12. Creative Effectiveness, Testing and Analytics

Introduction and Context

Creative strategy is often judged by the visible quality of an advertisement: its idea, writing, design, emotional force, cultural relevance, or distinctiveness. Yet professional advertising cannot rely on aesthetic or intuitive judgement alone. An advertisement must also perform a communicative and commercial task. It may need to make a brand memorable, change a perception, stimulate consideration, generate a response, strengthen loyalty, or contribute to long-term brand value. Creative effectiveness is therefore the disciplined study of whether communication produces meaningful outcomes, for whom, under what conditions, and at what cost.

Testing and analytics provide evidence for making these judgements. Pre-testing can identify misunderstandings before substantial media expenditure. Experiments can establish whether exposure to advertising causes a change in behaviour. Brand-lift research can measure changes in awareness, associations, preference, or intent. Digital analytics can reveal attention, engagement, conversion, and customer journeys. Attribution and incrementality methods help determine the contribution of advertising rather than merely recording activity that occurred after exposure.

The central challenge is that advertising operates across different time horizons. A short-term click may be easy to count but strategically unimportant. A distinctive brand asset may produce few immediate conversions while improving future recognition and reducing the cost of later persuasion. Conversely, a campaign may create impressive reach and engagement without influencing brand choice. Effective evaluation therefore requires both numerical discipline and strategic judgement. Metrics should serve the creative and business objectives; they should not replace them.

Learning Objectives

After studying this chapter, students should be able to:

1. Define creative effectiveness and distinguish objectives, outputs, outcomes, and business results.

2. Select appropriate key performance indicators (KPIs) for different advertising objectives.

3. Explain pre-testing, experiments, A/B testing, brand-lift studies, attention measures, conversion analytics, attribution, and incrementality.

4. Interpret data while recognising validity, reliability, bias, causality, and statistical uncertainty.

5. Connect short-term response measures with long-term brand-building outcomes.

6. Design an integrated measurement plan for a campaign.

7. Identify ethical, privacy, organisational, and professional risks in advertising analytics.

From Creative Idea to Measurable Effect

Creative effectiveness begins with a chain of reasoning. A campaign has a business problem, a communication objective, a target audience, a creative proposition, a media environment, and an expected response. Measurement should reflect this chain.

An output is what the campaign delivers: impressions, reach, completed views, or published assets. An audience response is what people notice, understand, feel, remember, or do. A business outcome may be sales, qualified leads, retention, market share, or profitability. Outputs are necessary but not sufficient evidence of effectiveness. Ten million impressions demonstrate distribution, not persuasion.

A useful hierarchy is:

• Exposure: Was the audience technically served the message?
• Attention: Did people notice and actively process it?
• Comprehension: Did they understand the intended meaning?
• Memory: Was the brand or message encoded and later retrieved?
• Attitude: Did perceptions, associations, or preference change?
• Behaviour: Did people search, visit, enquire, buy, subscribe, or recommend?
• Business value: Did the activity create profitable and sustainable value?

The hierarchy is not strictly linear. People can buy with little conscious memory, and a memorable advertisement may not generate immediate action. Nevertheless, it prevents the common error of treating every measurable event as equivalent.

Objectives, Metrics, and Key Performance Indicators

A KPI is a selected measure used to judge progress toward an important objective. It is not simply any available number. A good KPI is relevant to the decision, sufficiently reliable, interpretable, and linked to a defined time horizon.

For a new product, the objective may be awareness and mental availability. Suitable measures can include aided and unaided awareness, brand recognition, category-entry-point association, reach among the target audience, and search behaviour. For a performance campaign, the objective may be qualified acquisition. Measures might include conversion rate, cost per qualified lead, completed application, revenue, contribution margin, or customer lifetime value. For a retention programme, repeat purchase, churn, usage frequency, and advocacy may be more appropriate.

Common formulas include:

Click-through rate (CTR) = clicks ÷ impressions × 100

Conversion rate = conversions ÷ relevant visits or clicks × 100

Cost per acquisition (CPA) = advertising spend ÷ attributed acquisitions

Return on advertising spend (ROAS) = attributed revenue ÷ advertising spend

These measures are useful, but each depends on definitions. A cheap acquisition may have low quality or high cancellation. A high ROAS may result from customers who would have purchased without advertising. A high CTR may reflect curiosity, accidental clicks, or an offer that attracts people outside the intended audience. Sound analysis therefore examines the numerator, denominator, data source, time window, and counterfactual behind every metric.

