📊 Professional Training
Marketing Analytics
Using Claude
Master marketing analytics with AI. From campaign performance to customer segmentation, SEO, attribution, and growth analytics — all with real-world examples and copy-paste Claude prompts.
MODULE 01 · 12 SLIDES
🏗️ Foundation
Claude basics · Workflow setup · Prompt engineering for analytics
MODULE 02 · 14 SLIDES
📈 Campaign Performance
CTR/ROAS/CPC · A/B testing · Budget allocation · Creative analysis
MODULE 03 · 10 SLIDES
👥 Customer Segmentation
RFM analysis · Cohort retention · LTV modelling · Churn prediction
MODULE 04 · 7 SLIDES
🔍 SEO & Content Analytics
Keyword opportunity · Content audit · Competitor gap analysis
MODULE 05 · 7 SLIDES
🎯 Attribution & Paid Ads
Attribution models · Multi-touch · Incrementality · ROAS optimisation
MODULE 06 · 5 SLIDES
📱 Social Media & Growth
Social metrics · Growth funnel · Viral coefficient · SaaS metrics
MODULE 07 · 14 SLIDES
📋 Reporting & Advanced Analytics
CMO dashboards · Influencer analytics · Viral coefficient · Forecasting
FINAL EXAM · 50 QUESTIONS
🎓 Certification Exam
Pass 70% · Max 2 attempts · 1-year certificate validity
Slide 1 of 12
Use Next → to navigate
Module 1 — Foundation
Why Claude for Marketing Analytics?
Marketing teams drown in data — Google Analytics, Meta Ads Manager, HubSpot, Klaviyo, SEMrush — yet most can't answer basic questions like "which campaign actually drove revenue?" Claude cuts through the noise and gives you answers in minutes, not days.
⚡
Speed
Airbnb's marketing team reduced weekly reporting from 2 days to 2 hours using AI-assisted analysis. Claude makes this possible for any team.
🎯
Pattern Recognition
Claude spots trends across 10,000 rows of campaign data that humans miss — declining CTR before it hits revenue, rising CPCs in specific segments.
💬
Natural Language
Ask "why did our ROAS drop in week 3?" in plain English. No SQL, no pivot tables, no dashboard configuration — just answers.
📊
Synthesis
Pull data from Google Ads, Meta, and email into one analysis. Claude synthesises cross-channel insights that no single dashboard provides.
✍️
Reporting
Turn raw data into executive-ready insights. What took 3 hours of slide-building takes 10 minutes with Claude drafting the narrative.
🔄
Iteration
Test hypotheses instantly. "If we increase budget on Campaign A by 20%, what happens to blended ROAS?" Claude walks through the logic in seconds.
The Shift That's Happening
Nike, Coca-Cola, and every top DTC brand now use AI in their analytics workflows. The marketers who thrive are not those who resist it — they are those who know how to direct it precisely. This course gives you that skill.
Module 1 — Foundation
Claude Basics for Marketers
Before using Claude for analytics, understand what it knows, how it thinks, and where it needs your help. This mental model will save you hours of frustration and dramatically improve your outputs.
1
Claude understands marketing deeply
CTR, ROAS, CAC, LTV, MQL, SQL, funnel stages, attribution models, cohort analysis, A/B testing — Claude knows these at a senior marketing analyst level. Use the jargon freely.
2
Claude does NOT have your real-time data
Claude cannot log into your Google Analytics, Meta Ads, or HubSpot. You paste the data into Claude. Think of it as a brilliant analyst sitting next to you — they need you to share your screen, not log in for them.
3
Claude reasons, not retrieves
When you ask "is a 2.4% CTR good?", Claude reasons from its knowledge of industry benchmarks, your specific channel, and the context you've provided — not from a live database lookup.
4
Context is everything
A 3x ROAS might be excellent for a luxury brand with high AOV or terrible for a low-margin FMCG product. Always give Claude context: industry, business model, margins, objectives.
🌍 Real World — How Zomato Uses AI Analytics
Zomato's growth team uses AI to analyse order patterns, restaurant performance, and ad campaign ROI across 800+ cities simultaneously. The same analytical thinking — applied at your scale — is what this course teaches.
Module 1 — Foundation
Setting Up Your Analytics Workflow
The best Claude analytics workflow follows a consistent 4-step pattern. Teams that follow this structure get 3x better insights than those who just paste data and ask vague questions.
1
Set the context (Business + Objective)
"You are a senior marketing analyst. I run a D2C skincare brand with ₹800 AOV and 45% gross margin. My objective is to optimise paid acquisition for ROAS above 3x."
2
2
Provide the data (Clean, labelled)
Paste your data clearly with column headers and date ranges: "Google Ads data, last 30 days. Columns: Campaign | Impressions | Clicks | CTR | CPC | Conversions | Conv Rate | Cost | Revenue | ROAS"
3
Ask a specific question
"Which 3 campaigns are dragging down our blended ROAS? For each, identify whether the issue is CTR, conversion rate, or CPC. Recommend one specific action per campaign."
4
Request the output format
"Output as: (1) an executive summary in 3 bullets, (2) a ranked table of campaigns by ROAS, (3) specific recommendations in order of expected impact."
Perfect Opening Prompt Template
You are a senior marketing analyst. Business context: - Brand: [DTC/B2B/FMCG/SaaS] - Objective: [ROAS optimisation / lead gen / brand awareness] - Key metric: [ROAS / CPL / CAC / CTR] - Target: [e.g. ROAS > 3x, CPL < ₹500] Data: [paste your data here] Question: [your specific analytical question] Output format: [table / bullets / narrative / recommendation list]
Module 1 — Foundation
Prompt Engineering for Marketing Analytics
The difference between a mediocre and a brilliant Claude analytics session is your prompt. These patterns, used by performance marketing teams at HubSpot, Shopify, and top Indian DTC brands, will transform your outputs.
| ❌ Vague Prompt | ✅ Precise Prompt |
|---|---|
| "Analyse my campaign data" | "Rank these 8 campaigns by ROAS. Flag any with ROAS below 2x. For the bottom 3, identify whether the problem is CTR, CVR, or CPC and suggest one fix each." |
| "Is my CTR good?" | "My Google Search CTR is 4.2%. Benchmark this against e-commerce industry average (typically 2-5%). Am I above or below? What would moving to 5.5% mean for monthly traffic at 50,000 impressions?" |
| "Why did revenue drop?" | "Revenue dropped 18% WoW (Week 8 vs Week 7). Data provided shows impressions up 5%, CTR down 12%, CVR down 8%, AOV flat. Diagnose the root cause and rank the contributing factors." |
| "Help with segmentation" | "Segment these 500 customers by RFM score using: Recency (days since last purchase), Frequency (orders in 6 months), Monetary (total spend). Create 4 segments and name each with a marketing persona." |
The Specificity Rule
Every vague word in your prompt produces a vague output. Replace "analyse" with a specific action (rank, compare, diagnose, flag, quantify). Replace "good" with a benchmark (vs industry average, vs last period, vs target). Replace "help" with a deliverable (table, recommendation, narrative).
Module 1 — Foundation
Marketing Metrics Claude Understands
You can use professional marketing terminology with Claude without defining it. Here's a reference of what Claude knows at an expert level across every major channel.
| Channel | Metrics Claude knows deeply |
|---|---|
| Paid Search (Google) | CTR, CPC, Quality Score, Ad Rank, Impression Share, Search Lost IS (Budget/Rank), ROAS, Conv Rate, CPA, POAS |
| Paid Social (Meta/LinkedIn) | CPM, CPC, CTR, Frequency, Reach, Engagement Rate, Video Views, ThruPlay, ROAS, Hook Rate, Hold Rate, CPL, CPAS |
| Email / CRM | Open Rate, Click Rate, CTOR, Bounce Rate, Unsubscribe Rate, Revenue per Email, List Growth Rate, Deliverability Score, Segment Health |
| SEO | Organic Sessions, Keyword Rankings, Click-Through Rate, Impressions, Domain Authority, Backlink Profile, Core Web Vitals, Featured Snippets |
| Growth / Product | DAU/MAU, Retention Rate, Churn, NPS, CAC, LTV, LTV:CAC, Payback Period, Viral Coefficient, Activation Rate, Conversion Funnel |
| Attribution | First-touch, Last-touch, Linear, Time-decay, Data-driven, MTA, MMM, Incrementality, View-through, Post-click, ROPI |
Module 1 — Foundation
💡 The Marketer's Edge
💡 Insight Card — Foundation
"The Marketer's Edge: Data Literacy + AI Fluency"
Here is the uncomfortable truth about marketing analytics in 2025: most marketing teams have more data than they can use. They have Google Analytics, Meta Pixel, HubSpot, Klaviyo, Google Search Console, Hotjar, and 12 other tools — and they still can't answer the question their CMO asked on Monday morning.
The problem is not data. The problem is analysis speed and synthesis.
Claude doesn't solve your data collection problem. But it solves your analysis problem spectacularly. A question that would take a junior analyst 4 hours — "which audience segments drove our best LTV customers last quarter?" — Claude can answer in 4 minutes with the right prompt and the right data.
The marketers who will be most valuable in the next 5 years: Those who understand marketing deeply enough to ask the right questions, AND know how to use Claude to answer them faster than any competitor.
Data literacy is your moat. AI fluency is your force multiplier. This course gives you both.
The problem is not data. The problem is analysis speed and synthesis.
Claude doesn't solve your data collection problem. But it solves your analysis problem spectacularly. A question that would take a junior analyst 4 hours — "which audience segments drove our best LTV customers last quarter?" — Claude can answer in 4 minutes with the right prompt and the right data.
The marketers who will be most valuable in the next 5 years: Those who understand marketing deeply enough to ask the right questions, AND know how to use Claude to answer them faster than any competitor.
Data literacy is your moat. AI fluency is your force multiplier. This course gives you both.
Claude doesn't replace marketing judgment. It eliminates the analysis latency between insight and action — which is where most marketing opportunities are lost.
Module 1 — Foundation
How to Feed Data to Claude
The quality of your Claude analytics session starts before you type a single question — it starts with how you prepare and format your data. These methods work for any marketing data source.
1
Copy-paste from spreadsheets (most common)
Export from Google Ads / Meta / Analytics to CSV. Open in Excel/Sheets. Copy the data table. Paste directly into Claude. It reads tab-separated and comma-separated data accurately. Always include the header row.
2
Ask Claude to structure messy data first
"I'm going to paste raw export data from Meta Ads Manager. It has inconsistent formatting and merged cells. First, parse this into a clean table with standardised column names, then I'll ask you to analyse it."
3
Provide benchmarks alongside your data
"Here is my data. For context, industry benchmarks for e-commerce on Google Search are: CTR 3.2%, Conv Rate 2.8%, CPC ₹45. Please flag any metric where I'm more than 20% below benchmark."
4
Multi-channel data: label each source clearly
When providing data from multiple channels, label each section: "=== GOOGLE ADS DATA ===" then "=== META ADS DATA ===". This prevents Claude from conflating metrics across channels.
🌍 Real World — How Nykaa Analyses Cross-Channel Data
Nykaa's performance team uses a weekly "marketing dashboard" prompt — a template that pastes data from 5 channels into Claude and asks for a blended performance summary, top and bottom performers, and weekly budget reallocation recommendations. You'll build this exact workflow in Module 5.
