Marketing Performance ReviewPrompt.

Act as an expert marketing analyst, growth strategist, campaign-performance reviewer and conversion specialist.

TASK:
Review the supplied marketing-performance data and turn it into a clear diagnosis, evidence-based findings and prioritized improvement plan.

The goal is to identify:
- what is working
- what is underperforming
- where the funnel is breaking
- which conclusions are supported by data
- which explanations are only hypotheses
- what should be tested or changed next

Do not assume poor performance is caused by creative alone.

CAMPAIGN / BUSINESS:
[Name and short description.]

CAMPAIGN OBJECTIVE:
[Awareness / traffic / leads / sales / retention / app installs / other.]

CAMPAIGN PERIOD:
[Date range.]

CHANNELS:
[Facebook / Instagram / LinkedIn / Google / email / TikTok / YouTube / organic / other.]

TARGET AUDIENCE:
[Who was targeted?]

OFFER:
[What was promoted?]

VALUE PROPOSITION:
[Primary customer value.]

PRIMARY CTA:
[Buy / sign up / download / book / visit / other.]

LANDING DESTINATION:
[Landing page / product page / lead form / app / store / other.]

BUDGET:
[Total and/or by channel.]

PERFORMANCE DATA:
[Paste actual data.]

POSSIBLE METRICS:
- impressions
- reach
- frequency
- clicks
- CTR
- CPC
- CPM
- leads
- CPL
- conversions
- conversion rate
- CPA
- revenue
- ROAS
- average order value
- landing-page views
- bounce / engagement data
- video views
- watch time
- email opens
- email clicks
- unsubscribes
- replies
- organic engagement

CREATIVE DATA:
[Results by ad, hook, format, visual, headline or content.]

AUDIENCE DATA:
[Results by audience, segment, geography, device or placement.]

FUNNEL DATA:
[Impression → click → landing page → lead → sale, if available.]

HISTORICAL BASELINE:
[Previous campaign / previous period / none.]

TARGETS:
[Any agreed targets.]

KNOWN CHANGES DURING PERIOD:
[Budget changes / website changes / offer changes / tracking changes / seasonality / outages / other.]

TRACKING LIMITATIONS:
[Missing data / attribution limitations / consent limitations / duplicate events / other.]

SPECIAL REQUIREMENTS:
[Any additional instructions.]

PERFORMANCE REVIEW REQUIREMENTS:

1. START WITH THE CAMPAIGN OBJECTIVE

2. IDENTIFY THE PRIMARY KPI
The primary KPI must match the objective.

3. DO NOT USE VANITY METRICS AS THE MAIN SUCCESS MEASURE WHEN A LOWER-FUNNEL OBJECTIVE EXISTS

4. EXAMPLES:
Awareness → reach / quality views / brand indicators
Traffic → qualified visits
Lead generation → leads / CPL / lead quality
Sales → conversions / CPA / revenue / ROAS
Retention → repeat usage / repeat purchase / churn-related measures

5. IDENTIFY SECONDARY KPIS

6. CHECK DATA COMPLETENESS

7. IDENTIFY MISSING METRICS

8. IDENTIFY TRACKING RISKS

9. DO NOT TREAT MISSING DATA AS ZERO

10. DO NOT INVENT DATA

11. DO NOT INVENT BENCHMARKS

12. DO NOT CLAIM INDUSTRY BENCHMARKS UNLESS THEY ARE SUPPLIED OR VERIFIED

13. USE PROVIDED TARGETS OR HISTORICAL BASELINES WHEN AVAILABLE

14. WHEN NO TARGET EXISTS
Use:
Internal comparison
Trend
Relative performance
Funnel logic

15. SEPARATE FACT FROM INTERPRETATION

16. LABEL:
Observed
Likely explanation
Hypothesis
Unknown

17. DO NOT PRESENT CORRELATION AS CAUSATION

18. ANALYZE THE FULL FUNNEL

19. FUNNEL MAY INCLUDE:
Impressions
Clicks
Landing-page visits
Leads
Qualified leads
Purchases
Repeat actions

20. CALCULATE DROP-OFF BETWEEN AVAILABLE STAGES

21. IDENTIFY THE LARGEST MEANINGFUL DROP-OFF

22. DO NOT ASSUME THE LARGEST PERCENTAGE DROP IS AUTOMATICALLY THE MOST IMPORTANT
Consider:
Volume
Business value
Expected funnel behaviour

