Copy-paste prompts for GA4 interpretation, campaign performance analysis, trend identification, reporting summaries, anomaly detection, and executive dashboards.
Last updated: March 2026 · 12 min read
ChatGPT turns raw data exports into insights, summaries, and recommendations in minutes instead of hours. But only if you prompt it correctly.
These 33 prompts cover the most common marketing data analysis tasks. Paste your data (CSV, table, or summary) along with the prompt for best results. ChatGPT-4o and ChatGPT with Advanced Data Analysis (Code Interpreter) handle tabular data well.A ChatGPT prompt for data analysis is a structured instruction that specifies your dataset, the business question you’re trying to answer, the format you need the output in, and any context (industry benchmarks, goals, time periods) that helps ChatGPT produce actionable insights instead of descriptive summaries.
You pull data from 5 platforms and spend Friday afternoons turning it into reports. These prompts produce the narrative layer so you can focus on the strategy layer.
You need to identify what’s working, what’s declining, and where to reallocate budget. These prompts turn raw campaign data into prioritized action lists.
You receive data exports from your team but need to translate them into executive summaries for leadership. These prompts produce C-suite-ready insights from raw numbers.
7 prompts for making sense of Google Analytics 4 exports and reports.
Prompt: “Here is my GA4 traffic data by source/medium for the past 90 days: [paste data]. Analyze this data and: (1) identify the top 3 sources by engaged sessions, not just sessions, (2) flag any source where engagement rate is below 40%, (3) identify month-over-month trends for each source, (4) recommend 2 actions to improve the worst-performing source. Present findings in a summary table, then detailed analysis.”
Prompt: “Here are my top 20 landing pages from GA4 with sessions, engagement rate, conversions, and bounce rate: [paste data]. Identify: (1) the 5 pages with the highest traffic but lowest conversion rate (these are optimization priorities), (2) pages where engagement rate is above 60% but conversions are below average (content is good, CTA is failing), (3) recommend specific improvements for the top 3 underperforming pages.”
Prompt: “Here is my GA4 conversion path data showing touchpoints before conversion: [paste data]. Analyze: (1) what percentage of conversions are first-touch vs. multi-touch, (2) which channels appear most often as first touch vs. last touch, (3) the average number of touchpoints before conversion, (4) which channel combinations produce the highest conversion rate. Recommend budget allocation changes based on this.”
Prompt: “Here is a list of custom events currently tracked in our GA4 setup: [paste event names and counts]. Our business goals are: [list 3-5 goals]. Identify: (1) which events directly map to business goals, (2) which events are firing but not providing actionable data (candidates for removal), (3) which goal-critical user actions are NOT being tracked, (4) suggest an event naming convention if the current names are inconsistent.”
Prompt: “Here is GA4 data comparing two user segments: [Segment A: description and data] vs. [Segment B: description and data]. Compare these segments across: sessions, engagement rate, pages per session, average engagement time, and conversion rate. Identify the 3 most significant behavioral differences. Explain what these differences suggest about content or experience needs for each segment.”
Prompt: “Here is our GA4 ecommerce funnel data: view_item ([number]), add_to_cart ([number]), begin_checkout ([number]), purchase ([number]). Calculate drop-off rates between each step. Industry benchmark for our category ([category]) is [X]% cart-to-purchase rate. Identify which step has the biggest gap vs. benchmark and suggest 3 specific fixes for that step.”
Prompt: “Here is this week’s GA4 data vs. last week: [paste comparison data]. Write a weekly analytics summary for my team. Include: (1) top-line metrics with week-over-week change and direction arrows, (2) the single biggest positive change and likely cause, (3) the single biggest negative change and recommended response, (4) one metric to watch next week. Keep it under 300 words. No jargon.”
7 prompts for paid media, email, and multi-channel campaign analysis.
Prompt: “Here is my Google Ads campaign data for the past 30 days: [paste data with campaign name, spend, clicks, impressions, conversions, CPA, ROAS]. Our target CPA is $[X] and target ROAS is [X]. Identify: (1) campaigns exceeding targets (scale candidates), (2) campaigns within 20% of targets (optimization candidates), (3) campaigns missing targets by more than 50% (pause candidates). Recommend budget reallocation with specific dollar amounts.”
