Skills

Analyse a Real Marketing Campaign's Performance

6 weeks · 0 milestones

Conduct a rigorous retrospective analysis of a real marketing campaign using real performance data — either a campaign you ran, or a real documented campaign with published performance metrics. The analysis must identify what drove results, what failed, and what a future campaign would do differently, grounded in the actual numbers rather than speculation. Proof requires review by someone who has run a real marketing campaign and can confirm the performance attribution is sound and the conclusions non-obvious.

Part of: Business & Management · Proves for: UG

Milestone map

3 milestones

Define the Measurement Framework for a Real Campaign

1–2 weeks

Select a real marketing campaign you have access to performance data for — either your own organisation's campaign or a publicly available campaign with documented results. Define the measurement framework: campaign objective (awareness, consideration, conversion), key performance indicators, attribution model used, and the counterfactual (what baseline are you comparing against).

Proof required

Submit: (a) campaign brief summary (one page) describing the campaign objective, target audience, channels used, and budget (approximate if confidential), (b) measurement framework document specifying the KPIs, attribution model, and baseline for comparison, and (c) confirmation that you have access to the actual performance data for this campaign (access to real data is required — this outcome cannot be satisfied with hypothetical data).

What gets checked

Campaign brief is for a real campaign with real performance data — not a hypothetical
Measurement framework specifies the attribution model explicitly — last click, multi-touch, or data-driven attribution produce very different results
Baseline for comparison is stated — without a counterfactual, campaign performance cannot be assessed

Common mistakes

Using a case study from a textbook rather than a real campaign with real data
Attribution model not specified — 'we tracked conversions' without specifying which touchpoints received credit is not a measurement framework

Powstik Guide

A measurement framework for one real campaign.

Steps

Summarise the campaign on one page.
Set KPIs tied to its objective.
Name the data source for each KPI.
Set a target for each KPI.

Template

Campaign:        Objective:
Audience:        Channels:
| KPI | Target | Data source |
|-----|--------|-------------|

What gets sent back

A hypothetical campaign.
KPIs not linked to the objective.
No data sources.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Verify that the submitter has access to real campaign data — ask them to share a sample data table or screenshot with any sensitive data redacted.
Check the attribution model specification: can the submitter explain what touchpoints receive credit and why that model was chosen over alternatives?

You'll sign in first, then come straight back here.

Analyse Campaign Performance and Produce Findings

2–3 weeks

Conduct the campaign performance analysis: compute all specified KPIs, compare against baseline, segment performance by channel and audience where data permits, and produce findings on what drove performance. The analysis must go beyond headline metrics to explain why performance was what it was.

Proof required

Submit a campaign performance analysis (minimum 1,500 words) covering: (a) KPI performance table showing actuals vs. targets for all specified metrics, (b) channel and audience segment performance comparison (minimum three segments), (c) key drivers analysis explaining the top two or three factors that drove performance above or below target, and (d) data limitations section acknowledging measurement gaps and their effect on conclusions.

What gets checked

KPI table shows actuals vs targets — not just actuals
Driver analysis explains causation, not just correlation — 'social media drove 40% of conversions' needs an explanation of why
Data limitations are specific — 'we cannot distinguish brand vs performance attribution in the display channel because view-through windows are uncertain'

Common mistakes

Analysis reports headline metrics without segmentation — a campaign that performed at target on average may have significantly over- and under-performed by segment
Driver analysis assigns causation without evidence — correlation between channel exposure and conversion does not confirm the channel caused the conversion

Powstik Guide

A 1,500-word campaign analysis with a KPI table.

Steps

Fill actual vs target for each KPI.
Explain the biggest gaps.
Compare channels on cost per result.
Recommend what to change next time.

Template

| KPI | Target | Actual | Gap | Why |
|-----|--------|--------|-----|-----|
Channel comparison:
Recommendations:

What gets sent back

Under 1500 words.
Claims without sources.
No clear recommendation.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Challenge the driver analysis: 'You attributed the above-target conversion rate to the email channel — but email and social ran simultaneously. How do you know email drove the lift rather than social?'
Check the data limitations section: is it specific to this campaign's data, or generic?

You'll sign in first, then come straight back here.

Present Analysis and Recommendations to a Marketing Analytics Professional

1 week

Present the campaign performance analysis and recommendations for future campaigns to a marketing analytics professional (media analyst, marketing strategist, or digital marketing manager) who will challenge your attribution assumptions or driver conclusions.

Proof required

Submit: (a) a recommendations document (minimum 500 words) for the next campaign based on your analysis, (b) a Q&A log from the review session (minimum 300 words with at least two challenges and your responses), and (c) attendance record with reviewer name, credentials, and date.

What gets checked

Recommendations are specific and grounded in the analysis — not generic marketing best practices
Q&A log shows the reviewer challenged attribution or causal reasoning
Reviewer has marketing analytics expertise

Common mistakes

Recommendations are general ('invest more in social media') without specific targets or rationale from the data
Review session with someone who approves everything without challenge

Powstik Guide

Next-campaign recommendations, tested by a marketing analyst.

Steps

Turn each finding into one recommendation with a target number.
State the attribution model you used and its blind spot.
Book a media analyst, strategist or digital marketing manager.
Ask them to attack your attribution and causal claims.
Log the session and adjust the recommendations.