Module 1 — Foundation
Output Formats for Marketing Reports
Claude can deliver marketing analysis in any format you need — from quick Slack messages to board-ready decks. Always specify the format upfront so you get output you can use directly.
| Use Case | Best Format | Prompt instruction |
|---|---|---|
| CMO/CEO weekly update | 3-5 bullet executive summary | "Output as 5 bullets: performance vs target, top win, biggest concern, key insight, and recommended action this week" |
| Campaign optimisation | Ranked table + recommendations | "Output a table ranked by ROAS, then below it, 3 specific optimisation actions in priority order" |
| Agency/client report | Narrative + data | "Write a 200-word performance narrative suitable for a monthly client report, professional tone" |
| Internal Slack update | Short, scannable bullets | "Summarise in 5 bullets, casual tone, emoji ok, suitable for posting in #marketing-weekly" |
| Budget presentation | Structured argument | "Write a business case for increasing Google Ads budget by 30%, citing the data. Include expected ROI." |
| A/B test result | Statistical summary | "Summarise the A/B test result: winner, confidence level, expected uplift, and recommendation to scale or kill" |
Module 1 — Foundation
💡 Garbage Data, Garbage Insights
💡 Insight Card — Data Quality
"Garbage Data, Garbage Insights — Even With Claude"
The most common mistake marketers make with Claude: they paste data without cleaning it first, then blame Claude when the analysis is off.
The typical culprits:
• Mismatched date ranges (comparing 28-day periods to 30-day periods)
• Different attribution windows in different data sources (Meta 7-day click vs Google last-click)
• Untracked revenue (users who called instead of bought online)
• Bot traffic inflating impressions and dragging down CTR
• UTM parameter inconsistencies causing campaign revenue to appear in "Direct"
Real example: A fashion brand couldn't understand why their Meta ROAS showed 4x in Meta Ads Manager but Google Analytics showed the same campaigns driving only 1.8x revenue. The discrepancy? Meta was using a 7-day click + 1-day view attribution window; GA was using last-click. Both were technically correct — but incomparable.
The fix: Before every Claude analytics session, state your data sources, attribution windows, and date ranges explicitly. "All data is Google Analytics 4, last-click attribution, last 30 days (Jun 1-30). Meta data is 7-day click only."
The typical culprits:
• Mismatched date ranges (comparing 28-day periods to 30-day periods)
• Different attribution windows in different data sources (Meta 7-day click vs Google last-click)
• Untracked revenue (users who called instead of bought online)
• Bot traffic inflating impressions and dragging down CTR
• UTM parameter inconsistencies causing campaign revenue to appear in "Direct"
Real example: A fashion brand couldn't understand why their Meta ROAS showed 4x in Meta Ads Manager but Google Analytics showed the same campaigns driving only 1.8x revenue. The discrepancy? Meta was using a 7-day click + 1-day view attribution window; GA was using last-click. Both were technically correct — but incomparable.
The fix: Before every Claude analytics session, state your data sources, attribution windows, and date ranges explicitly. "All data is Google Analytics 4, last-click attribution, last 30 days (Jun 1-30). Meta data is 7-day click only."
Tell Claude your data limitations upfront. A good prompt includes not just what the data shows, but what it cannot show.
Module 1 — Foundation
Limitations — When Not to Rely on Claude
❌ Real-time data access
Claude cannot log into your Google Analytics, Meta Business Manager, HubSpot, or any platform. You must paste or describe the data. It cannot pull live numbers.
❌ Your proprietary benchmarks
Claude knows industry averages but not your company's specific historical performance. Tell Claude: "Our internal benchmark is 3.5% CTR based on last 6 months performance."
❌ Causation without context
If you don't tell Claude what changed (new creative, budget shift, competitor promotion, seasonality), it cannot diagnose the cause — only the symptoms. Always provide context around changes.
❌ Statistical significance alone
Claude can calculate statistical significance if you provide the data, but it cannot run tests independently. Provide: control variant performance, test variant performance, sample sizes, and duration.
✅ Always reliable for
Pattern identification, metric calculation, framework application, report drafting, hypothesis generation, benchmark comparison (with data you provide), and recommendation prioritisation.
Module 1 — Key Distinctions
Key Distinctions — Foundation
Metrics vs KPIs
Metrics are any measurable value (impressions, clicks, sessions). KPIs are the metrics directly tied to business objectives (ROAS if you're a DTC brand; CPL if you're B2B). Claude should always anchor its analysis to your KPIs, not just any metric.
Correlation vs Causation
Claude can identify correlations in your data (CTR fell when CPM rose). It cannot confirm causation without controlled testing. Always say "this suggests" not "this proves" when acting on correlation-based insights.
Vanity Metrics vs Action Metrics
Impressions and follower counts are vanity metrics — they feel good but don't drive decisions. CTR, ROAS, CAC, and LTV are action metrics. Ask Claude to focus analysis on action metrics unless you explicitly need vanity metrics for stakeholder reporting.
Last-Click vs Multi-Touch Attribution
Last-click attribution gives 100% credit to the final touchpoint before purchase. This systematically undervalues top-of-funnel channels (SEO, display, social) and overvalues bottom-of-funnel (branded search). Tell Claude which model your data uses so it interprets ROAS correctly.
Module 1 — Quiz
Quiz — Foundation (8 Questions)
Q1. What is the most important context to give Claude before a marketing analytics session?
A. Your company name and logo
B. Business model, objective, key metric, target, and data source/attribution window
C. Your marketing budget
D. The number of campaigns running
Q2. A fashion brand sees 4x ROAS in Meta Ads Manager but only 1.8x in Google Analytics. The most likely explanation is:
A. Claude made an error in the calculation
B. Different attribution windows — Meta uses 7-day click + 1-day view; GA uses last-click. Both are correct but incomparable without aligning the windows.
C. The campaigns are underperforming
D. Google Analytics is always more accurate
Q3. Which is an "action metric" vs a "vanity metric"?
A. Follower count — action metric
B. Impressions — action metric
C. CAC (Customer Acquisition Cost) — action metric; Impressions — vanity metric
D. Both are vanity metrics
Q4. When pasting multi-channel data into Claude, the best practice is:
A. Combine all data into one table
B. Label each data source clearly (e.g., "=== GOOGLE ADS ===") and specify attribution windows for each
C. Only paste the summary row, not raw data
D. Convert all metrics to percentages first
Q5. Claude identifies that CTR rose when you increased budget. This is:
A. Proof that budget drives CTR
B. A correlation — it suggests a relationship but requires controlled testing to confirm causation
C. A KPI improvement
D. A vanity metric change
Q6. The best prompt for campaign analysis is:
A. "Analyse my campaign data"
B. "Rank these 8 campaigns by ROAS. Flag any below 2x. For the bottom 3, identify whether the issue is CTR, CVR, or CPC and suggest one specific fix each."
C. "Is my campaign good?"
D. "Help with marketing"
Q7. What does Claude need that it cannot get on its own?
A. Knowledge of marketing metrics
B. Your actual data — Claude cannot log into Google Analytics, Meta, or any platform. You must paste or describe the data.
C. Understanding of attribution models
D. Knowledge of A/B testing
Q8. Last-click attribution systematically:
A. Overvalues top-of-funnel channels like SEO and display
B. Undervalues top-of-funnel channels and overvalues bottom-of-funnel channels like branded search
C. Equally distributes credit across all touchpoints
D. Only counts paid channels
Slide 1 of 12
Slide 1 of 14
Use Next → to navigate
Module 2 — Campaign Performance
Reading Campaign Data with Claude
Most marketers look at their campaign dashboard and see numbers. Claude helps you see stories — why a campaign is underperforming, which lever to pull first, and what the data would look like if your hypothesis is correct.
📈
Paid Search
CTR, Quality Score, Impression Share — Claude diagnoses Google Ads performance at campaign, ad group, and keyword level.
📱
Paid Social
Hook Rate, Hold Rate, CPM, ROAS — Claude reads Meta/TikTok/LinkedIn data and tells you if your creative or your targeting is the bottleneck.
📧
Email
Open Rate, CTOR, Revenue per Email — Claude analyses your email performance and identifies deliverability, subject line, or content issues.
🔄
Cross-Channel
Blended ROAS, channel contribution, incrementality — Claude synthesises performance across all your channels into one coherent story.
🚩
Red Flags
Rising CPCs, falling Quality Scores, frequency fatigue, click-to-purchase gap — Claude spots warning signals before they hit your revenue.
💡
Budget Allocation
Which campaigns deserve more budget? Claude applies marginal ROAS analysis to recommend reallocation with expected revenue impact.
🌍 Real World — Swiggy's Campaign Optimisation
Swiggy's performance marketing team reviews 200+ active campaigns weekly. Their workflow: export all campaign data → paste into AI analysis → get top 10 and bottom 10 by efficiency → redistribute budget. Claude makes this workflow accessible to any team, not just unicorns with data science teams.
Module 2 — Campaign Performance
CTR, CPC, ROAS — Interpreting the Core Metrics
These three metrics tell 80% of the performance story for most paid campaigns. Claude can interpret them, benchmark them, and tell you what to fix — but only if you ask precisely.
Core Metrics Analysis Prompt
I run Google Search campaigns for a B2C e-commerce brand (electronics). Target ROAS: 4x. Here is last 30 days data: Campaign | Impressions | Clicks | CTR | CPC | Conv | CVR | Revenue | ROAS Brand Keywords | 45,000 | 3,150 | 7.0% | ₹18 | 210 | 6.7% | ₹8,82,000 | 13.1x Generic-TV | 1,20,000 | 2,160 | 1.8% | ₹65 | 32 | 1.5% | ₹1,92,000 | 1.2x Generic-Mobile | 85,000 | 1,530 | 1.8% | ₹72 | 28 | 1.8% | ₹1,68,000 | 1.0x Competitor KW | 32,000 | 640 | 2.0% | ₹95 | 19 | 3.0% | ₹1,14,000 | 1.9x Shopping-Premium | 65,000 | 1,950 | 3.0% | ₹48 | 97 | 5.0% | ₹5,82,000 | 6.2x 1. Rank campaigns by ROAS. Flag any below target (4x). 2. For each underperformer: is the problem CTR, CVR, or CPC? 3. Which campaigns should get more budget? Which should be paused? 4. What is the blended ROAS across all campaigns?
What Claude Will Tell You
From this data, Claude will identify that Generic-TV and Generic-Mobile are destroying blended ROAS (dragging a 13x brand campaign down to a blended 3.8x). Shopping-Premium at 6.2x deserves more budget. Competitor Keywords at 1.9x need a conversion rate fix or a pause. This analysis takes 30 seconds with Claude vs 2 hours manually.
Module 2 — Campaign Performance
Diagnosing Campaign Underperformance
When a campaign underperforms, there are only 3 possible root causes: not enough people see the ad (reach/impression problem), not enough people click it (CTR/creative problem), or not enough people buy after clicking (CVR/landing page problem). Claude helps you find which one.
The Performance Diagnostic Framework
Step 1 — Check impressions: If impressions are low → budget, bidding, or targeting issue
Step 2 — Check CTR: If impressions ok but CTR low → creative or ad copy issue
Step 3 — Check CVR: If clicks ok but conversions low → landing page, offer, or audience mismatch
Step 4 — Check CPC: If all else ok but ROAS low → CPC too high, need better Quality Score or different keyword strategy
Step 2 — Check CTR: If impressions ok but CTR low → creative or ad copy issue
Step 3 — Check CVR: If clicks ok but conversions low → landing page, offer, or audience mismatch
Step 4 — Check CPC: If all else ok but ROAS low → CPC too high, need better Quality Score or different keyword strategy
Diagnostic Prompt
Campaign "Women's Activewear - Prospecting" has dropped ROAS from 3.2x (last month) to 1.8x (this month). Data comparison: Metric | Last Month | This Month | Change Impressions | 2,40,000 | 2,85,000 | +18.8% CTR | 2.8% | 1.9% | -32% CPC | ₹42 | ₹58 | +38% CVR | 3.1% | 2.9% | -6% AOV | ₹1,850 | ₹1,820 | -1.6% ROAS | 3.2x | 1.8x | -44% Diagnose: What is the primary driver of ROAS decline? Is it a reach problem, creative problem, or conversion problem? What is the ONE change that would have the biggest impact?