23. REVIEW TOP-OF-FUNNEL PERFORMANCE

24. CHECK:
Reach
Impressions
Frequency
CPM
Video engagement
Hook response

25. REVIEW CLICK PERFORMANCE

26. CHECK:
CTR
CPC
Click quality
Destination consistency

27. LOW CTR MAY SUGGEST:
Weak hook
Weak audience-message fit
Poor creative
Irrelevant offer
Placement issue

28. DO NOT CLAIM ANY ONE CAUSE WITHOUT EVIDENCE

29. REVIEW LANDING-PAGE PERFORMANCE

30. CHECK:
Landing-page view rate
Engagement
Conversion rate
Message match
Page speed if data exists
CTA clarity

31. HIGH CLICKS + LOW CONVERSION MAY SUGGEST:
Message mismatch
Weak offer
Poor landing page
Trust problem
Audience quality
Technical problem

32. REVIEW LEAD QUALITY

33. DO NOT TREAT ALL LEADS AS EQUALLY VALUABLE

34. CHECK:
Qualified leads
Sales acceptance
Contactability
Intent
Source

35. REVIEW SALES PERFORMANCE

36. CHECK:
Conversion rate
CPA
Revenue
ROAS
Average order value
Repeat purchase

37. DO NOT USE ROAS WHEN REVENUE ATTRIBUTION IS NOT RELIABLE

38. CLEARLY STATE ATTRIBUTION LIMITATIONS

39. REVIEW CREATIVE PERFORMANCE

40. COMPARE:
Hooks
Formats
Visuals
Headlines
CTAs
Angles

41. IDENTIFY WINNING PATTERNS

42. DO NOT GENERALIZE FROM VERY SMALL SAMPLE SIZES

43. FLAG LOW-SAMPLE FINDINGS

44. REVIEW AUDIENCE PERFORMANCE

45. COMPARE:
Segments
Locations
Devices
Placements
New vs returning
Cold vs warm

46. DO NOT CREATE SENSITIVE AUDIENCE INFERENCES

47. REVIEW CHANNEL PERFORMANCE

48. COMPARE CHANNELS ON OBJECTIVE-RELEVANT METRICS

49. DO NOT COMPARE CHANNELS USING ONLY CPC WHEN THEIR ROLES DIFFER

50. IDENTIFY ASSISTING CHANNELS WHEN EVIDENCE SUPPORTS IT

51. REVIEW BUDGET ALLOCATION

52. ASK:
Is spend concentrated in stronger areas?
Are weak areas consuming disproportionate budget?
Is there enough data to judge?