Template

Recommendation table
| Finding (with number) | Recommendation | Target | How we will measure |
|-----------------------|----------------|--------|---------------------|

Review session log
Date:            Duration:
Reviewer name:
Reviewer role and years of experience:

Challenge 1 (attribution or causality):
  What they asked:
  My answer:
  Did it hold? What would I change?

Challenge 2:
  What they asked:
  My answer:
  Did it hold? What would I change?

Change I made after the session:
Reviewer's overall view (their words):

What gets sent back

Generic advice like 'invest more in social'.
No challenge to the attribution in the log.
A reviewer without marketing analytics experience.

Resources

Foundationstart here

Depthgo deeper

Masteryfor the dedicated

What a verifier looks for

Challenge the recommendations: 'You're recommending increasing the email budget — but your own analysis showed email had the highest unsubscribe rate during this campaign. How do you reconcile that?'
Provide a written confirmation (minimum 150 words) of your marketing analytics credentials and the challenges you raised.

You'll sign in first, then come straight back here.

Part of

Prove Your Marketing Management Chops →

Related outcomes

Milestone map

3 milestones

Define the Measurement Framework for a Real Campaign

1–2 weeks

Proof required

What gets checked

Campaign brief is for a real campaign with real performance data — not a hypothetical
Measurement framework specifies the attribution model explicitly — last click, multi-touch, or data-driven attribution produce very different results
Baseline for comparison is stated — without a counterfactual, campaign performance cannot be assessed

Common mistakes

Using a case study from a textbook rather than a real campaign with real data
Attribution model not specified — 'we tracked conversions' without specifying which touchpoints received credit is not a measurement framework

Powstik Guide

A measurement framework for one real campaign.

Steps

Summarise the campaign on one page.
Set KPIs tied to its objective.
Name the data source for each KPI.
Set a target for each KPI.

Template

Campaign:        Objective:
Audience:        Channels:
| KPI | Target | Data source |
|-----|--------|-------------|

What gets sent back

A hypothetical campaign.
KPIs not linked to the objective.
No data sources.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Verify that the submitter has access to real campaign data — ask them to share a sample data table or screenshot with any sensitive data redacted.
Check the attribution model specification: can the submitter explain what touchpoints receive credit and why that model was chosen over alternatives?

You'll sign in first, then come straight back here.

Analyse Campaign Performance and Produce Findings

2–3 weeks

Proof required

What gets checked

KPI table shows actuals vs targets — not just actuals
Driver analysis explains causation, not just correlation — 'social media drove 40% of conversions' needs an explanation of why
Data limitations are specific — 'we cannot distinguish brand vs performance attribution in the display channel because view-through windows are uncertain'

Common mistakes

Analysis reports headline metrics without segmentation — a campaign that performed at target on average may have significantly over- and under-performed by segment
Driver analysis assigns causation without evidence — correlation between channel exposure and conversion does not confirm the channel caused the conversion

Powstik Guide

A 1,500-word campaign analysis with a KPI table.

Steps

Fill actual vs target for each KPI.
Explain the biggest gaps.
Compare channels on cost per result.
Recommend what to change next time.

Template

| KPI | Target | Actual | Gap | Why |
|-----|--------|--------|-----|-----|
Channel comparison:
Recommendations:

What gets sent back

Under 1500 words.
Claims without sources.
No clear recommendation.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Challenge the driver analysis: 'You attributed the above-target conversion rate to the email channel — but email and social ran simultaneously. How do you know email drove the lift rather than social?'
Check the data limitations section: is it specific to this campaign's data, or generic?

You'll sign in first, then come straight back here.

Present Analysis and Recommendations to a Marketing Analytics Professional

1 week

Proof required

What gets checked

Recommendations are specific and grounded in the analysis — not generic marketing best practices
Q&A log shows the reviewer challenged attribution or causal reasoning
Reviewer has marketing analytics expertise

Common mistakes

Recommendations are general ('invest more in social media') without specific targets or rationale from the data
Review session with someone who approves everything without challenge

Powstik Guide

Next-campaign recommendations, tested by a marketing analyst.

Steps

Turn each finding into one recommendation with a target number.
State the attribution model you used and its blind spot.
Book a media analyst, strategist or digital marketing manager.
Ask them to attack your attribution and causal claims.
Log the session and adjust the recommendations.

Template

Recommendation table
| Finding (with number) | Recommendation | Target | How we will measure |
|-----------------------|----------------|--------|---------------------|

Review session log
Date:            Duration:
Reviewer name:
Reviewer role and years of experience:

Challenge 1 (attribution or causality):
  What they asked:
  My answer:
  Did it hold? What would I change?

Challenge 2:
  What they asked:
  My answer:
  Did it hold? What would I change?

Change I made after the session:
Reviewer's overall view (their words):

What gets sent back

Generic advice like 'invest more in social'.
No challenge to the attribution in the log.
A reviewer without marketing analytics experience.

Resources

Foundationstart here

Depthgo deeper

Masteryfor the dedicated

What a verifier looks for

Challenge the recommendations: 'You're recommending increasing the email budget — but your own analysis showed email had the highest unsubscribe rate during this campaign. How do you reconcile that?'
Provide a written confirmation (minimum 150 words) of your marketing analytics credentials and the challenges you raised.

You'll sign in first, then come straight back here.