🌍 Real World — How Myntra Diagnoses Campaign Issues
Myntra's performance team uses a "waterfall diagnostic" — they check funnel metrics from top to bottom (impressions → clicks → add-to-cart → purchase) to isolate where the drop-off is happening. This is exactly what the prompt above does using Claude.
Module 2 — Campaign Performance
💡 Statistical Significance in A/B Testing
💡 Insight Card — A/B Testing
"Your A/B Test Winner Is Probably Wrong"
This is the most expensive mistake in performance marketing: declaring an A/B test winner based on insufficient data.
The typical mistake: A brand runs two ad creatives for 3 days. Creative A: 2.1% CTR. Creative B: 2.8% CTR. They declare B the winner and scale it. Two weeks later, B is underperforming. Why? Because 3 days of data with 500 clicks per variant is not statistically significant.
What statistical significance actually means: If you ran this test 100 times, the same winner would emerge at least 95 times (at 95% confidence). Below 95% confidence, your "winner" might just be random variation.
The minimum viable test: For a typical e-commerce conversion rate of 3%, you need approximately 1,000+ conversions per variant to detect a 10% improvement at 95% confidence. Most teams don't wait this long.
Use Claude to check significance: "We ran an A/B test for 14 days. Variant A: 12,000 visitors, 348 conversions (2.9%). Variant B: 11,800 visitors, 389 conversions (3.3%). Is this result statistically significant at 95% confidence? What is the expected revenue uplift if we scale B?"
Claude will calculate significance, confidence interval, and expected uplift in seconds.
The typical mistake: A brand runs two ad creatives for 3 days. Creative A: 2.1% CTR. Creative B: 2.8% CTR. They declare B the winner and scale it. Two weeks later, B is underperforming. Why? Because 3 days of data with 500 clicks per variant is not statistically significant.
What statistical significance actually means: If you ran this test 100 times, the same winner would emerge at least 95 times (at 95% confidence). Below 95% confidence, your "winner" might just be random variation.
The minimum viable test: For a typical e-commerce conversion rate of 3%, you need approximately 1,000+ conversions per variant to detect a 10% improvement at 95% confidence. Most teams don't wait this long.
Use Claude to check significance: "We ran an A/B test for 14 days. Variant A: 12,000 visitors, 348 conversions (2.9%). Variant B: 11,800 visitors, 389 conversions (3.3%). Is this result statistically significant at 95% confidence? What is the expected revenue uplift if we scale B?"
Claude will calculate significance, confidence interval, and expected uplift in seconds.
Never call a test winner until Claude confirms statistical significance. Gut instinct plus insufficient data is the most expensive combination in marketing.
Module 2 — Campaign Performance
A/B Test Analysis with Claude
A/B testing is the scientific method applied to marketing. Claude can calculate statistical significance, interpret results correctly, and tell you whether you have enough data to make a decision — before you make an expensive mistake.
A/B Test Analysis Prompt
We tested two email subject lines for our Black Friday campaign: Variant A: "Black Friday Sale — Up to 50% Off" - Sent: 45,000 | Opens: 8,550 | Open Rate: 19.0% - Clicks: 1,710 | CTOR: 20.0% | Revenue: ₹8,55,000 Variant B: "Only 6 hours left — your cart is waiting 🛒" - Sent: 45,000 | Opens: 11,250 | Open Rate: 25.0% - Clicks: 2,812 | CTOR: 25.0% | Revenue: ₹14,06,000 1. Is this result statistically significant at 95% confidence? 2. Calculate: revenue uplift from B vs A 3. What is the expected annual revenue impact if we use urgency/personalisation subject lines for all campaigns? 4. What 3 hypotheses does this data suggest for future tests?
✓
Testing hierarchy: what to test first
Ask Claude: "Given limited testing bandwidth (we can run 2 tests per month), what should we test first — subject lines, send time, CTA button colour, or landing page headline? Rank by expected impact based on our current open rate of 19% and CVR of 2.1%."
Module 2 — Campaign Performance
Budget Allocation with Claude
Budget allocation is the highest-leverage marketing decision you make. Moving ₹1 lakh from a 1x ROAS campaign to a 6x ROAS campaign doesn't just improve efficiency — it multiplies revenue. Claude makes this analysis systematic.
Budget Reallocation Prompt
Monthly budget: ₹10,00,000. Current allocation and performance: Channel | Budget | Revenue | ROAS | Notes Google Brand | ₹1,20,000 | ₹15,60,000 | 13x | Near saturation Google Shopping | ₹2,40,000 | ₹14,40,000 | 6x | Room to scale Google Generic | ₹3,00,000 | ₹3,00,000 | 1x | Poor performance Meta Prospecting | ₹2,00,000 | ₹4,00,000 | 2x | New creative testing Meta Retargeting | ₹80,000 | ₹4,80,000 | 6x | Audience size limited Email/CRM | ₹60,000 | ₹9,00,000 | 15x | Limited scalability 1. Recommend optimal budget reallocation keeping total at ₹10L 2. Calculate projected revenue under new allocation 3. What is the revenue opportunity cost of keeping Google Generic alive? 4. At what ROAS threshold should we pause a channel?
The Marginal ROAS Principle
Never evaluate campaigns on average ROAS alone. Ask Claude to calculate marginal ROAS — "if we add ₹50,000 more to Google Shopping, what is the expected incremental return?" High-ROAS campaigns near saturation have low marginal ROAS. Low-ROAS campaigns with untapped audience may have high marginal ROAS. Claude can model this.
Module 2 — Campaign Performance
Creative Performance Analysis
Great targeting with bad creative is still a losing strategy. Claude helps you analyse creative performance data — Hook Rate, Hold Rate, CTR by creative — to identify what's working and why.
Meta Creative Analysis Prompt
Analyse Meta ad creative performance (last 30 days): Creative | Type | Impressions | Hook% | Hold% | CTR | CPM | ROAS | Theme Creative-A | Video | 2,10,000 | 38% | 65% | 2.8% | ₹180 | 3.2x | Founder story Creative-B | Video | 1,85,000 | 22% | 45% | 1.4% | ₹195 | 1.8x | Product demo Creative-C | Carousel| 1,40,000 | N/A | N/A | 3.1% | ₹165 | 2.9x | UGC/Reviews Creative-D | Static | 90,000 | N/A | N/A | 1.9% | ₹210 | 2.1x | Discount offer Creative-E | Video | 1,60,000 | 45% | 72% | 3.4% | ₹172 | 4.1x | Before/After 1. Which creative is top performer? Why? 2. Creative-B has low Hook Rate (22%) — what does this mean and how to fix? 3. What creative themes should we test next based on this data? 4. How do we scale Creative-E without audience fatigue?
🌍 Real World — How boAt Uses Creative Analytics
boAt Lifestyle, India's top audio brand, systematically tracks Hook Rate (first 3 seconds views / impressions) for every video ad. Anything below 30% Hook Rate gets paused immediately. Anything above 40% gets scaled. Claude makes this analysis instant — paste the data, get the decision.
Module 2 — Campaign Performance
Email Campaign Analysis
Email is still the highest-ROAS channel for most brands — but only if you use it right. Claude helps you diagnose email performance issues, identify your best-performing segments, and optimise send strategy.
Email Performance Analysis Prompt
Email campaign performance — last 8 sends (e-commerce brand): Date | Campaign | Sent | Opens | OR% | Clicks | CR% | Revenue | RPE Aug 1 | Weekly Newsletter | 45,000 | 7,650 | 17% | 918 | 12% | ₹2,75,400 | ₹6.1 Aug 5 | Flash Sale 24hr | 45,000 | 13,050| 29% | 3,267 | 25% | ₹16,33,500| ₹36.3 Aug 8 | New Arrivals | 44,200 | 7,514 | 17% | 1,503 | 20% | ₹4,51,000 | ₹10.2 Aug 12 | Abandoned Cart | 8,400 | 4,536 | 54% | 1,361 | 30% | ₹8,16,000 | ₹97.1 Aug 15 | Weekend Offer | 44,800 | 6,272 | 14% | 940 | 15% | ₹2,82,000 | ₹6.3 Aug 19 | VIP Segment Only | 5,200 | 3,224 | 62% | 1,096 | 34% | ₹7,67,000 | ₹147.5 Aug 22 | Product Review Ask| 43,500 | 5,220 | 12% | 522 | 10% | ₹52,200 | ₹1.2 Aug 26 | Re-engagement | 12,000 | 2,040 | 17% | 306 | 15% | ₹1,83,600 | ₹15.3 (RPE = Revenue Per Email sent) 1. Rank campaigns by RPE. What pattern emerges? 2. What 3 campaign types should get priority in next month? 3. The Weekly Newsletter has 17% OR — is this good or concerning? 4. How do we increase VIP Segment revenue without burning the list?
Module 2 — Campaign Performance
Weekly Performance Review Framework
The best marketing teams have a weekly rhythm of data review. Claude makes this review faster, more consistent, and more insightful than any dashboard. Here is the exact framework used by top performance teams.
Weekly Marketing Review Prompt
Weekly marketing performance review — Week 32 (Aug 5-11): vs Week 31 (Jul 29 - Aug 4) PAID CHANNELS: Total Spend: ₹8,20,000 (+5% WoW) Total Revenue: ₹28,70,000 (+2% WoW) Blended ROAS: 3.5x (-0.1x WoW) New Customers Acquired: 487 (+3%) CAC: ₹1,684 (+2%) TOP PERFORMER: Google Shopping ROAS 6.2x (+0.4x) BOTTOM PERFORMER: Meta Prospecting ROAS 1.9x (-0.6x) ORGANIC: SEO Sessions 42,100 (+8% WoW) EMAIL: Revenue ₹4,20,000 (1 campaign sent) Generate weekly review output: 1. Executive summary (3 bullets) — wins, concerns, priority action 2. Budget recommendation for next week 3. One hypothesis to test this week based on the data 4. One question the CMO will ask that this data cannot answer
Save This Prompt as a Template
Build a weekly review template that your team fills in every Monday morning. The same structure, every week, means Claude's outputs are comparable week-over-week and trends become visible over time.
Module 2 — Campaign Performance
💡 Last-Click is Lying to You
💡 Insight Card — Attribution
"Last-Click Attribution Is Lying to You — And Costing You Money"
Last-click attribution is the default in most analytics tools. It gives 100% of the conversion credit to the last touchpoint before purchase. This creates a systematic illusion that distorts every budget decision you make.
What it makes you do: Over-invest in branded search (the last click before purchase for almost every customer) and under-invest in top-of-funnel channels that generate demand in the first place (Instagram, YouTube, SEO, influencer).
A real example: A DTC skincare brand was about to cut their Instagram budget because Instagram showed only 1.8x ROAS in Meta Ads Manager. Before cutting, they ran an incrementality test — pausing Instagram for 2 weeks in one region. Branded search conversions dropped 34% in that region. Instagram was generating demand; branded search was just capturing it. Last-click gave all the credit to search.
Use Claude to pressure-test your attribution: "Our last-click attribution shows SEO contributing 12% of revenue and branded search contributing 38%. But SEO likely drives branded search intent. How would you estimate SEO's true contribution? What test would prove it?"
Claude can help you design incrementality tests, interpret multi-touch attribution data, and build the business case for proper attribution — even if you can't run full MMM yet.