Prompt: “Here is performance data for [X] ad creatives: [paste data with headline, description, CTR, conversion rate, cost per conversion]. Identify: (1) the top 3 performing creatives by conversion rate, (2) common patterns in the top performers (length, tone, CTA type, offer), (3) common patterns in the bottom 3, (4) suggest 3 new creative concepts based on what’s working.”
Prompt: “Here are the results of our last 10 email campaigns: [paste data with subject line, send date, open rate, click rate, unsubscribe rate, revenue]. Industry benchmark open rate: [X]%, click rate: [X]%. Identify: (1) which subject line patterns drive highest opens, (2) which send days/times perform best, (3) any correlation between email length and click rate, (4) the campaign with the best revenue-per-send and what made it different.”
Prompt: “Here is our marketing spend and revenue by channel for Q[X] [year]: [paste data with channel, spend, leads, customers, revenue, CAC, LTV]. Calculate ROI for each channel. Rank channels by: (1) lowest CAC, (2) highest LTV:CAC ratio, (3) highest total revenue. Identify which channel deserves more budget and which is underperforming relative to spend. Present as a table with recommendations.”
Prompt: “Here are the results of an A/B test: Control ([metric data]) vs. Variant ([metric data]). Sample size: [X per group]. Test duration: [X days]. Calculate: (1) statistical significance (is this a real difference?), (2) confidence interval for the lift, (3) projected annual impact if we implement the variant, (4) your recommendation: ship, iterate, or retest. Explain in plain English, not statistical jargon.”
Prompt: “Here is our social media data by platform for the past quarter: [paste data with platform, followers, engagement rate, posts published, link clicks, conversions, time invested]. Calculate: (1) cost per engagement (factoring in time at $[hourly rate]), (2) cost per conversion by platform, (3) which platform has the best and worst ROI. Recommend whether we should increase, maintain, or reduce effort on each platform.”
Prompt: “Here is our marketing budget vs. actual spend for [month/quarter]: [paste data with category, budgeted, actual, variance]. Total budget: $[X]. Identify: (1) categories that overspent and by how much, (2) categories with unspent budget, (3) whether overspending categories show proportional results improvement, (4) recommended reallocations for next period. Format as an executive summary table with commentary.”
6 prompts for spotting what’s changing before it becomes a problem or opportunity.
Prompt: “Here is daily website traffic data for the past 60 days: [paste data with date, sessions, source]. Identify any days where traffic deviated more than 2 standard deviations from the 7-day rolling average. For each anomaly: (1) state the date and magnitude of deviation, (2) check if a specific traffic source caused it, (3) suggest likely causes (algorithm update, viral content, bot traffic, technical issue).”
Prompt: “Here is our weekly conversion rate data for the past 26 weeks: [paste data]. Identify: (1) the overall trend direction (improving, declining, flat), (2) any inflection points where the trend changed direction, (3) weeks with conversion rates more than 15% above or below the 4-week average, (4) correlation with any external factors I mention: [list known events like site changes, campaigns, seasonality].”
Prompt: “Here are our keyword rankings this month vs. last month: [paste data with keyword, current position, previous position, search volume]. Identify: (1) keywords that moved up 5+ positions (momentum), (2) keywords that dropped 5+ positions (risk), (3) keywords on page 2 with high search volume (near-miss opportunities). For the top 5 drops, suggest one action each to recover the position.”
Prompt: “Here is monthly [metric: revenue/traffic/leads] data for the past 24 months: [paste data]. Identify: (1) recurring seasonal patterns (which months consistently peak/trough), (2) year-over-year growth rate by month, (3) whether seasonality is strengthening or weakening, (4) recommended planning adjustments for the next 6 months based on these patterns.”
Prompt: “Here is retention data by monthly cohort: [paste data showing cohort month, month-1 retention, month-2 retention, etc.]. Calculate: (1) average retention curve across all cohorts, (2) which cohorts retained significantly better or worse than average, (3) the month where the biggest drop-off occurs, (4) whether retention is improving or declining for newer cohorts vs. older ones.”