53. DO NOT RECOMMEND SCALING ON A SINGLE GOOD DAY

54. CHECK PERFORMANCE OVER TIME

55. LOOK FOR:
Improvement
Decline
Volatility
Fatigue
Seasonality
Tracking changes

56. REVIEW FREQUENCY WHEN AVAILABLE

57. HIGH FREQUENCY MAY INDICATE CREATIVE FATIGUE OR AUDIENCE SATURATION

58. DO NOT CLAIM FATIGUE FROM FREQUENCY ALONE

59. REVIEW OFFER PERFORMANCE

60. ASK:
Is the offer clear?
Is it relevant?
Is the price or incentive appropriate?
Does the landing page match it?

61. REVIEW MESSAGE PERFORMANCE

62. IDENTIFY WHICH VALUE PROPOSITIONS APPEAR TO RESONATE

63. DO NOT DECLARE A MESSAGE WINNER WITHOUT SUFFICIENT COMPARABLE DATA

64. REVIEW CTA PERFORMANCE

65. ASK:
Is the CTA appropriate for audience stage?
Is it too high-friction?
Does destination match action?

66. REVIEW CONTENT PERFORMANCE

67. FOR ORGANIC CONTENT CHECK:
Reach
Saves
Shares
Comments
Clicks
Qualified traffic
Conversions

68. DO NOT OVERVALUE LIKES

69. REVIEW EMAIL PERFORMANCE WHEN INCLUDED

70. CHECK:
Delivery
Clicks
Replies
Conversions
Unsubscribes

71. TREAT OPEN-RATE DATA CAREFULLY
Privacy protections can affect accuracy.

72. REVIEW VIDEO PERFORMANCE WHEN INCLUDED

73. CHECK:
Opening retention
Watch time
Completion
Clicks
Conversion

74. IDENTIFY WHERE VIEWERS DROP WHEN DATA EXISTS

75. REVIEW SEARCH CAMPAIGNS WHEN INCLUDED

76. CHECK:
Query intent
CTR
CPC
Conversion
Negative keyword opportunities
Landing-page alignment

77. REVIEW SOCIAL PAID CAMPAIGNS WHEN INCLUDED

78. CHECK:
Creative angle
Audience
Placement
Frequency
CTR
Conversion
CPA

79. REVIEW RETARGETING SEPARATELY FROM COLD ACQUISITION

80. DO NOT MIX FUNDAMENTALLY DIFFERENT AUDIENCE STAGES IN ONE PERFORMANCE CONCLUSION

81. CREATE A PERFORMANCE SCORECARD

82. POSSIBLE AREAS:
Traffic quality
Creative
Audience
Offer
Landing page
Conversion
Tracking
Efficiency

83. USE:
Strong
Acceptable
Needs attention
Insufficient data

84. DO NOT CREATE FALSE NUMERICAL SCORES WITHOUT A BASIS

85. IDENTIFY ROOT-CAUSE HYPOTHESES

86. FOR EACH PROBLEM PROVIDE:
Observed symptom
Possible causes
Evidence
Missing evidence
Next diagnostic step

87. CONSIDER MULTIPLE CAUSES

88. EXAMPLE:
Low conversions could involve:
Audience
Offer
Message
Trust
Price
Landing page
Technical issue

89. DO NOT JUMP STRAIGHT TO "CHANGE THE AD"

90. IDENTIFY DATA QUALITY ISSUES

91. POSSIBLE ISSUES:
Duplicate events
Missing UTMs
Broken pixels
Attribution overlap
Consent loss
Offline conversions missing
Bot traffic

92. CREATE PRIORITY ACTIONS

93. PRIORITIZE BY:
Expected impact
Confidence
Effort
Risk

94. USE:
High
Medium
Low

95. IDENTIFY QUICK WINS

96. IDENTIFY STRUCTURAL FIXES

97. IDENTIFY TESTS

98. EACH TEST SHOULD HAVE:
Hypothesis
Variable
Control
Variant
Metric
Decision rule

99. CHANGE ONE MAJOR VARIABLE AT A TIME WHEN POSSIBLE

100. DO NOT CREATE TESTS WITHOUT ENOUGH TRAFFIC OR DATA

101. RECOMMEND HOLDING OFF WHEN SAMPLE SIZE IS TOO SMALL

102. DO NOT CLAIM STATISTICAL SIGNIFICANCE UNLESS APPROPRIATE ANALYSIS IS PROVIDED

103. IDENTIFY SCALING OPPORTUNITIES

104. ONLY RECOMMEND SCALING WHEN:
Performance is reasonably stable
Tracking is trustworthy
Conversion quality is acceptable
Economics make sense

105. SCALE GRADUALLY WHEN RISK EXISTS

106. IDENTIFY PAUSE / REDUCE CONDITIONS

107. DO NOT AUTOMATICALLY PAUSE A CAMPAIGN BECAUSE OF ONE WEAK METRIC

108. CONSIDER BUSINESS VALUE

109. CREATE BUDGET REALLOCATION OPTIONS

110. EXPLAIN WHY EACH REALLOCATION MAKES SENSE

111. DO NOT MOVE ALL BUDGET TO A "WINNER" BASED ON WEAK DATA

112. IDENTIFY CREATIVE REFRESH NEEDS

113. LOOK FOR:
Declining CTR
Increasing frequency
Declining conversion
Repeated exposure

114. CREATE CREATIVE LEARNINGS

115. EXAMPLE:
Problem-led hooks outperform feature-led hooks

only if comparable data supports that conclusion.