A Measurement Architecture for Creative Strategy

A campaign measurement plan should be developed before launch rather than assembled after results appear. It normally contains five connected decisions.

First, define the strategic question. Examples include: Can the new platform make the brand more distinctive? Does the revised message increase consideration? Does retargeting add incremental conversions? Which execution communicates the benefit most clearly?

Second, define the success criteria. These may include a minimum lift in aided awareness, a reduction in cost per qualified lead, or a statistically credible increase in sales relative to a control group.

Third, identify the unit of analysis. This might be an individual, household, store, region, campaign, impression, session, or customer cohort. Confusing units can produce false conclusions; for example, campaign-level averages cannot automatically explain individual behaviour.

Fourth, specify the comparison or counterfactual. Measurement is strongest when the advertiser can compare what happened with advertising to what would have happened without it. This is the foundation of causal inference.

Fifth, establish the decision rule. Results should lead to an action: continue, revise, scale, pause, investigate, or retain for long-term brand reasons despite weak immediate response.

A practical dashboard can be organised into three layers. The first records delivery and quality of exposure. The second measures communication effects such as attention, memory, and brand associations. The third measures behavioural and financial outcomes. The layers should be read together rather than collapsed into a single score.

Pre-Testing: Improving Work Before It Enters the Market

Pre-testing evaluates advertising before or during limited release. Its purpose is not to predict the market with perfect accuracy. Rather, it reduces avoidable failure and helps the team understand how audiences interpret the work.

Qualitative methods include individual interviews, focus groups, accompanied browsing, projective exercises, and think-aloud tasks. They can reveal whether the consumer notices the brand, interprets the proposition correctly, finds the tone credible, or experiences an unintended cultural meaning. Qualitative research is particularly valuable for diagnosing why a communication succeeds or fails. It is less suitable for estimating the exact size of an effect in the population.

Quantitative pre-testing may use structured surveys, monadic exposure designs, forced or spontaneous recall, recognition, message takeout, brand linkage, purchase intention, and comparative evaluations. In a monadic design, each respondent evaluates one execution rather than comparing several advertisements directly. This reduces artificial preference effects caused by side-by-side comparison.

Implicit and behavioural measures may include response latency, eye tracking, facial expression coding, scroll behaviour, or simulated choice. These can supplement stated responses, but they should not be treated as direct windows into the mind. Eye movement indicates visual attention, not necessarily meaningful comprehension; a facial reaction does not automatically indicate persuasion.

Pre-testing should examine at least four questions:

1. Attention: Is the advertisement noticed in its intended environment?

2. Brand linkage: Is the correct brand connected to the message?

3. Communication: Is the proposition understood and believable?

4. Effect: Does the work produce the intended emotional, attitudinal, or behavioural response?

A common professional mistake is to use pre-testing as a popularity contest. Strong creative work may be unfamiliar, challenging, or initially polarising. The task is not to remove all risk or make an advertisement universally liked. It is to determine whether the risk is strategically productive and whether the central idea survives contact with the audience.

Experiments and A/B Testing

An experiment changes one or more conditions and compares outcomes with a control condition. Random assignment is important because it distributes known and unknown differences between groups, making causal interpretation more credible.

In a simple A/B test, audience members are randomly exposed to version A or version B. The versions might differ in headline, visual, call to action, landing page, offer, or complete creative concept. The outcome could be qualified leads, completed purchases, brand search, or another pre-specified measure.

A valid test requires more than observing that B received more conversions. The groups should be comparable, exposure should be sufficiently controlled, the sample should be large enough for the expected effect, and the test should run for an appropriate period. Analysts should avoid repeatedly checking results and stopping the test when a favourable difference appears. This practice increases false-positive risk.

The absolute difference is the difference between the two observed rates. The relative lift is the difference divided by the control rate. If version A converts at 4% and version B at 5%, the absolute lift is one percentage point and the relative lift is 25%. Both descriptions are accurate but communicate different impressions.

Statistical significance indicates that the observed result would be relatively unlikely under a specified null hypothesis. It does not establish strategic importance, profitability, or universal applicability. Confidence intervals, effect size, sample quality, and practical consequences should also be considered. A tiny improvement may be statistically reliable but commercially irrelevant; a meaningful effect may require more data before it becomes statistically conclusive.

A/B testing is powerful for optimisation but has limits. It often tests isolated elements in a particular channel and short time period. It can favour direct-response executions that produce immediate action while underestimating ideas that build memory. It may also produce interaction effects: a headline that performs well with one image may perform poorly with another. Multivariate testing can examine combinations, but it requires much larger samples and careful experimental design.