What it makes you do: Over-invest in branded search (the last click before purchase for almost every customer) and under-invest in top-of-funnel channels that generate demand in the first place (Instagram, YouTube, SEO, influencer).
A real example: A DTC skincare brand was about to cut their Instagram budget because Instagram showed only 1.8x ROAS in Meta Ads Manager. Before cutting, they ran an incrementality test — pausing Instagram for 2 weeks in one region. Branded search conversions dropped 34% in that region. Instagram was generating demand; branded search was just capturing it. Last-click gave all the credit to search.
Use Claude to pressure-test your attribution: "Our last-click attribution shows SEO contributing 12% of revenue and branded search contributing 38%. But SEO likely drives branded search intent. How would you estimate SEO's true contribution? What test would prove it?"
Claude can help you design incrementality tests, interpret multi-touch attribution data, and build the business case for proper attribution — even if you can't run full MMM yet.
Before cutting any top-of-funnel budget based on last-click ROAS, ask Claude to model what branded search would look like without that channel. The answer will change your decision.
Module 2 — Lab Exercise
Lab: Analyse a Google Ads Campaign
🔬 Lab Exercise
Full Google Ads Campaign Audit
Using the data below (or your own real campaign data), conduct a complete Google Ads performance audit with Claude. This mirrors exactly what a senior performance marketer would do.
1
Paste the campaign data table from Slide 2 into Claude. Ask it to calculate blended ROAS and rank campaigns. Identify the top performer and worst performer and explain why.
2
Ask Claude: "Generic-TV has 1.8% CTR. The industry benchmark for electronics on Google Search is 3.2%. What are the 5 most likely reasons for this gap, and what would you test first?"
3
Ask Claude to build a budget reallocation recommendation: "If we have ₹5,00,000 to reallocate from underperforming to overperforming campaigns, where should it go and what revenue uplift should we expect?"
4
Ask Claude to write a 150-word performance summary suitable for presenting to a CMO — what's working, what's not, and one clear action to take this week.
Module 2 — Campaign Performance
Key Distinctions — Campaign Analytics
ROAS vs ROI
ROAS = Revenue / Ad Spend. ROI = (Revenue - Total Cost) / Total Cost. A campaign with 4x ROAS may have negative ROI if COGS + fulfilment + ad spend exceeds revenue. Always tell Claude your gross margin so it can calculate true ROI, not just ROAS.
CTR vs Hook Rate
CTR (Click-Through Rate) applies to all ad formats — it measures clicks/impressions. Hook Rate is specific to video ads — it measures what % of viewers watched the first 3 seconds. Both tell you about creative appeal, but at different stages. A high Hook Rate with low CTR means the opening grabbed attention but the call-to-action failed.
Average ROAS vs Marginal ROAS
Average ROAS = total revenue / total spend. Marginal ROAS = incremental revenue from next ₹1 of spend. A campaign with 6x average ROAS might have 2x marginal ROAS if it's near audience saturation. Claude can model this distinction if you provide historical spend-vs-revenue curves.
Statistical Significance vs Practical Significance
A result can be statistically significant (unlikely to be random) but practically insignificant (the effect size is too small to matter). A 0.1% improvement in CVR at 95% confidence is real but not worth acting on unless you have massive volume. Ask Claude for both.
Module 2 — Quiz
Quiz — Campaign Performance (8 Questions)
Q1. A campaign has good impressions but low CTR. The primary issue is:
A. Budget — not enough spend
B. Targeting — wrong audience
C. Creative or ad copy — people see the ad but don't click it
D. Landing page — people arrive but don't convert
Q2. ROAS = 4x but the brand has 30% gross margin and 10% fulfilment cost. The campaign is:
A. Highly profitable
B. Marginally profitable or breakeven — Revenue × 40% (margin - fulfilment) must exceed ad spend. At 4x ROAS, profit margin is 40% × Revenue - 25% of Revenue = 15% profit margin.
C. Unprofitable — ROAS below 5x is always a loss
D. Cannot determine without more data
Q3. A/B test: Creative A 2.1% CTR, Creative B 2.8% CTR, after 2 days and 200 clicks each. You should:
A. Declare B the winner and scale immediately
B. Continue the test — 200 clicks per variant is insufficient for statistical significance. Need minimum 500-1,000 conversions per variant for reliable results.
C. Pause A immediately
D. Ask Claude to pick the winner
Q4. Hook Rate of 22% on a video ad means:
A. 22% of people who saw the ad clicked it
B. 22% of people who saw the ad watched the first 3 seconds — the opening is not compelling enough to stop the scroll. Industry benchmark is typically 30%+.
C. 22% of people converted after watching
D. The ad reached 22% of the target audience
Q5. Last-click attribution systematically causes brands to:
A. Over-invest in top-of-funnel channels like Instagram and YouTube
B. Over-invest in bottom-of-funnel channels like branded search, because they capture the last click before purchase regardless of what drove the customer's intent
C. Equally distribute budget across all channels
D. Under-invest in email marketing
Q6. Revenue Per Email (RPE) is highest for which campaign type typically?
A. Weekly newsletter to full list
B. Abandoned cart and VIP/high-value customer segments — they have higher purchase intent and personalisation, driving higher open rates, click rates, and conversion rates
C. Product review requests
D. New arrival announcements
Q7. Marginal ROAS differs from Average ROAS because:
A. Marginal ROAS only counts new customers
B. Marginal ROAS measures the return on the next incremental ₹1 of spend — as campaigns saturate, marginal ROAS falls even if average ROAS remains high
C. Marginal ROAS includes all costs, not just ad spend
D. They are calculated using different attribution windows
Q8. When diagnosing a ROAS drop, the correct order to check is:
A. Creative → CTR → CVR → Budget
B. Impressions (reach) → CTR (creative) → CVR (landing page/offer) → CPC (bidding/competition) — working down the funnel from top to bottom
C. Revenue → ROAS → CPC → Impressions
D. Budget → Targeting → Creative → Landing page
Module 2 — Campaign Performance
Google Ads Quality Score Deep Dive
Quality Score is Google's 1-10 rating of your ad relevance. It directly determines your ad rank and CPC — a Quality Score of 8 gets you the same position as a competitor with Quality Score 5, at 40% lower cost. Claude helps you diagnose and fix it systematically.
Quality Score Formula
Ad Rank = Quality Score × Max CPC × Expected Impact of Extensions
Quality Score = Expected CTR (40%) + Ad Relevance (40%) + Landing Page Experience (20%)
A QS of 10 vs QS of 5 means you pay approximately 50% less per click for the same position.
Quality Score = Expected CTR (40%) + Ad Relevance (40%) + Landing Page Experience (20%)
A QS of 10 vs QS of 5 means you pay approximately 50% less per click for the same position.
Quality Score Diagnostic Prompt
Diagnose Quality Score issues for these ad groups: Ad Group | Keywords | QS | Expected CTR | Ad Relevance | LP Experience Brand-Exact | [brand terms] | 9 | Above avg | Above avg | Above avg Running Shoes | running shoes | 5 | Below avg | Average | Below avg Trail Running | trail running | 4 | Below avg | Below avg | Below avg Nike Air Max | nike air max | 7 | Average | Above avg | Average Budget Running | cheap running | 3 | Below avg | Below avg | Below avg 1. Which component is dragging QS most for each ad group? 2. "Trail Running" has low ad relevance — what does this mean? 3. "Budget Running" has low LP experience — what 5 things should we check? 4. If we improve QS from 4 to 7 on Trail Running (current CPC ₹85), what is the estimated new CPC?
The QS CPC Relationship
CPC you pay = (Competitor Ad Rank / Your QS) + ₹0.01. So a competitor with Ad Rank 40 against your QS of 5 means you pay ₹8.01. Improve QS to 8 and you pay ₹5.01 for the same position — 37% cheaper. This is why QS improvement is the highest-ROI Google Ads optimisation.
Module 2 — Campaign Performance
Meta Ads Frequency & Audience Fatigue
Frequency is the average number of times a person has seen your ad. Too low and you don't build recall. Too high and you waste budget on people who've already decided. Claude helps you find the sweet spot and identify fatigue before it destroys ROAS.
Frequency Fatigue Prompt
Analyse audience fatigue across Meta ad sets (last 30 days): Ad Set | Frequency | CTR | CPM | ROAS | Audience Size Lookalike-1% | 2.1 | 2.8% | ₹185 | 3.8x | 4,50,000 Lookalike-2-5% | 1.8 | 3.1% | ₹175 | 4.2x | 18,00,000 Interest-Fitness | 4.2 | 1.9% | ₹225 | 2.4x | 8,00,000 Interest-Nutrition | 5.8 | 1.4% | ₹268 | 1.7x | 6,00,000 Retargeting-30day | 8.4 | 4.1% | ₹310 | 5.1x | 85,000 Broad-18-35 | 2.4 | 2.5% | ₹195 | 3.2x | 45,00,000 1. Which ad sets show clear fatigue signals (high freq + falling CTR)? 2. Interest-Nutrition at frequency 5.8 — should we pause or refresh? 3. What is the optimal frequency range for prospecting vs retargeting? 4. Recommend: new creative, new audience, or pause for each ad set 5. At what frequency should we automatically trigger a creative refresh alert?
🌍 Real World — How Mamaearth Manages Frequency
Mamaearth, one of India's fastest-growing D2C brands, sets automatic frequency caps at 3.5 for prospecting and 7 for retargeting. When either cap is hit, their team rotates creatives within 48 hours. This discipline keeps their CPMs stable and ROAS predictable even at scale.
Module 2 — Campaign Performance
LinkedIn Ads B2B Analytics
LinkedIn is the dominant B2B advertising platform — but it's expensive (CPCs of ₹500-2,000) and requires different success metrics than B2C platforms. Claude helps you evaluate LinkedIn performance correctly and justify the premium CPCs to stakeholders.
LinkedIn B2B Campaign Analysis Prompt
Analyse LinkedIn Ads performance for our SaaS product (B2B): Product: Project management tool, ₹8,999/month, ACV ₹1,07,988 Campaign | Spend | Impressions | Clicks | CTR | CPL | Leads | MQL% Sponsored Content | ₹1,20,000| 2,85,000 | 855 | 0.3% | ₹2,400 | 50 | 44% Message Ads | ₹80,000 | 95,000 | 2,850 | 3.0% | ₹1,600 | 50 | 38% Lead Gen Forms | ₹95,000 | 1,85,000 | 4,625 | 2.5% | ₹950 | 100 | 28% Retargeting | ₹40,000 | 65,000 | 1,300 | 2.0% | ₹800 | 50 | 65% 1. Calculate Cost per MQL for each campaign 2. Sponsored Content has highest CPL but 44% MQL rate — is it best value? 3. Lead Gen Forms has lowest CPL but 28% MQL — what explains this? 4. If average deal close rate is 18% and ACV ₹1,07,988, which campaign has the best pipeline ROI?
B2B LinkedIn Benchmarks
For SaaS B2B: CTR 0.3-0.5% (Sponsored Content), CPL ₹1,500-5,000, MQL rate 25-50%. The right metric is Cost per Pipeline Opportunity, not Cost per Lead. Always tell Claude the ACV and close rate so it can calculate true pipeline ROI.
Module 2 — Campaign Performance
Seasonal Campaign Analysis
Seasonality affects every business — but the pattern is different for each. Claude helps you identify your brand's seasonal patterns, predict peak periods, and plan budget allocation in advance rather than reacting.