Prompt: “Here is customer activity data for the past 90 days: [paste data with customer ID, last purchase date, purchase frequency, total spend, support tickets]. Define ‘at risk’ as: no purchase in [X] days and decreasing frequency. Identify: (1) the number of at-risk customers, (2) their total revenue at stake, (3) common patterns among at-risk customers, (4) suggested reactivation actions for each risk segment.”
7 prompts for turning raw data into reports that stakeholders actually read.
Prompt: “Here is this month’s marketing performance data: [paste key metrics with month-over-month change]. Write a 200-word executive summary for [audience: CEO/CMO/board]. Include: the single most important metric and its trajectory, one win worth celebrating, one concern requiring attention, and one recommended action for next month. Use plain language. No acronyms without definition.”
Prompt: “Here is our monthly marketing data across all channels: [paste comprehensive data]. Write a monthly report narrative covering: (1) top-line performance vs. targets (table format), (2) channel-by-channel highlights (3-4 sentences each), (3) what drove the biggest wins this month, (4) what underperformed and hypothesized reasons, (5) priorities for next month. 600-800 words. Professional tone.”
Prompt: “Here is our Q[X] marketing data: [paste quarterly data]. Create content for a 10-slide QBR presentation. For each slide, provide: slide title, 3-4 bullet points, and speaker notes (2-3 sentences). Slide order: quarterly highlights, traffic, leads, revenue, channel performance, top campaigns, budget utilization, competitive environment changes, challenges, and next quarter priorities.”
Prompt: “Here is this month’s performance data for our client [client industry/type]: [paste data]. Write a client-facing monthly summary. Include: a 3-sentence executive overview, a performance dashboard table (metric, target, actual, status), top 3 accomplishments with specific numbers, 2 areas of focus for next month, and a recommended strategy adjustment if any. Professional but approachable. Under 500 words.”
Prompt: “Here is competitive data I’ve gathered: [paste data on competitors including traffic estimates, keyword overlap, social following, ad spend estimates]. Our company metrics: [your data]. Create a competitive analysis summary: (1) side-by-side comparison table, (2) where we’re winning and why, (3) where competitors are ahead and what we can learn, (4) 3 specific opportunities based on competitor gaps.”
Prompt: “I need to present this marketing data to [audience]: [describe the data and key message]. Recommend: (1) the best chart type for this data (and why), (2) what should be on each axis, (3) which data points to highlight or annotate, (4) a color scheme that makes the key insight obvious, (5) a one-sentence chart title that conveys the main finding.”
Prompt: “Here is our marketing data for [current period] vs. [same period last year]: [paste comparison data]. Analyze: (1) which metrics improved and by how much, (2) which metrics declined, (3) whether growth is accelerating or decelerating vs. prior year trends, (4) external factors that may explain major changes (market conditions, algorithm updates, competitor moves). Write in a narrative format, not bullet points. Under 400 words.”
6 prompts for forecasting, segmentation, and statistical analysis.
Prompt: “Here is 18 months of monthly revenue data: [paste or upload data]. Using a time-series model: (1) forecast the next 6 months of revenue, (2) calculate confidence intervals (80% and 95%), (3) identify the seasonal component if any, (4) estimate the impact of [planned marketing initiative] on the forecast. Show the methodology and output a table and chart.”
Prompt: “Here is customer data with: purchase frequency, average order value, recency of last purchase, and total lifetime spend: [paste or upload data]. Run an RFM analysis and: (1) create 4-5 customer segments with descriptive names, (2) calculate the size and revenue contribution of each segment, (3) recommend a specific marketing action for each segment, (4) identify the segment with the highest growth potential.”
Prompt: “Here is marketing data with multiple variables: [paste data with columns for spend, impressions, clicks, leads, revenue, and other metrics]. Calculate the correlation between: (1) ad spend and revenue, (2) content published and organic traffic, (3) email frequency and unsubscribe rate. Identify the strongest and weakest correlations. Flag any surprising correlations. Note: correlation is not causation, but flag areas worth investigating.”