Brand Lift and the Measurement of Mental Availability

Brand-lift studies estimate whether advertising changes brand-related attitudes or knowledge. They commonly compare people who were exposed to advertising with a control group that was not, using survey responses collected before and after exposure or at a single post-exposure point with appropriate controls.

Measures may include unaided awareness, aided awareness, ad recall, message association, brand consideration, perceived relevance, trust, preference, and purchase intent. The selection should follow the communication objective. If the campaign is intended to establish a distinctive association, measuring only purchase intent is inadequate. If the campaign is designed to drive immediate sign-ups, brand lift should be complemented by behavioural data.

Brand-lift results are vulnerable to contamination. Control respondents may encounter the advertisement elsewhere, exposed respondents may not actually notice it, and self-reported exposure may be inaccurate. Survey wording, sample composition, timing, and social desirability can also influence findings. For these reasons, brand lift is best treated as one component of evidence rather than a definitive verdict.

Long-term brand effects are especially difficult to observe because they accumulate gradually and interact with distribution, pricing, competition, product experience, and other communications. Consistent measurement of distinctive assets, category associations, mental availability, and future consideration can provide a more useful longitudinal view than a single campaign score.

Attention, Engagement, and the Problem of Proxy Metrics

Attention is a scarce resource and an important intermediate condition for advertising effectiveness. However, exposure is not attention, and attention is not persuasion. A viewable impression may be technically present on screen without being consciously processed. A completed video may be autoplayed or watched while the viewer is distracted.

Attention can be estimated through active time, viewability, completion, scroll depth, screen position, device signals, eye tracking, biometric measures, or controlled self-report. Each method captures a different dimension. Active time may indicate opportunity to process; eye tracking indicates gaze allocation; recall indicates later retrieval. No single measure is a complete representation of attention.

Engagement is similarly broad. Likes, comments, shares, saves, dwell time, and video completion may indicate involvement, but they can also reflect controversy, entertainment unrelated to the brand, or platform behaviour. A provocative advertisement can generate conversation while damaging trust. Engagement should therefore be interpreted in relation to sentiment, brand linkage, audience quality, and downstream outcomes.

Creative teams should ask: What kind of attention does the idea require? Is the brand visible early enough to receive credit? Does the execution make the intended claim easy to understand? Does the emotional response support the brand’s positioning? These questions connect analytics to craft rather than treating metrics as independent from creative judgement.

Conversion Analytics, Attribution, and Incrementality

Conversion analytics tracks actions such as visits, registrations, downloads, purchases, or applications. It depends on reliable event definitions, tagging, identity resolution, and consistent treatment of duplicate or invalid events. A conversion is not automatically a valuable conversion. Quality, margin, retention, and customer lifetime value may matter more than volume.

Attribution assigns credit for an outcome to one or more touchpoints. Common rules include last-click, first-click, linear, time-decay, and position-based models. Platform-reported attribution can be useful for operational optimisation but often reflects the platform’s own view of exposure and may overstate its contribution. Last-click attribution, for example, tends to reward the final interaction while ignoring earlier brand-building contact.

Attribution answers a descriptive question: which recorded touchpoints were associated with a conversion? Incrementality asks a causal question: how many additional outcomes occurred because of the advertising? The difference can be substantial. A person who searches for a brand, clicks a paid search advertisement, and purchases may already have been intending to buy. Attribution credits the search click; an incrementality test may find little additional effect.

Incrementality can be estimated through user-level holdouts, geo-experiments, matched-market tests, conversion-lift studies, and carefully designed time-series models. A holdout group does not receive the treatment, while the treatment group does. A geo-test compares selected regions, ideally with pre-period balancing and controls for external shocks. These methods require operational discipline: exclusion must be genuine, spillover must be assessed, and the test must run long enough to capture delayed effects.

The distinction is strategically important. Attribution is often useful for allocating effort within a campaign. Incrementality is more appropriate for deciding whether a channel, audience, or tactic creates additional demand. Neither method alone captures all long-term brand effects, so short-term experiments should be combined with brand and market indicators.

Interpreting Results Without Destroying Creativity

Analytics can improve creative work when it clarifies a problem. If people remember the advertisement but not the brand, the issue may be brand linkage. If they notice the brand but misunderstand the proposition, the problem may be copy clarity. If an execution performs well with existing customers but poorly with new prospects, the targeting or entry point may need revision.