Seasonal Analysis Prompt
Analyse seasonal patterns and plan next year's budget for our fashion e-commerce brand. Monthly data (last 2 years): Month | FY23 Revenue | FY24 Revenue | FY23 Spend | FY24 Spend | FY23 ROAS | FY24 ROAS Jan | ₹42L | ₹58L | ₹8L | ₹11L | 5.3x | 5.3x Feb | ₹38L | ₹52L | ₹7L | ₹9L | 5.4x | 5.8x Mar | ₹45L | ₹61L | ₹9L | ₹12L | 5.0x | 5.1x Apr | ₹35L | ₹48L | ₹8L | ₹10L | 4.4x | 4.8x May | ₹28L | ₹38L | ₹7L | ₹9L | 4.0x | 4.2x Jun | ₹32L | ₹44L | ₹8L | ₹10L | 4.0x | 4.4x Jul | ₹38L | ₹52L | ₹9L | ₹11L | 4.2x | 4.7x Aug | ₹42L | ₹58L | ₹10L | ₹13L | 4.2x | 4.5x Sep | ₹48L | ₹65L | ₹11L | ₹14L | 4.4x | 4.6x Oct | ₹82L | ₹1.1Cr | ₹18L | ₹24L | 4.6x | 4.6x Nov | ₹95L | ₹1.3Cr | ₹22L | ₹28L | 4.3x | 4.6x Dec | ₹68L | ₹92L | ₹15L | ₹20L | 4.5x | 4.6x 1. Identify peak, shoulder, and off-peak periods 2. ROAS is consistently lower in peak months — why? 3. Build FY25 budget recommendation by month (total budget ₹1.8Cr) 4. Which month offers best ROAS efficiency? Should we spend more there?
Module 2 — Campaign Performance
Competitor Spend Intelligence
Understanding what your competitors spend on advertising helps you size your own budget and anticipate ROAS changes. Claude helps you use publicly available signals — auction insights, SEMrush/SimilarWeb data — to build a competitive intelligence picture.
Competitor Intelligence Prompt
Build a competitive ad spend intelligence report using this data: (From Google Auction Insights + SEMrush estimates) Competitor | Impression Share | Overlap Rate | Outrank% | Est Monthly Spend | Top Keywords Our Brand | 48% | - | - | ₹8L | [our brand] Competitor A | 72% | 68% | 45% | ₹22L (est) | [category terms] Competitor B | 35% | 42% | 28% | ₹9L (est) | [niche terms] Competitor C | 28% | 31% | 19% | ₹6L (est) | [budget terms] 1. Competitor A has 72% impression share vs our 48% — what does this mean? 2. We are being outranked 45% of the time by Competitor A — should we increase bids or improve Quality Score? 3. Where is our impression share being lost — to budget or to rank? 4. What budget increase would get us to 60% impression share?
🌍 Real World — How Razorpay Monitors Competition
Razorpay's growth team monitors Google Auction Insights weekly to track when Paytm or PayU increases search spend (usually before major product launches or seasonal campaigns). This early warning allows them to defend their impression share proactively. The same insight is available to any Google Ads advertiser for free.
Module 2 — Campaign Performance
Cross-Channel Performance Dashboard
Most brands run 5-8 marketing channels simultaneously. Without a unified view, budget decisions are made in silos. Claude helps you build a cross-channel performance dashboard that tells the complete story in one prompt.
Cross-Channel Dashboard Prompt
Build a unified marketing performance dashboard — Month of August. Business: D2C supplements brand | Target blended ROAS: 3.5x Gross Margin: 58% | Total Revenue Target: ₹1.5Cr Channel | Spend | Revenue | ROAS | New Cust | CAC | Notes Google Search | ₹2,80,000| ₹11,20,000 | 4.0x | 142 | ₹1,972 | Brand heavy Google Shopping | ₹1,60,000| ₹8,00,000 | 5.0x | 98 | ₹1,633 | Top performer Meta Prospecting | ₹2,40,000| ₹4,80,000 | 2.0x | 285 | ₹842 | New customers Meta Retargeting | ₹80,000 | ₹4,80,000 | 6.0x | 28 | ₹2,857 | Limited audience Email/CRM | ₹40,000 | ₹8,00,000 | 20x | 0 | - | Existing base Influencer | ₹1,20,000| ₹2,40,000 | 2.0x | 156 | ₹769 | New customers SEO (est) | ₹60,000 | ₹6,00,000 | 10x | 84 | ₹714 | Low CAC Total Spend: ₹9,80,000 | Total Revenue: ₹45,20,000 1. What is blended ROAS? Are we hitting target? 2. Which channel has best CAC for new customer acquisition? 3. We need ₹15L more revenue next month. Where should the extra budget go? 4. Write a 5-bullet CMO summary of this month's performance
Slide 1 of 19
Slide 1 of 13
Use Next → to navigate
Module 3 — Segmentation & Cohorts
Customer Segmentation with Claude
The biggest insight in modern marketing: not all customers are equal, and treating them equally is the fastest way to destroy both margins and loyalty. Claude helps you segment, profile, and act on customer differences at scale.
🎯
RFM Segmentation
Recency, Frequency, Monetary — the gold standard for e-commerce customer segmentation. Claude calculates RFM scores and creates actionable segments.
📊
Cohort Analysis
Group customers by when they first purchased. See how retention, LTV, and behaviour differ across acquisition cohorts.
💰
LTV Modelling
Predict which customers will be most valuable over 12 months. Allocate acquisition budget based on predicted LTV, not just first-order value.
🔄
Churn Prediction
Identify customers who are about to churn before they actually do — and trigger the right retention intervention.
🌍 Real World — How Amazon Uses Segmentation
Amazon Prime customers spend 4.5x more than non-Prime customers annually. This insight came from cohort analysis — not from aggregate revenue reports. The same analytical approach, applied to your customer data using Claude, can identify your equivalent "high-value segment" and the marketing levers that convert customers into it.
Module 3 — Segmentation
RFM Analysis Using Claude
RFM (Recency, Frequency, Monetary) is the most actionable segmentation framework for any e-commerce or subscription business. Give Claude your customer transaction data and it will build your entire RFM model.
RFM Segmentation Prompt
Segment these customers using RFM analysis. Score each dimension 1-5 (5=best). Create 5 named segments. Data: Customer transactions last 12 months CustomerID | Last Purchase | # Orders | Total Spend C001 | 5 days ago | 12 | ₹48,000 C002 | 180 days ago | 2 | ₹3,200 C003 | 12 days ago | 8 | ₹28,000 C004 | 95 days ago | 1 | ₹1,800 C005 | 3 days ago | 18 | ₹95,000 C006 | 45 days ago | 5 | ₹18,500 C007 | 240 days ago | 1 | ₹900 [paste full dataset] Output: (1) RFM scores for each customer, (2) Segment they belong to, (3) Segment names with description, (4) Recommended action for each segment, (5) % of customers and % of revenue in each segment
The 5 Standard RFM Segments
Champions (555) — recent, frequent, high spend. Loyal Customers (X4X) — frequent buyers. At-Risk (411) — once high-value, haven't returned. Hibernating (111) — long gone, low value. Prospects (511) — recent but only one purchase. Each needs a completely different marketing strategy.
Module 3 — Segmentation
Cohort Retention Analysis
Cohort analysis answers the question every subscription and repeat-purchase business must answer: of customers who first bought in Month X, what % are still buying 3, 6, 12 months later? This is the true measure of product-market fit.
Cohort Analysis Prompt
Analyse customer retention by acquisition cohort (monthly). Show retention rate = % of cohort who purchased again. Cohort | M0 | M1 | M2 | M3 | M4 | M5 | M6 Jan | 1,200 | 38% | 24% | 18% | 14% | 12% | 11% Feb | 980 | 35% | 22% | 17% | 13% | 11% | - Mar | 1,450 | 42% | 28% | 22% | 17% | - | - Apr | 1,100 | 40% | 26% | 20% | - | - | - May | 1,380 | 44% | 30% | - | - | - | - Jun | 1,600 | 46% | - | - | - | - | - 1. Is retention improving or declining across cohorts? 2. The biggest drop is between M0 and M1 — what are the likely causes? 3. If we improve M1 retention from 38% to 50%, what is the 12-month revenue impact assuming AOV ₹1,500 and 1,200 customers/cohort? 4. Which cohort shows the healthiest retention? What happened that month?
🌍 Real World — How Spotify Uses Cohort Analysis
Spotify tracks cohort retention obsessively — what % of Month 1 listeners are still active in Month 6? When retention improved by just 2 percentage points one quarter, they traced it to a single product feature (personalised playlists). This is the power of cohort thinking — it connects product decisions to revenue outcomes.
💡 Insight Card — Segmentation
"Not All Customers Are Equal — And Treating Them Equally Is Costing You"
Here is a pattern that appears in almost every e-commerce business: 20% of customers generate 80% of revenue. And within that 20%, the top 5% are often the only reason the business is profitable at all.
The calculation most brands never do: What is the revenue per marketing email for your top 10% vs your bottom 50%? For a typical brand, top 10% customers generate 8-12x more revenue per email than bottom 50% customers — yet they receive the exact same emails at the same frequency.
What to do about it:
• Identify your Champions segment (high RFM) and give them VIP treatment — early access, exclusive products, personal outreach
• Identify your At-Risk high-value customers (dropped off recently) and trigger a personal win-back campaign
• Stop over-communicating with Hibernating low-value customers — the cost of unsubscribes and spam complaints outweighs the revenue
Ask Claude: "Given this RFM distribution, design a differentiated communication strategy for each segment — frequency, channel, offer type, and tone."
The calculation most brands never do: What is the revenue per marketing email for your top 10% vs your bottom 50%? For a typical brand, top 10% customers generate 8-12x more revenue per email than bottom 50% customers — yet they receive the exact same emails at the same frequency.
What to do about it:
• Identify your Champions segment (high RFM) and give them VIP treatment — early access, exclusive products, personal outreach
• Identify your At-Risk high-value customers (dropped off recently) and trigger a personal win-back campaign
• Stop over-communicating with Hibernating low-value customers — the cost of unsubscribes and spam complaints outweighs the revenue
Ask Claude: "Given this RFM distribution, design a differentiated communication strategy for each segment — frequency, channel, offer type, and tone."
Segmentation is not a reporting exercise. It is the bridge between data and personalised marketing action. Every segment needs a different campaign, not just a different subject line.
Module 3 — Segmentation
Customer LTV Modelling
LTV (Lifetime Value) is the most important number in marketing — because it tells you how much you can afford to spend acquiring a customer. Claude can model LTV, segment it, and connect it to your acquisition strategy.
LTV Calculation Prompt
Calculate Customer LTV and segment by acquisition channel: Business: D2C nutrition brand | Gross Margin: 52% Channel | Customers | AOV | Orders/yr | Retention | CAC Google Ads | 2,400 | ₹2,100 | 3.2 | 42% yr2 | ₹680 Meta Ads | 1,850 | ₹1,800 | 2.8 | 35% yr2 | ₹520 Influencer | 980 | ₹2,400 | 4.1 | 58% yr2 | ₹890 Organic SEO | 620 | ₹2,200 | 3.8 | 65% yr2 | ₹180 Referral | 440 | ₹2,600 | 4.8 | 71% yr2 | ₹240 Calculate for each channel: 1. Year 1 LTV = AOV × Orders × Gross Margin 2. Year 2 LTV = Year 1 LTV × Retention Rate 3. 2-Year LTV total 4. LTV:CAC ratio (target > 3x) 5. Which channel should get more budget? Rank by LTV:CAC.
Module 3 — Segmentation
Churn Analysis & Win-Back Strategy
For subscription businesses and repeat-purchase brands, churn is the silent killer. Claude helps you identify churning customers before they leave and design data-driven win-back campaigns.