Prompt: “Here is customer purchase data: [paste data with customer ID, first purchase date, total purchases, total revenue, last purchase date]. Calculate: (1) average customer lifetime value, (2) LTV by acquisition channel if the data includes source, (3) LTV:CAC ratio if CAC data is provided: [CAC by channel], (4) projected LTV for customers acquired in the last 6 months based on early behavior patterns.”
Prompt: “Here is 12 months of marketing data with weekly spend by channel and weekly revenue: [paste data]. Run a simplified marketing mix model: (1) estimate the revenue contribution of each channel, (2) identify diminishing returns thresholds for each channel, (3) calculate the optimal budget split to maximize revenue at $[total budget], (4) show what happens to revenue if we shift 20% of budget from [channel A] to [channel B].”
Prompt: “Here is a raw data export from [platform]: [paste or upload data]. Audit this data for: (1) missing values by column (count and percentage), (2) duplicate rows, (3) inconsistent formats (date formats, number formats, text casing), (4) outliers that may indicate tracking errors, (5) columns with zero variance (useless for analysis). Provide a data quality score (% of clean rows) and specific cleaning recommendations.”
“The analysts on our team who adopted these prompt patterns cut their weekly reporting time from 4 hours to under 90 minutes. The time savings didn’t come from faster number-crunching. It came from ChatGPT writing the narrative and recommendations that used to take the longest to produce.”
Hardik Shah, Founder of ScaleGrowth.Digital
| Platform | Best Export Format | What to Analyze |
|---|---|---|
| Google Analytics 4 | CSV from Explorations | Traffic, engagement, conversions, paths |
| Google Ads | CSV report download | Campaign performance, keywords, ad groups |
| Meta Ads Manager | CSV export | Ad performance, audience insights, creative |
| Google Search Console | CSV from Performance | Rankings, CTR, impressions by query |
| HubSpot/Salesforce | CSV export | Leads, pipeline, attribution, lifecycle |
| Email platforms | Campaign report CSV | Open rates, clicks, revenue, engagement |
| Social platforms | Native analytics CSV | Engagement, reach, follower growth |
Tip: Use ChatGPT’s Advanced Data Analysis (Code Interpreter) for CSV uploads over 5,000 rows. Paste smaller datasets directly in the chat.
Calculate ROI across channels with our free calculator. Input spend and revenue, get ROI, ROAS, and breakeven analysis. Use Calculator →
Pre-built report structure with charts, commentary sections, and executive summary format. Works with the reporting prompts above. Get Template →
Set up GA4 event tracking correctly so the data you feed into ChatGPT is clean and complete from the start. Read Guide →
Remove personally identifiable information (names, emails, phone numbers) before pasting. Use customer IDs instead. OpenAI’s ChatGPT Team and Enterprise plans offer data privacy protections including no training on your data. For sensitive financial data, use the API with a data processing agreement in place.
No. ChatGPT is a batch analysis and narrative tool, not a real-time dashboard. Tableau and Looker Studio connect to live data sources and update automatically. ChatGPT excels at ad-hoc analysis, insight generation, and report writing. Use dashboards for monitoring, ChatGPT for deep-dive analysis.
For basic statistics (mean, median, percentages, growth rates), ChatGPT is highly accurate. For complex models (regression, time-series forecasting), use the Advanced Data Analysis feature which runs actual Python code. Always verify critical business decisions against a dedicated analytics tool.
Pasted data in the chat window is limited to about 3,000-4,000 rows depending on column count. For larger datasets, use the file upload feature with Advanced Data Analysis, which handles CSV files up to 512 MB. For very large datasets (millions of rows), pre-aggregate in SQL or Excel, then paste the summary.
The free plan handles basic analysis from pasted data. ChatGPT Plus ($20/month as of March 2026) gives you GPT-4o and Advanced Data Analysis with file uploads. For team use with data privacy protections, ChatGPT Team ($25/user/month) or Enterprise (custom pricing) is recommended.
Data analysis tells you what happened. AI visibility strategy determines what happens next. We do both. Explore AI Visibility Services →