Analytics becomes harmful when teams optimise only what is easiest to count. Excessive focus on CTR can encourage sensational headlines. Overreliance on immediate CPA can eliminate upper-funnel activity that creates future demand. Treating average results as universal can conceal differences by audience, context, device, culture, or customer value.

A balanced interpretation considers:

• the original objective and time horizon;
• the quality and representativeness of the data;
• the size and uncertainty of the effect;
• causal evidence versus simple association;
• short-term response and possible delayed effects;
• brand distinctiveness and consistency;
• commercial value, not merely activity volume;
• ethical and privacy implications.

Creative judgement remains essential because data describes observed behaviour under particular conditions. It does not automatically explain meaning, cultural interpretation, future potential, or strategic fit. The strongest organisations create a dialogue between creative, media, research, data, finance, and client teams. No department should own effectiveness in isolation.

An Illustrative Case: The “Fresh Start” Campaign

Background

The following is a fictional illustrative case. Fresh Start is a regional subscription-based meal service seeking younger urban customers. Its previous advertising emphasised discounts and generated efficient first purchases, but repeat orders were weak. The brand wanted to become associated with convenient, healthy weekday routines rather than cheap trial offers.

Situation and Problem

The creative team developed two routes. Route A used a direct price message: “Dinner sorted for less.” Route B used a more distinctive narrative showing a person moving from a chaotic workday to a calm evening, ending with the line: “Make room for your evening.” The media team predicted that Route A would generate a higher immediate conversion rate, while the brand team believed Route B could build stronger associations.

Measurement Design

The campaign objective was divided into two horizons. The short-term objective was qualified first orders, defined as a completed order that was not cancelled within the initial period. The longer-term objective was association with easy weekday routines, consideration among non-customers, and repeat purchase.

Before launch, both routes were tested using separate respondent groups. Route A produced stronger stated deal appeal. Route B produced higher intended association with “making evenings easier” and stronger brand distinctiveness, but some participants initially found the ending less explicit. The team revised the final frame to improve brand linkage without removing the narrative idea.

The live campaign used a geo-experiment. Comparable areas were assigned to treatment or control, with treatment areas receiving the new campaign and control areas receiving existing low-intensity activity. A/B testing within treatment areas compared two final-frame copy versions. Brand-lift surveys measured consideration and the weekday-routine association. Sales data tracked qualified first orders and repeat orders.

Analysis and Outcome

During the first weeks, the price-led control generated a higher attributed conversion rate. However, the geo-experiment showed that the difference in total qualified orders was smaller than platform attribution suggested. This indicated that some conversions credited to the price-led advertising would probably have occurred without it.

The narrative campaign produced a modest but credible increase in the intended brand association and consideration. Its immediate acquisition cost was higher, but customers acquired in treatment areas displayed stronger early repeat-order behaviour. The revised final frame improved brand linkage without materially reducing attention. The company therefore retained the narrative platform, used a clearer response mechanism in selected placements, and continued measurement over a longer period.

Learning Points

The case illustrates five principles. First, a campaign can have more than one legitimate KPI when objectives operate over different time horizons. Second, platform attribution may overstate contribution. Third, pre-testing is valuable when it diagnoses a specific issue, such as weak brand linkage, rather than rewarding the most popular execution. Fourth, creative effectiveness is not equivalent to immediate conversion efficiency. Finally, the best decision may be a portfolio approach: brand-building narrative supported by appropriately designed response assets.

Risks, Limitations, and Governance

Measurement systems can fail through poor data quality, inconsistent definitions, sample bias, missing conversions, duplicated events, bot activity, or changes in platform algorithms. Organisations should document taxonomy, tracking rules, ownership, data lineage, and revision history. A dashboard containing precise numbers can still be conceptually weak.

Privacy is a central professional responsibility. Personalisation, identity matching, location data, behavioural tracking, and customer lists should be collected and used lawfully, transparently, proportionately, and with appropriate consent where required. Data minimisation, access controls, retention limits, security, and responsible vendor management are essential. Sensitive inferences should not be used to exploit vulnerable audiences or exclude people unfairly.

AI-based optimisation introduces additional risks. Automated systems may reproduce historical bias, favour high-volume audiences, generate misleading creative variations, or optimise toward a proxy that conflicts with brand and social objectives. Human review is required for claims, cultural representation, targeting, privacy, and brand safety. Experimentation should not expose consumers to materially harmful or deceptive experiences merely to obtain a statistically interesting result.