Churn Analysis Prompt
Identify at-risk customers and design a win-back strategy. Context: SaaS product, monthly subscription ₹2,999/mo Churn signals (from product data): Behaviour | Churned Rate | Active Rate Login frequency < 2/week | 68% | 12% Feature usage dropped 50%+| 71% | 8% Support tickets > 3/month | 58% | 14% No login in 14 days | 82% | 2% Downgraded plan | 74% | 11% 1. Build a churn risk score (0-100) using these signals 2. Design a 3-step intervention for high-risk customers (score > 70) 3. What is the revenue impact of reducing churn from 8% to 6% monthly on a base of 5,000 subscribers at ₹2,999/mo? 4. Write a personalised win-back email for a customer who hasn't logged in 21 days and had 3 support tickets last month
Module 3 — Segmentation
Building Customer Personas with Claude
Data-driven personas are more actionable than gut-feel personas. Claude can build rich customer personas from your actual transaction, demographic, and behavioural data — personas that the entire team can use to make better creative and channel decisions.
Persona Building Prompt
Build 4 customer personas from this segmentation data. Brand: Premium skincare D2C, AOV ₹3,200 Segment | Age | Gender | City Tier | AOV | Freq | Top Category | Acquisition Champions | 28-38| 78% F | Tier 1 | ₹6,200| 6/yr | Anti-aging | Influencer Loyalists | 32-45| 71% F | Tier 1-2 | ₹3,800| 4/yr | Moisturisers | Google Search Occasionals|24-32| 65% F | Tier 1-3 | ₹1,900| 2/yr | Sunscreen/basic | Meta At-Risk | 35-50| 69% F | Tier 1 | ₹4,100| prev | Anti-aging | Influencer For each segment: 1. Name the persona (memorable, specific) 2. Write a 50-word persona description 3. Their primary motivation for buying skincare 4. Their biggest purchase barrier 5. Best channel and message to reach them 6. One campaign idea tailored to this persona
Module 3 — Lab
Lab: Build a Customer Segment Report
🔬 Lab Exercise
Full RFM Segmentation + Action Plan
Using your own customer data (or the sample data in this slide), build a complete RFM segmentation and derive an action plan using Claude. This is the most valuable exercise in this module.
1
Export your last 12 months of customer transaction data from your e-commerce platform (Shopify, WooCommerce, etc.) as a CSV. Columns needed: CustomerID, Order Date, Order Value. Paste into Claude and ask for RFM scores.
2
Ask Claude to name your 5 segments with marketing personas, calculate the size and revenue contribution of each, and identify which segment has the most untapped potential.
3
For your At-Risk segment (high historical value, haven't bought recently), ask Claude to write a personalised win-back email sequence — 3 emails, 7 days apart, with different angles (offer, emotional, exclusivity).
4
Build an LTV model: ask Claude to calculate the revenue impact of moving 10% of your Occasionals segment into Loyalists. This is your "retention investment business case."
Module 3 — Key Distinctions
Key Distinctions — Segmentation
RFM vs Demographic Segmentation
Demographic segmentation (age, gender, location) describes who customers are. RFM segmentation describes how they behave. Behavioural segmentation is always more predictive of future purchases than demographic data. Ask Claude to segment by RFM first, then overlay demographics to understand who is in each segment.
Average LTV vs Segment LTV
A brand's average LTV of ₹8,500 might mask Champions with ₹45,000 LTV and Hibernating customers with ₹900 LTV. Blended averages hide the insight. Always ask Claude to calculate LTV by segment, acquisition channel, and product category.
Cohort Retention vs Overall Retention
Overall retention rate = returning customers / total customers. This is distorted by growth — a fast-growing business always has low overall retention even if individual cohorts retain well. Cohort retention measures the true health of customer relationships.
Churn vs Dormancy
A churned customer has actively cancelled. A dormant customer has simply stopped buying without telling you. For e-commerce, dormancy is the equivalent of churn. Define your dormancy threshold (e.g., no purchase in 90 days) and ask Claude to size the at-risk revenue.
Module 3 — Quiz
Quiz — Segmentation & Cohorts (8 Questions)
Q1. RFM stands for:
A. Reach, Frequency, Message
B. Recency (last purchase date), Frequency (number of purchases), Monetary (total spend) — the three behavioural dimensions that predict future customer value
C. Revenue, Funnel, Marketing
D. Retention, Frequency, Margin
Q2. A customer with RFM score 511 (high recency, low frequency, high monetary on first order) is best classified as:
A. A Champion — buy everything
B. A Prospect — recently acquired, made a large first purchase, but hasn't returned yet. Priority: trigger second purchase quickly.
C. At-Risk — about to churn
D. Loyal Customer — reliable repeat buyer
Q3. A cohort shows M0: 1,000 customers, M1: 38% retention, M2: 24%. What is the absolute number of customers still active at M2?
A. 240 customers
B. 240 customers — 1,000 × 24% = 240. Note: this assumes M2 retention is % of original cohort, not % of M1 survivors.
C. 380 customers
D. 91 customers
Q4. LTV:CAC ratio of 1.5x means:
A. Excellent — more than 1x is always profitable
B. Concerning — the business is earning only 1.5x what it costs to acquire customers. The benchmark for healthy unit economics is 3x or higher. Operating costs are not yet covered.
C. The business should scale immediately
D. CAC is too low and needs to be increased
Q5. Why is cohort retention more useful than overall retention rate for a growing business?
A. Cohort retention is easier to calculate
B. Overall retention is distorted by rapid growth — a fast-growing business always has low overall retention because new customers dominate the denominator. Cohort retention shows the true loyalty of each customer group.
C. Cohort retention includes more customers
D. Overall retention counts churned customers twice
Q6. The most valuable customer to target for a win-back campaign is:
A. All lapsed customers equally
B. At-Risk customers who were formerly Champions — they had high RFM scores historically but haven't purchased recently. They know the brand, have high LTV potential, and are most recoverable.
C. The most recently lapsed customers regardless of historical value
D. Customers who never bought but opened emails
Q7. Average LTV of ₹8,500 for a brand means:
A. Every customer will spend ₹8,500
B. The average masks segment differences — Champions may have ₹45,000 LTV while Hibernating customers have ₹900. Always ask Claude to calculate LTV by segment, not just overall average.
C. CAC must be below ₹8,500
D. The business is profitable if COGS is below ₹8,500
Q8. The biggest drop in most cohort retention curves is between M0 and M1. This suggests:
A. The product has a quality problem
B. The post-purchase experience and onboarding is failing to create habit or drive a second purchase. The fix is typically a first-purchase email sequence, product education, or a second-purchase offer within 30 days.
C. The acquisition cost is too high
D. The product is priced incorrectly
Module 3 — Segmentation
Behavioural Segmentation Beyond RFM
RFM tells you what customers bought and when. Behavioural segmentation tells you how they shop — device, time of day, category preference, promotion sensitivity. Claude helps you build multi-dimensional segments that drive truly personalised marketing.
Behavioural Segmentation Prompt
Build behavioural segments from this customer data (e-commerce brand): Segment Behaviour | % Customers | AOV | Conv Rate | Pref Channel | Discount Response Mobile-only shoppers | 42% | ₹1,200 | 1.8% | App/Mobile | High (28% lift) Desktop research-buy | 28% | ₹3,400 | 3.2% | Desktop | Low (4% lift) Discount hunters | 18% | ₹1,800 | 0.8% base | Email/Push | Very high (65% lift) Loyal full-price | 8% | ₹4,200 | 4.1% | Email | None needed Impulse-social | 4% | ₹1,500 | 2.9% | Instagram | Medium (18% lift) 1. Which segment is most profitable (Revenue - Discount cost)? 2. "Discount hunters" have 65% lift with offers — but at what margin cost? 3. Design a different marketing strategy for each behavioural segment 4. How should we avoid training "Loyal full-price" customers to expect discounts?
🌍 Real World — How Flipkart Segments Behaviourally
Flipkart's "Big Billion Days" strategy evolved from mass discounting to surgical behavioural targeting — price-sensitive customers get bigger discounts, loyal customers get early access and exclusivity, lapsed customers get win-back offers. Result: same promotional budget, 35% higher net revenue. Claude helps you build this same analytical foundation.
Module 3 — Segmentation
Geographic Segmentation Analysis
Geography is one of the most underused segmentation dimensions in Indian marketing. Tier 1 vs Tier 2 vs Tier 3 customers have fundamentally different AOVs, channel preferences, and product affinities. Claude helps you decode the geographic opportunity.
Geographic Analysis Prompt
Analyse geographic performance for our beauty brand (last 6 months): City Tier | Orders | Revenue | AOV | Top Products | Preferred Channel | Return Rate Tier 1 (6 cities) | 18,400 | ₹8.3Cr | ₹4,511 | Premium skincare | Website/App | 12% Tier 2 (28 cities) | 22,100 | ₹6.2Cr | ₹2,805 | Mass + mid-premium | Website | 8% Tier 3 (120+ cities)| 14,200 | ₹2.8Cr | ₹1,972 | Entry-level | COD preferred| 14% International | 1,800 | ₹1.1Cr | ₹6,111 | Premium/gift sets | Website | 6% 1. Which tier has best unit economics (AOV vs return rate)? 2. Tier 3 has 14% return rate — top 3 likely causes and solutions 3. How should product assortment differ by tier? 4. We want to grow Tier 2 by 40% — what marketing strategy changes? 5. International at ₹6,111 AOV — should we invest more there?
Module 3 — Segmentation
NPS Analysis with Claude
Net Promoter Score (NPS) is the gold standard for measuring customer loyalty — and the verbatim comments from NPS surveys are among the most valuable marketing research you can do. Claude can analyse hundreds of NPS responses in minutes.
NPS Analysis Prompt
Analyse our NPS survey results (last quarter, n=482): Promoters (9-10): 198 | Passives (7-8): 168 | Detractors (0-6): 116 NPS = (198-116)/482 × 100 = 17 Sample verbatim comments: PROMOTERS: "Fast delivery, exactly as described", "Best quality I've found", "Customer service resolved my issue in minutes", "Love the subscription model" PASSIVES: "Good but slightly overpriced", "Delivery took 5 days, expected 3", "Product is fine, nothing exceptional", "App crashes sometimes" DETRACTORS: "Wrong item delivered, took 2 weeks to resolve", "Quality worse than photos", "Impossible to cancel subscription", "No response from support for 4 days" 1. What is our NPS? Is 17 good or bad for our category? 2. Categorise all comments into themes (delivery, quality, service, product, tech) 3. Which theme has the most Detractor mentions? This is the #1 priority fix. 4. Write 3 specific actions to move Passives to Promoters 5. What marketing claims should we amplify based on Promoter feedback?
Module 3 — Segmentation
Customer Journey Mapping with Claude
A customer journey map visualises every touchpoint from awareness to advocacy. Claude helps you build data-driven journey maps that reveal where customers drop off, what messaging works at each stage, and where to invest next.
Customer Journey Analysis Prompt
Map the customer journey for our fitness app (subscription ₹999/month). Data from 1,000 paying subscribers (survey + analytics): Stage | How they found us | % each path | Time in stage | Key action Awareness | Instagram Reel | 38% | 1-3 days | Saved/shared content | Google Search | 28% | same day | Clicked ad or organic | Friend referral | 22% | immediate | Downloaded app | YouTube | 12% | 2-5 days | Subscribed to channel Consideration | Free trial started | 100% | 3-14 days | Used 2+ features | Abandoned trial | 62% | Day 3-5 | Stopped logging in Decision | Converted to paid | 38% | Day 7-14 | Credit card added Retention | Still active Month 3 | 58% | ongoing | 4+ sessions/week Advocacy | Referred a friend | 22% | Month 2-4 | Used referral code 1. Where is the biggest opportunity — improving trial conversion or retention? 2. 62% abandon trial — what interventions at Day 3-5 could reduce this? 3. Instagram has 38% of awareness but what % of paid conversions? 4. Design a personalised message for each journey stage
Slide 1 of 14
Slide 1 of 7
Use Next to navigate
Module 4 — SEO & Content Analytics
SEO Analytics with Claude
SEO is the highest-LTV acquisition channel for most businesses — customers who find you via organic search have lower CAC, higher intent, and better retention than paid channels. Claude helps you analyse SEO performance, find opportunities, and prioritise content investments.
