The Best ChatGPT Prompts for Campaign Performance Reports

Analyzing campaign performance represents one of the most critical and time-consuming activities in the arsenal of a digital marketing professional. Data collection, interpretation, coherent storytelling of results, and the formulation of strategic recommendations require cross-functional skills and hours of concentrated work. ChatGPT has profoundly transformed this process, enabling marketing teams to accelerate report production without sacrificing analytical quality or strategic depth.

ChatGPT's ability to process large quantities of structured data, identify narrative patterns, and translate complex metrics into accessible language for diverse stakeholders makes it a particularly suitable tool for this use case. Unlike other automation tools, ChatGPT can adapt the tone, level of technical detail, and report structure based on the target audience, whether it's the board of directors, the creative team, or external clients.

This article presents tested and refined prompts for creating campaign performance reports with ChatGPT. The reader will find ready-to-use examples, explanations of the underlying prompt engineering principles, customization tips, and the most common pitfalls to avoid.

Ready-to-Use Prompts for Campaign Performance Reports

Prompt 1: General Campaign Report Structure

You are an expert digital marketing analyst with over 10 years of experience.
Create a structured performance report for a campaign [campaign type: e.g. Google Ads / Meta Ads / Email Marketing]
for the period [start date] - [end date].

The available data is as follows:
- Impressions: [value]
- Clicks: [value]
- CTR: [value]
- Conversions: [value]
- CPA (Cost Per Acquisition): [value]
- ROAS (Return on Ad Spend): [value]
- Total budget spent: [value]

The report must include:
1. Executive Summary (3-4 sentences)
2. Analysis of key metrics with critical commentary
3. Comparison with industry benchmarks (industry: [specify])
4. Identification of strengths and critical issues
5. Strategic recommendations for the next cycle

The report recipient is [e.g. the marketing director / the client / the creative team].
Use a [professional / technical / accessible] tone.

Why this prompt works: This prompt leverages the principle of role assignment combined with the provision of structured data and explicit output format. Specifying the recipient activates automatic calibration of language register and technical level in the model.

Expected output: A complete report in Markdown or structured text format, with well-defined sections, critical metric analysis, and concrete recommendations.

Customization tip: Add a "campaign objectives" field (e.g. brand awareness, lead generation, direct sales) to obtain recommendations more relevant to the specific context.

Prompt 2: Multi-Campaign Comparative Analysis

Analyze and compare the performance of the following campaigns from the same period [month/quarter/year]:

Campaign A - [Name]:
- Channel: [e.g. Google Search]
- Budget: [€]
- Conversions: [n]
- CPA: [€]
- ROAS: [x]

Campaign B - [Name]:
- Channel: [e.g. Meta Ads]
- Budget: [€]
- Conversions: [n]
- CPA: [€]
- ROAS: [x]

Campaign C - [Name]:
- Channel: [e.g. Email Marketing]
- Budget: [€]
- Conversions: [n]
- CPA: [€]
- ROAS: [x]

Analysis objective: identify which channel delivered the best return on investment
and provide a recommendation on budget allocation for the next quarter.

Present the results in a comparative table followed by a narrative analysis of 200-300 words.

Why this prompt works: The parallel structure of the data facilitates comparative reasoning by the model. The explicit request for a table followed by narrative leverages ChatGPT's ability to organize mixed output (tabular and textual), increasing the legibility of the final report.

Expected output: A Markdown table with main metrics side-by-side, followed by an analysis that identifies the best-performing channel and justifies the proposed budget allocation.

Customization tip: Include additional metrics specific to your industry (e.g. open rate for email, CPM for awareness campaigns) to refine the comparative analysis.

Prompt 3: Executive Summary for Non-Technical Stakeholders

Based on the following campaign performance data [campaign name]:

[Paste here raw data or complete technical report]

Write an Executive Summary of maximum 150 words intended for the Board of Directors
(or senior client) that:
- Does not use technical acronyms without explaining them
- Expresses results in terms of business impact (revenue generated, leads acquired, cost per customer)
- Clearly highlights whether campaign objectives were met or not
- Concludes with a single priority action recommendation

Tone: formal, direct, results-oriented. Avoid "agency speak" and vague statements.

Why this prompt works: The explicit word limit constraint (150 words) forces the model to prioritize the most relevant information. The instruction "avoid agency speak" is a negative prompting technique that reduces generic and promotional output, producing more credible and authoritative text.

Expected output: A concise paragraph in business language that communicates essential results without excessive jargon and concludes with clear strategic direction.

Customization tip: Add the industry context and agreed-upon KPI objectives with the client to make the comparison between expectations and results more accurate.

Prompt 4: Identifying Causes of Underperformance

The following campaign [name and type] recorded results below the set objectives:

Objectives:
- Target CPA: [€X]
- Target conversions: [n]
- Target ROAS: [x]

Actual results:
- Actual CPA: [€Y]
- Actual conversions: [n]
- Actual ROAS: [x]

Additional data available:
- Average Quality Score: [value] (if Google Ads)
- Average frequency: [value] (if Meta Ads)
- Audience segments used: [list]
- Creatives used: [describe briefly]
- Landing page: [URL or description]
- Period and seasonality: [specify]

Analyze the possible causes of underperformance by applying a structured diagnostic framework
(e.g. MECE - Mutually Exclusive, Collectively Exhaustive).
For each identified cause, assign a probability level (High / Medium / Low)
and propose a specific, measurable correction.