Common mistakes include changing several variables at once, testing without a control, stopping early, ignoring seasonality, interpreting correlation as causation, selecting only favourable metrics, and reporting averages without audience segments. Another error is post-rationalisation: deciding what the campaign objective was after seeing which metric performed best. Pre-registration of hypotheses and decision rules can reduce this problem.

Emerging Directions in Effectiveness Practice

Advertising measurement is moving toward privacy-conscious, outcome-oriented systems. As individual-level tracking becomes less reliable or less acceptable, advertisers are combining consented first-party data with contextual signals, media mix modelling, geo-experiments, clean-room analysis, and aggregated conversion measurement. These approaches are not interchangeable. Aggregated models can estimate broad contribution but may lack granular diagnostic power; experiments provide stronger causal evidence but can be costly and operationally difficult.

Attention research is becoming more sophisticated, particularly in connected television, digital video, retail media, and mobile environments. The important development is not the creation of one universal attention score, but a better understanding of how format, context, creative structure, and exposure quality interact.

Generative AI is also changing testing workflows. It can produce copy variants, classify comments, summarise qualitative responses, and identify patterns in large datasets. Its outputs require human validation because fluent language is not evidence of strategic or ethical suitability. The future professional will need both creative literacy and measurement literacy: the ability to ask meaningful questions of data and to protect the qualities of advertising that numbers do not fully capture.

Best Practices for Professional Use

A robust effectiveness practice follows several principles:

1. Begin with the communication and business objective, not the available dashboard.

2. Select a small number of decision-relevant KPIs and define them precisely.

3. Combine pre-testing, live experimentation, brand measures, behavioural data, and financial outcomes.

4. Distinguish exposure, attention, engagement, attribution, and incrementality.

5. Use control groups or credible counterfactuals whenever causal claims are made.

6. Report effect size, uncertainty, sample limitations, and practical significance.

7. Protect long-term brand-building work from short-term metric bias.

8. Review results across audiences, contexts, time periods, and customer quality segments.

9. Treat privacy, fairness, transparency, and security as part of effectiveness practice.

10. Convert findings into creative learning rather than merely ranking advertisements.

Chapter Summary

Creative effectiveness is the relationship between advertising activity and meaningful communication or business outcomes. Evaluation should proceed from objectives to appropriate KPIs and should distinguish outputs from outcomes. Pre-testing diagnoses problems before launch; experiments and A/B tests compare alternatives; brand-lift research examines changes in mental and attitudinal responses; attention and engagement measures assess intermediate processes; conversion analytics records actions; attribution assigns descriptive credit; and incrementality estimates additional effects caused by advertising.

No single metric captures effectiveness. Strong interpretation considers causality, uncertainty, data quality, time horizon, brand distinctiveness, customer value, and ethical responsibility. Short-term conversion metrics are useful but can undervalue long-term brand-building creativity. The most reliable practice combines evidence from controlled tests, brand measures, behavioural outcomes, and strategic judgement. Analytics should make creative decisions more intelligent, not make creativity subordinate to whatever is easiest to count.

Key Terms

Creative effectiveness: The extent to which advertising creative produces intended communication, behavioural, or business outcomes.

KPI: A selected indicator used to evaluate progress toward an important objective.

Output: A delivered campaign activity, such as impressions, reach, or completed views.

Outcome: A response or result produced among audiences or in the market.

Pre-testing: Evaluation of advertising before or during limited release to diagnose likely strengths and weaknesses.

A/B testing: A controlled comparison between two versions of an advertising variable or execution.

Control group: A group not receiving the tested treatment, used as a comparison condition.

Brand lift: A measured change in awareness, associations, consideration, preference, or related brand responses.

Attention: The degree to which an audience notices and processes an advertisement.

Engagement: Observable or reported involvement with advertising, such as commenting, sharing, saving, or active viewing.

Conversion: A defined desired action, such as purchase, registration, enquiry, or application.

Attribution: Assignment of credit for an outcome to one or more recorded touchpoints.

Incrementality: The additional outcome caused by advertising compared with what would have occurred without it.

Counterfactual: The estimated outcome that would have occurred in the absence of the intervention.

Statistical significance: Evidence that an observed result is unlikely under a specified null hypothesis.

Effect size: The magnitude of a difference or relationship, independent of whether it is statistically significant.

ROAS: Revenue attributed to advertising divided by advertising spend.

CPA: Advertising spend divided by attributed acquisitions.

Mental availability: The probability that a brand comes to mind in relevant buying situations.