🔍
Keyword Analysis
Find high-opportunity keywords your competitors rank for that you don't. Claude identifies gaps and prioritises by traffic potential.
📄
Content Audit
Which pages are driving organic revenue? Which are ranking on page 2 and need a push? Claude audits your content performance data.
🏆
Competitor Analysis
What content strategy is driving your competitor's organic growth? Claude analyses competitor data and identifies replicable patterns.
Real World — How HubSpot Uses Content Analytics
HubSpot generates 60%+ of its leads from organic search. Their content team uses analytics to identify "decaying content" — posts that once ranked well but are losing position. They refresh these with Claude-assisted analysis, recovering rankings in weeks. This is a repeatable process any team can implement.
Module 4 — SEO Analytics
Keyword Opportunity Analysis
Not all keywords are equal. The best SEO opportunities combine commercial intent, reasonable competition, and meaningful search volume. Claude helps you find these opportunities in your niche.
Keyword Opportunity Prompt
Analyse these keywords for our fitness equipment brand (India). We currently rank for keywords marked with our position. Keyword | Vol/mo | Difficulty | Our Rank | CPC "home gym equipment" | 22,000 | 65 | 8 | ₹45 "adjustable dumbbells" | 18,000 | 58 | 4 | ₹38 "resistance bands" | 45,000 | 72 | 15 | ₹22 "protein powder" | 85,000 | 82 | Not ranked| ₹65 "yoga mat india" | 28,000 | 55 | 2 | ₹18 "home workout equipment" | 15,000 | 48 | 12 | ₹40 "kettlebell 20kg" | 8,500 | 42 | 6 | ₹52 "treadmill under 30000" | 12,000 | 60 | Not ranked| ₹95 1. Rank keywords by opportunity score (Vol × (1-Difficulty) × Intent) 2. Which keywords are on page 2 (rank 11-20) — quick wins with content update? 3. Which keywords are too competitive given our current authority? 4. Suggest 5 long-tail variants of our best opportunity keywords
Module 4 — SEO Analytics
Content Performance Audit
Most brands have blog posts and landing pages that were once valuable but are now decaying — losing rankings, traffic, and conversions. Claude helps you identify and prioritise which content to refresh for maximum impact.
Content Audit Prompt
Audit our blog content performance (Google Search Console + GA4 data): URL | Impressions | Clicks | CTR | Avg Pos | Sessions | Conv /blog/home-gym-guide | 45,000 | 1,800 | 4.0%| 5.2 | 1,650 | 28 /blog/protein-myths | 28,000 | 560 | 2.0%| 12.1 | 490 | 4 /blog/resistance-training | 18,000 | 1,080 | 6.0%| 3.8 | 980 | 41 /blog/yoga-for-beginners | 65,000 | 975 | 1.5%| 8.4 | 850 | 6 /blog/best-dumbbells-2022 | 32,000 | 640 | 2.0%| 11.3 | 580 | 12 /blog/morning-workout | 12,000 | 840 | 7.0%| 2.9 | 790 | 35 1. Which posts are quick-win opportunities (good impressions, poor CTR)? 2. Which posts are on the cusp of page 1 (position 8-15)? 3. Which posts convert best — use this to guide future content creation 4. The "best-dumbbells-2022" post — why is it underperforming and what fix?
Insight Card — SEO Analytics
"Traffic Is Vanity, Conversion Is Sanity"
The most common mistake in SEO content strategy: optimising for traffic volume instead of commercial intent.
A blog post ranking #1 for "what is a dumbbell?" might get 50,000 sessions per month but convert at 0.1%. A product comparison post ranking #4 for "best dumbbells under ₹5,000" might get 3,000 sessions but convert at 4%.
Revenue from Post 1: 50,000 × 0.1% × ₹4,000 AOV = ₹2,00,000
Revenue from Post 2: 3,000 × 4% × ₹4,000 AOV = ₹4,80,000
The lower-traffic post generates 2.4x more revenue.
Use Claude to calculate revenue per session for every content piece: "Given these content performance metrics and my 2% average conversion rate and ₹4,000 AOV, calculate revenue per session for each blog post. Rank by revenue contribution, not traffic."
This single reframe will change your entire content strategy.
A blog post ranking #1 for "what is a dumbbell?" might get 50,000 sessions per month but convert at 0.1%. A product comparison post ranking #4 for "best dumbbells under ₹5,000" might get 3,000 sessions but convert at 4%.
Revenue from Post 1: 50,000 × 0.1% × ₹4,000 AOV = ₹2,00,000
Revenue from Post 2: 3,000 × 4% × ₹4,000 AOV = ₹4,80,000
The lower-traffic post generates 2.4x more revenue.
Use Claude to calculate revenue per session for every content piece: "Given these content performance metrics and my 2% average conversion rate and ₹4,000 AOV, calculate revenue per session for each blog post. Rank by revenue contribution, not traffic."
This single reframe will change your entire content strategy.
Rank your content by revenue per session, not page views. Then invest your content team's time in creating more of what converts, not what just gets clicks.
Module 4 — SEO Analytics
Competitor Gap Analysis
Your competitors have already done the work of discovering what content drives organic revenue in your industry. Claude helps you analyse their strategy and identify the gaps you should fill.
Competitor Gap Prompt
Identify content gaps vs our top competitor (data from SEMrush/Ahrefs export). Our brand: FitZone. Competitor: HealthKart. HealthKart ranks for (we do NOT rank for): Keyword | Volume | HealthKart Pos "whey protein benefits" | 35,000 | 3 "creatine dosage" | 28,000 | 5 "pre workout india" | 22,000 | 2 "bcaa vs eaa" | 15,000 | 4 "intermittent fasting guide" | 42,000 | 6 "gym diet plan india" | 38,000 | 3 We rank for (HealthKart does NOT): "home gym setup" | 18,000 | 4 "resistance band workout" | 22,000 | 3 1. Which competitor keywords should we target first? Rank by priority. 2. What type of content does HealthKart create for these keywords? 3. Our strength is home fitness. How do we use this angle to compete on HealthKart's strong keywords? 4. Suggest 3 content ideas that combine our home fitness strength with their high-volume nutrition keywords.
Module 4 — Lab
Lab: SEO Performance Analysis
Lab Exercise
Complete SEO Content Audit
Using Google Search Console data (free, available to any website owner), conduct a complete content audit with Claude in under 30 minutes.
1
Go to Google Search Console → Performance → Pages. Export last 3 months data. Paste into Claude with the prompt: "Audit this content — identify quick wins (good impressions, poor CTR), page 2 opportunities (position 8-15), and dead content (declining impressions YoY)."
2
For your top 5 ranking pages, ask Claude: "Calculate revenue per session assuming [your CVR] and [your AOV]. Which page is most undermonetised relative to its traffic?"
3
Take your weakest-performing content piece (high impressions, low CTR). Ask Claude: "Suggest 5 alternative title tags and meta descriptions that would improve CTR for this page, based on the target keyword [X]."
Module 4 — Quiz
Quiz — SEO & Content Analytics (6 Questions)
Q1. A blog post has 65,000 impressions, 975 clicks, and position 8.4. The priority action is:
A. Delete it — too many impressions with low clicks
B. Improve CTR — good impressions show strong search demand, position 8 is top of page 2, better title tags and meta descriptions could double clicks.
C. Increase its word count immediately
D. Build backlinks to improve position
Q2. Content ranking position 12 for a keyword with 18,000 monthly searches is best described as:
A. A failure — not on page 1
B. A quick-win opportunity — position 12 is just off page 1. Content improvement or link building could move it to page 1 and 5-10x its traffic with relatively low effort.
C. Too competitive to pursue
D. An average performer with no action needed
Q3. "Revenue per session" as a content metric is more useful than "pageviews" because:
A. It is easier to calculate
B. It connects content performance to business outcomes — high-traffic informational content often generates less revenue than lower-traffic commercial intent content. This metric correctly prioritises investment.
C. Google Analytics tracks it automatically
D. It is the metric Google uses for ranking
Q4. A competitor gap keyword is most worth targeting when:
A. It has the highest search volume regardless of difficulty
B. It has commercial intent aligned with your product, reasonable difficulty relative to your domain authority, and your competitor's content is outdated or low quality
C. Your competitor ranks #1 for it
D. It contains your brand name
Q5. A blog post titled "Best Dumbbells 2022" is underperforming in 2025 because:
A. Dumbbells are no longer popular
B. The outdated year in the title and content signals to both users and Google that the content is stale. Users see "2022" in the SERP and skip it. The fix: update year, refresh content, update URL if possible.
C. The keyword is too competitive
D. It should be targeting a different audience
Q6. CTR of 1.5% for a keyword at position 8 in SEO means:
A. Performance is excellent
B. CTR is below expected (~3-4% at position 8) — the title tag or meta description is not compelling enough. Testing different title formats can significantly improve clicks without improving ranking.
C. The keyword is too broad
D. There is a technical SEO error
Module 4 — SEO & Content Analytics
Core Web Vitals Analysis
Core Web Vitals are Google's user experience metrics that directly impact search rankings. LCP (loading), FID/INP (interactivity), and CLS (visual stability) — poor scores cost you organic visibility. Claude helps you interpret these metrics and prioritise fixes.
Core Web Vitals Analysis Prompt
Analyse Core Web Vitals for our e-commerce site (Google Search Console data): Page Type | LCP | INP | CLS | Mobile Pass% | Desktop Pass% Homepage | 3.2s | 180ms | 0.12 | 62% | 88% Product Pages | 4.8s | 245ms | 0.28 | 38% | 72% Category Pages | 3.9s | 195ms | 0.15 | 55% | 81% Blog Posts | 2.8s | 120ms | 0.05 | 78% | 95% Checkout | 5.2s | 320ms | 0.08 | 28% | 68% Benchmarks: LCP good = <2.5s | INP good = <200ms | CLS good = <0.1 1. Which pages are failing Core Web Vitals and at what risk to rankings? 2. Product Pages have CLS of 0.28 — what typically causes high CLS? 3. Checkout at 5.2s LCP — what is the likely revenue impact? 4. Prioritise fixes by: SEO impact × revenue impact 5. What 5 technical changes would have highest impact on LCP?
LCP and Revenue Connection
Every 1-second delay in page load reduces conversions by approximately 7% (Google data). For a checkout page with ₹50L monthly revenue, improving LCP from 5.2s to 2.5s could recover ₹10-12L in monthly revenue. Always frame Core Web Vitals improvements in revenue terms for stakeholder buy-in.
Module 4 — SEO & Content Analytics
Local SEO Analytics
For businesses with physical locations — retail, restaurants, clinics, real estate — local SEO is the highest-intent acquisition channel. Claude helps you analyse Google Business Profile performance and identify local search opportunities.