Why this prompt works: Explicitly requesting the application of a known analytical framework (MECE) guides the model toward systematic reasoning rather than intuitive analysis. The request to assign probability levels introduces a prioritization dimension immediately useful for the operational team.

Expected output: A structured list of potential causes, each with estimated probability, explanation, and concrete corrective action. The MECE format ensures that causes don't overlap and cover the entire problem space.

Customization tip: If you have historical data from previous similar campaigns, include it in the prompt as a comparison point to refine the diagnosis.

Prompt 5: Generating Recommendations for the Next Cycle

Based on the following performance report for the quarter [Q1/Q2/Q3/Q4] [year]:

[Insert summary of results or complete report]

Generate an optimization plan for the next quarter structured as follows:

1. **Quick Wins** (actions implementable within 2 weeks, estimated impact on CPA/ROAS)
2. **Strategic Optimizations** (medium-term actions, 1-3 months, with measurement KPIs)
3. **Experiments to Test** (A/B tests or new channels to explore, with clear hypothesis and success metric)

For each recommendation specify:
- The concrete action to take
- The suggested responsible party (e.g. creative team, media buyer, web developer)
- The success measurement metric
- Impact estimate (in % expected improvement)

Industry: [specify]
Available budget for next quarter: [€]
Strategic priority: [e.g. CPA reduction / increase conversion volume / expansion to new markets]

Why this prompt works: The temporal segmentation of recommendations (short, medium-term, experimentation) reflects the operational reality of marketing teams and produces output directly usable in an editorial or project plan. Including available budget and strategic priorities contextualizes recommendations, reducing impractical suggestions.

Expected output: A structured action plan with three levels of intervention, each with responsibilities, metrics, and impact estimates. Ideal for insertion directly into a client presentation or project management tool.

Customization tip: Specify any constraints (e.g. "we cannot increase creative budget" or "the website cannot be modified before [date]") to obtain realistically implementable recommendations.

Prompt 6: Data Storytelling for Client-Facing Presentations

Transform the following campaign performance data into an engaging narrative for a client presentation
lasting 15 minutes:

Campaign data [name]:
[Insert data]

The narrative must follow this structure:
1. **The Context** (initial challenge and agreed objectives)
2. **The Journey** (what was done, key strategic choices)
3. **The Results** (data interpreted with comparison to objectives and benchmarks)
4. **Lessons Learned** (genuine insights, not just successes)
5. **The Next Step** (clear call to action or proposal)

Tone: professional but not sterile. The client is [briefly describe the client: industry, size, level of sophistication in digital marketing].
Length: approximately 400-500 words.

Why this prompt works: The narrative structure (context → journey → results → lessons → next step) follows the principles of strategic storytelling, which increase data memorability and persuasiveness. The instruction to include "genuine lessons learned, not just successes" produces more credible output and builds trust with the client.

Expected output: Flowing narrative text, suitable for reading as a presentation script or inserting as an introductory page of a PDF report. The tone will be calibrated to the client profile specified.

Customization tip: Add two or three examples of tone the client prefers (e.g. "the client appreciates concrete facts and numbers, not metaphors") to further refine the output style.

Prompt Engineering Principles for Performance Reports

1. Always provide structured and contextualized data

ChatGPT does not have real-time access to advertising platform data. Output quality depends entirely on input quality. It's good practice to organize data in tabular format or with clear labels before inserting it into the prompt. Always include context (industry, initial objectives, period, budget) to enable the model to produce meaningful comparative analysis rather than superficial descriptions of numbers.

2. Specify the recipient and output format

A report intended for the media buying team requires technical language, industry acronyms, and granular details on campaign optimizations. A report for the CEO or senior client requires synthesis, business impact, and absence of jargon. Explicitly specifying the recipient in the prompt is one of the single most effective interventions to improve output relevance. Similarly, requesting a specific format (table + narrative, bullet points, titled paragraphs) reduces post-editing work.

3. Use role assignment to increase specialization

Assigning ChatGPT a specific role and contextual credentials (e.g. "You are a senior performance marketing analyst specialized in e-commerce with experience in the fashion industry") activates more specialized reasoning patterns. This technique is particularly effective for obtaining more accurate industry benchmarks, appropriate terminology, and recommendations more relevant to the specific operational context.

4. Explicitly request certainty levels and limitations

ChatGPT can produce statements that sound authoritative even when based on inferences or general data. For performance reports, it's essential to include instructions like "if you don't have sufficient data for a certain conclusion, indicate it explicitly" or "distinguish between observations based on the provided data and interpretive hypotheses." This practice increases output reliability and reduces the risk of presenting uncertain conclusions to the client.