Brand linkage: The extent to which an audience connects an advertisement’s message to the correct brand.

Review Questions

Short Answer Questions

1. Distinguish between campaign outputs, audience responses, and business outcomes.

2. Why is a KPI more than simply an available performance number?

3. What is the main purpose of advertising pre-testing?

4. Explain the difference between attribution and incrementality.

5. Why should attention and engagement not be treated as identical measures?

Long Answer Questions

1. Explain how a campaign measurement architecture can connect business objectives, communication effects, behavioural outcomes, and financial results.

2. Discuss the strengths and limitations of qualitative and quantitative pre-testing methods.

3. Explain how A/B testing should be designed and interpreted, including the importance of randomisation, effect size, and uncertainty.

4. Analyse why short-term conversion metrics can undervalue long-term brand-building creativity.

5. Discuss privacy, fairness, transparency, and governance issues in contemporary advertising analytics.

Analytical / Application Questions

1. An advertisement has a high completion rate and strong social sharing but low brand recall and no measurable sales lift. Diagnose at least three possible explanations and recommend further tests.

2. A paid-search platform reports a high ROAS, but a geo-experiment shows little difference between exposed and control regions. Explain the discrepancy and advise the advertiser.

3. Design an integrated measurement plan for a new brand campaign whose objectives are awareness, distinctive association, qualified acquisition, and repeat purchase. Specify methods, KPIs, controls, and decision rules.

22. Multiple Choice Questions with Answers

Question 1

Which measure most directly addresses whether advertising created additional conversions beyond those that would have occurred naturally?

A. Last-click attribution

B. Incrementality

C. Click-through rate

D. Ad recall

Correct Answer: B. Incrementality

Question 2

A campaign receives ten million impressions. What is the most accurate interpretation of this result?

A. The campaign persuaded ten million people

B. The campaign generated ten million purchases

C. The campaign delivered a measured volume of exposure opportunities

D. The campaign created ten million incremental conversions

Correct Answer: C. The campaign delivered a measured volume of exposure opportunities

Question 3

Why might a monadic pre-test be preferred when comparing advertising executions?

A. It guarantees market sales

B. It prevents respondents from evaluating any advertising

C. It reduces artificial preferences caused by direct side-by-side comparison

D. It eliminates the need for a control group

Correct Answer: C. It reduces artificial preferences caused by direct side-by-side comparison

Question 4

Version A has a conversion rate of 4% and version B has a conversion rate of 5%. What is the relative lift of B over A?

A. 1%

B. 5%

C. 20%

D. 25%

Correct Answer: D. 25%

Question 5

Which finding most strongly indicates a brand-linkage problem?

A. The advertisement is noticed and the brand is correctly remembered, but sales are delayed

B. The advertisement is remembered, but respondents cannot identify which brand made it

C. The advertisement is not served to the intended audience

D. The advertisement produces qualified repeat purchases

Correct Answer: B. The advertisement is remembered, but respondents cannot identify which brand made it

Question 6

What is the principal weakness of last-click attribution?

A. It cannot record any digital event

B. It always produces statistically insignificant results

C. It tends to credit the final interaction while ignoring earlier influences

D. It measures only brand awareness

Correct Answer: C. It tends to credit the final interaction while ignoring earlier influences

Question 7

A statistically significant result is found in an A/B test. Which conclusion is still not justified automatically?

A. The observed difference is unlikely under the specified null hypothesis

B. The difference should be examined for effect size

C. The result is necessarily strategically important and profitable

D. The test conditions should be considered before generalising

Correct Answer: C. The result is necessarily strategically important and profitable

Question 8

Which combination best reflects a balanced effectiveness dashboard?

A. Impressions only

B. CTR only

C. Delivery, communication effects, behavioural outcomes, and business value

D. Social shares only

Correct Answer: C. Delivery, communication effects, behavioural outcomes, and business value

Question 9

Why can a high ROAS fail to demonstrate advertising incrementality?

A. Revenue can never be measured

B. Some credited purchasers may have bought without the advertising

C. ROAS measures only awareness

D. Incrementality applies only to television

Correct Answer: B. Some credited purchasers may have bought without the advertising

Question 10

What is the most appropriate role of analytics in creative strategy?

A. To replace all creative judgement with automated rankings

B. To optimise only the cheapest immediate response

C. To diagnose performance and inform decisions while respecting long-term brand objectives

D. To remove all distinctive or potentially polarising ideas

Correct Answer: C. To diagnose performance and inform decisions while respecting long-term brand objectives