 Analytical / Application Questions

1. A retailer observes that customers who receive personalized recommendations purchase more than customers who do not. Identify at least three alternative explanations and propose a design for estimating the recommendation’s causal effect.
2. A subscription business has 12,000 customers at the beginning of a quarter, acquires 2,000 new customers, and ends the quarter with 12,600 customers. Calculate the simple retention rate using the formula in this chapter, and explain one limitation of the result.
3. Design a decision-centred analytics project for reducing churn. Specify the decision, lifecycle stage, data sources, descriptive and predictive measures, a possible causal test, and two ethical safeguards.

Multiple Choice Questions with Answers

Question 1

Which statement best describes descriptive analytics?

A. It recommends the optimal action under constraints.

B. It summarizes what has happened or is happening.

C. It proves that one intervention caused an outcome.

D. It estimates the probability of a future event only.

Correct Answer: B. It summarizes what has happened or is happening.

Question 2

A company compares treatment and control groups after randomly assigning customers to receive a new offer. What is the primary purpose of this design?

A. To estimate an incremental causal effect.

B. To create a customer segment.

C. To calculate website traffic.

D. To replace data governance.

Correct Answer: A. To estimate an incremental causal effect.

Question 3

Which activity is most closely associated with market orientation?

A. Generating, disseminating, and responding to market intelligence.

B. Maximizing clicks regardless of customer outcomes.

C. Collecting data without assigning decision responsibility.

D. Using only internal financial data.

Correct Answer: A. Generating, disseminating, and responding to market intelligence.

Question 4

A website has 500 eligible checkout users and 25 purchases. What is the conversion rate?

A. 2%

B. 5%

C. 20%

D. 25%

Correct Answer: B. 5%

Question 5

Which statement about predictive analytics is correct?

A. A predictive score automatically proves why an outcome will occur.

B. Predictive analytics estimates likely outcomes using available evidence.

C. Predictive analytics is limited to financial accounting data.

D. A prediction cannot be evaluated using performance measures.

Correct Answer: B. Predictive analytics estimates likely outcomes using available evidence.

Question 6

Why can last-touch attribution be misleading?

A. It gives equal credit to every interaction.

B. It ignores all digital interactions.

C. It may assign excessive credit to the final interaction in a multistep journey.

D. It can only be used for offline purchases.

Correct Answer: C. It may assign excessive credit to the final interaction in a multistep journey.

Question 7

Which denominator is most appropriate for a clearly defined checkout conversion rate?

A. All employees in the organization.

B. Eligible checkout opportunities during the specified period.

C. All social-media followers from every year.

D. Only customers who completed a purchase.

Correct Answer: B. Eligible checkout opportunities during the specified period.

Question 8

Which practice best supports responsible customer analytics?

A. Collecting every available attribute indefinitely.

B. Removing all human oversight from consequential decisions.

C. Documenting data lineage, access rules, assumptions, and limitations.

D. Optimizing a single engagement metric regardless of harm.

Correct Answer: C. Documenting data lineage, access rules, assumptions, and limitations.

Question 9

What is the main purpose of cohort analysis?

A. To group observations by a shared starting event or characteristic.

B. To guarantee that a result is causal.

C. To eliminate the need for customer research.

D. To ensure every customer receives identical treatment.

Correct Answer: A. To group observations by a shared starting event or characteristic.

Question 10

Which question is prescriptive analytics designed to address?

A. What happened last month?

B. Why did the conversion rate change?

C. What is likely to happen next quarter?

D. What action should be taken given objectives and constraints?

Correct Answer: D. What action should be taken given objectives and constraints?