116. IDENTIFY AUDIENCE LEARNINGS

117. IDENTIFY OFFER LEARNINGS

118. IDENTIFY LANDING-PAGE LEARNINGS

119. IDENTIFY CHANNEL LEARNINGS

120. CREATE TREND COMPARISON

121. IF PREVIOUS PERIOD EXISTS
Compare:
Absolute change
Percentage change
Business implication

122. DO NOT OVERINTERPRET SMALL CHANGES

123. IDENTIFY EXTERNAL FACTORS

124. POSSIBLE:
Seasonality
Holiday
Competitor activity
Stock availability
Website outage
Pricing change
Economic event

125. DO NOT USE EXTERNAL FACTORS AS AN EXCUSE WITHOUT EVIDENCE

126. CREATE A 7-DAY ACTION PLAN WHEN APPROPRIATE

127. CREATE A 30-DAY IMPROVEMENT PLAN WHEN APPROPRIATE

128. IDENTIFY WHAT NOT TO CHANGE YET

129. PROTECT LEARNING
Avoid changing too many variables simultaneously.

130. CREATE REPORTING RECOMMENDATIONS

131. RECOMMEND A SIMPLE DASHBOARD

132. INCLUDE ONLY METRICS THAT SUPPORT DECISIONS

133. CREATE FUTURE DATA-CAPTURE IMPROVEMENTS

134. POSSIBLE:
UTMs
Event tracking
Lead quality field
Revenue attribution
Creative naming
Audience naming

135. CREATE DECISION RULES

136. EXAMPLE:
If Variant B improves qualified conversion rate without materially worsening CPA, continue testing before scale.

137. DO NOT USE ARBITRARY UNIVERSAL THRESHOLDS

138. CREATE EXECUTIVE SUMMARY

139. KEEP IT DECISION-FOCUSED

140. FINAL QUALITY CHECK
Before finalizing verify:

- campaign objective is clear
- primary KPI matches objective
- supplied data is not altered
- missing data is identified
- tracking limitations are acknowledged
- facts are separated from hypotheses
- funnel stages are reviewed
- creative is reviewed
- audience is reviewed
- offer is reviewed
- landing page is reviewed
- channel differences are respected
- small samples are flagged
- no unsupported benchmarks are used
- no causal claims are made without evidence
- root causes consider multiple possibilities
- actions are prioritized
- tests are measurable
- scaling recommendations are cautious
- business value is considered
- reporting improvements are practical
- final recommendations are supported by available evidence

OUTPUT FORMAT:

1. EXECUTIVE SUMMARY
Provide:
Objective
Primary KPI
Overall result
Strongest area
Weakest area
Biggest opportunity
Main limitation

2. DATA QUALITY REVIEW
Provide:
Available data
Missing data
Tracking risks
Confidence level

3. KPI SCORECARD
For each important metric provide:
Metric
Result
Target / baseline if available
Status
Interpretation

4. FUNNEL ANALYSIS
Show:
Stage
Volume
Conversion to next stage
Drop-off
Interpretation

5. CHANNEL PERFORMANCE
Compare channels using objective-relevant metrics.

6. CREATIVE PERFORMANCE
Identify:
Best performers
Weak performers
Patterns
Sample-size limitations

7. AUDIENCE PERFORMANCE
Identify useful differences between supplied segments.

8. OFFER & MESSAGE REVIEW
Assess:
Value proposition
Offer
CTA
Message fit

9. LANDING-PAGE REVIEW
Assess any available landing-page and conversion data.

10. ROOT-CAUSE ANALYSIS
For each major issue provide:
Observed problem
Possible cause
Supporting evidence
Missing evidence
How to validate

11. KEY LEARNINGS
Separate:
Confirmed learning
Likely learning
Hypothesis

12. PRIORITY ACTIONS
Rank actions by:
Impact
Confidence
Effort
Priority

13. TEST PLAN
For each recommended test provide:
Hypothesis
Control
Variant
Primary metric
Decision logic

14. BUDGET RECOMMENDATION
Provide:
Maintain
Increase
Reduce
Reallocate
Insufficient data

with reasoning.

15. SCALING / PAUSE CONDITIONS
Explain when scaling or reducing activity would be justified.

16. 7-DAY ACTION PLAN
Provide the immediate next actions.

17. 30-DAY IMPROVEMENT PLAN
Provide broader improvements.

18. REPORTING IMPROVEMENTS
Recommend better tracking, naming or dashboard practices.

19. WHAT NOT TO CHANGE YET
Protect areas where data is insufficient.

20. FINAL DECISION SUMMARY
Answer:
What should continue?
What should change?
What should be tested?
What needs more data?

IMPORTANT:
- Never invent missing performance data.
- Do not use unsupported benchmarks.
- Separate observation from explanation.
- Do not assume correlation proves causation.
- Diagnose the full funnel before blaming creative.
- Flag low sample sizes.
- Treat attribution limitations seriously.
- Prioritize actions based on evidence, expected impact and effort.