Local SEO Analysis Prompt
Analyse Google Business Profile performance for our 5 store locations: Location | Views | Searches | Direction Req | Calls | Website Clicks | Reviews | Avg Rating Mumbai-BKC | 8,400 | 3,200 | 680 | 420 | 850 | 284 | 4.6 Mumbai-Andheri| 5,200 | 1,980 | 320 | 195 | 420 | 142 | 4.2 Delhi-CP | 6,800 | 2,600 | 540 | 380 | 680 | 198 | 4.8 Bangalore-MG | 7,100 | 2,800 | 610 | 440 | 720 | 312 | 4.7 Pune-FC Road | 3,100 | 1,200 | 180 | 110 | 240 | 68 | 3.9 1. Which location has best conversion (views → actions)? 2. Mumbai-Andheri has 4.2 rating vs Delhi-CP at 4.8 — revenue impact? 3. Pune-FC Road has only 68 reviews — what is the priority action? 4. How do we improve "Searches" → "Direction Requests" conversion? 5. Write a response template for negative reviews that protects brand reputation
Module 4 — SEO & Content Analytics
Backlink Analysis with Claude
Backlinks remain one of the strongest ranking signals. Understanding your backlink profile — and your competitors' — helps you identify link building opportunities and diagnose ranking drops. Claude interprets backlink data to guide your strategy.
Backlink Analysis Prompt
Analyse our backlink profile vs top competitor (Ahrefs/SEMrush data): Metric | Our Site | Competitor A | Gap Domain Authority | 38 | 54 | -16 Total Backlinks | 4,200 | 18,400 | -14,200 Referring Domains | 380 | 1,240 | -860 Dofollow Links | 68% | 72% | -4% Avg Link DR | 32 | 48 | -16 New Links (last 30d) | 12 | 85 | -73 Lost Links (last 30d) | 8 | 14 | - Top Link Sources (Competitor A that we don't have): - TechCrunch India (DR 82) - YourStory (DR 78) - Inc42 (DR 74) - ET Tech (DR 88) 1. Is the DA gap a serious problem for our target keywords? 2. Competitor gets 85 new links/month vs our 12 — what content drives theirs? 3. We are missing links from TechCrunch, YourStory, Inc42 — what type of content gets coverage from these outlets? 4. Build a 3-month link building strategy to close 30% of the gap
Module 4 — SEO & Content Analytics
Featured Snippet Optimisation
Featured snippets (Position 0) appear above all organic results — they get 8% of clicks on average but dominate voice search. For question-based queries in your niche, winning the featured snippet can double your organic visibility. Claude helps you identify and win these opportunities.
Featured Snippet Opportunity Prompt
Identify featured snippet opportunities from our rankings data: Keyword | Our Rank | Snippet Exists | Snippet Type | Our Snippet? "what is ROAS in marketing" | 4 | Yes (def box) | Definition | No "how to calculate CAC" | 2 | Yes (numbered) | List | No "meta ads vs google ads" | 6 | Yes (table) | Table | No "best time to send email" | 3 | Yes (para) | Paragraph | No "what is customer LTV" | 1 | No | - | - "marketing attribution models" | 5 | Yes (list) | List | No 1. Which keywords have the best featured snippet opportunity given our rank? 2. We rank #1 for "customer LTV" but no snippet — how do we create one? 3. "Meta ads vs google ads" needs a table snippet — what should the table include? 4. Write the optimised content format for "how to calculate CAC" to win the numbered list snippet (max 40 words, 5 steps)
Module 5 — Attribution & Paid Ads
Marketing Mix Modelling Basics
Marketing Mix Modelling (MMM) is the gold standard for measuring the true contribution of every marketing channel — including channels that attribution tools can't track like TV, OOH, and brand awareness spend. Claude helps you understand and apply MMM thinking even without a data science team.
What is MMM?
MMM uses statistical regression to model the relationship between all marketing inputs (spend by channel, pricing, distribution) and business outputs (revenue, sales volume). It answers: "What % of our revenue came from each channel, including the unmeasured ones?" It's how P&G, Unilever, and all major FMCG brands allocate budget.
MMM Thinking Prompt
Apply MMM thinking to our brand's revenue drivers (FMCG brand). Monthly data (18 months): Month | Revenue | Digital Spend | TV Spend | Promotion | Seasonal Index | Avg Price Jan | ₹4.2Cr | ₹28L | ₹45L | No promo | 0.82 | ₹299 Feb | ₹3.8Cr | ₹24L | ₹45L | No promo | 0.74 | ₹299 Mar | ₹5.1Cr | ₹32L | ₹55L | Yes | 0.91 | ₹269 [...] Oct | ₹8.2Cr | ₹48L | ₹65L | Yes | 1.42 | ₹269 Nov | ₹9.5Cr | ₹55L | ₹75L | Yes | 1.65 | ₹249 Using regression thinking, estimate: 1. What % of revenue is "base" (would exist without any marketing)? 2. What incremental revenue is driven by Digital vs TV? 3. What is the revenue impact of promotions? 4. If we cut TV by 30% and increase Digital by 30%, what happens to revenue?
Module 5 — Attribution & Paid Ads
UTM Tracking Setup & Analysis
UTM parameters are the foundation of accurate marketing attribution. Without clean UTM tracking, revenue gets mis-attributed to "Direct" and channel performance is invisible. Claude helps you audit your UTM structure and fix attribution gaps.
UTM Audit Prompt
Audit our UTM tracking quality. These are actual UTM parameters found in our GA4 (last month top sources): utm_source | utm_medium | utm_campaign | Sessions | Revenue | Issue? google | cpc | brand_aug24 | 12,400 | ₹8.2L | OK facebook | social | summer_sale | 8,200 | ₹4.1L | Medium unclear Instagram | (none) | (none) | 3,400 | ₹1.8L | Missing UTMs! email | Email | weekly_newsletter | 2,100 | ₹1.2L | Case inconsistent FACEBOOK | CPC | BRAND | 1,800 | ₹0.9L | Duplicate of fb! newsletter | email | aug_promo | 1,600 | ₹0.8L | Source/medium swapped (direct) | (none) | (none) | 9,200 | ₹5.1L | Untracked traffic 1. How much revenue is currently "lost" to untracked / miscategorised? 2. "FACEBOOK" and "facebook" are showing as different sources — impact? 3. Instagram has 3,400 sessions with no UTMs — what should the UTM be? 4. Build a UTM naming convention for our brand (all channels, formats) 5. How do we retroactively clean this data in GA4?
Module 5 — Attribution & Paid Ads
Cross-Device Attribution
A customer who sees your Instagram ad on mobile, researches on desktop, and buys on mobile again — how do you attribute that purchase? Cross-device attribution is one of the hardest problems in marketing analytics. Claude helps you understand the gap and practical solutions.
The Cross-Device Problem
The average Indian digital consumer uses 2.8 devices. Without cross-device attribution, the same customer journey appears as 3 separate users in your analytics. This inflates new visitor counts, deflates conversion rates, and mis-attributes revenue to mobile when the research happened on desktop.
Cross-Device Analysis Prompt
Analyse cross-device journey data from GA4 (User-ID enabled): Device Path | Sessions | Conversions | Revenue | % of Total Mobile only | 45,200 | 890 | ₹26.7L | 58% Desktop only | 18,400 | 620 | ₹31.0L | 40% Mobile → Desktop → Buy | 8,200 | 185 | ₹11.1L | 24% Desktop → Mobile → Buy | 3,800 | 98 | ₹4.9L | 13% Mobile → Mobile → Buy | 12,400 | 245 | ₹7.4L | 16% App → Website → Buy | 2,100 | 68 | ₹4.8L | 9% 1. Without User-ID, how would GA4 mis-count unique users? 2. Desktop-only has lower sessions but higher revenue — what does this mean? 3. "Mobile → Desktop → Buy" path — how do we optimise for this journey? 4. Our mobile conversion rate appears 1.8% but true rate may be higher — why?
Slide 1 of 11
Slide 1 of 7
Use Next to navigate
Module 5 — Attribution Modelling
Attribution Models Explained
Attribution is the most debated and most misunderstood topic in marketing analytics. How you attribute conversions determines where you invest budget — which means attribution errors directly cost you revenue. Claude helps you understand, choose, and use the right model for your business.
1️⃣
First-Touch
100% credit to the first touchpoint. Best for understanding awareness channels. Overvalues top-of-funnel.
🎯
Last-Touch
100% credit to the last touchpoint. Most common default. Overvalues bottom-of-funnel branded search.
➗
Linear
Equal credit to every touchpoint. Simple but ignores the different roles channels play in the funnel.
⏱️
Time-Decay
More credit to touchpoints closer to conversion. Good for short sales cycles with clear conversion moments.
🔢
Data-Driven
Machine learning assigns credit based on actual conversion path patterns. Most accurate but requires volume.
📊
MMM
Marketing Mix Modelling — statistical analysis of all marketing inputs vs revenue outputs. The gold standard but complex.
Real World — How Airbnb Handles Attribution
Airbnb uses a combination of data-driven attribution and incrementality testing. They discovered that SEO was worth 3x what last-click models showed, and display advertising was worth 40% less. This insight shifted hundreds of millions in budget. Claude can help you run the same analytical thinking at your scale.
Module 5 — Attribution
Multi-Touch Attribution with Claude
Real customers touch multiple channels before buying. Understanding this path is essential for allocating budget correctly. Claude can analyse your conversion path data and show you what last-click is hiding.
Multi-Touch Attribution Prompt
Analyse these customer conversion paths (last 30 days, 500 conversions). Format: Channel touchpoints in order → Converted Path | Conversions | Revenue Instagram Ad → Google Search → Purchase | 145 | ₹8,70,000 Google Search → Purchase | 98 | ₹5,88,000 YouTube Ad → Instagram → Search → Purchase | 72 | ₹5,76,000 Email → Purchase | 68 | ₹4,08,000 Instagram → Instagram → Search → Purchase | 52 | ₹4,16,000 Direct → Purchase | 38 | ₹2,28,000 Influencer → Search → Purchase | 27 | ₹2,43,000 Under last-click, Google Search gets 100% credit for paths 1, 3, 5, 7. Under linear attribution, recalculate channel credit. 1. Which channel is most undervalued by last-click? 2. How much of Search revenue actually originated from Instagram/YouTube? 3. What does this suggest about our Instagram budget allocation?
The Key Insight
In most multi-touch analyses, top-of-funnel channels (Instagram, YouTube, Display) initiate 40-60% of journeys that last-click attributes entirely to branded search. This means brands systematically underfund the channels that generate demand and overfund the channel that captures it.
Insight Card — Attribution
"Incrementality — The Only Attribution That Actually Tells the Truth"
All attribution models have one fundamental flaw: they assume every touchpoint contributed to a conversion that would not have happened otherwise. This is often wrong.
The real question is incrementality: If I had NOT run this campaign, how many fewer conversions would have occurred? This is the true value of a channel.
The ghost stores experiment: Meta ran an experiment for a major retailer — they showed ads to half the audience and nothing to the other half (the "ghost" group). The group that saw ads bought more — but the difference was only 18% more, not 100% more. The other 82% of attributed conversions would have happened anyway. Last-click attributed 100% of those sales to Meta.
How to run incrementality tests with Claude's help: "We want to test the incrementality of our Instagram budget. Design a geo holdout test — which regions should be our control (no Instagram ads) vs test (normal Instagram ads), and for how long do we need to run it to detect a 15% lift with 80% power?"
Claude can design the test, calculate required sample sizes, and analyse the results.
The real question is incrementality: If I had NOT run this campaign, how many fewer conversions would have occurred? This is the true value of a channel.
The ghost stores experiment: Meta ran an experiment for a major retailer — they showed ads to half the audience and nothing to the other half (the "ghost" group). The group that saw ads bought more — but the difference was only 18% more, not 100% more. The other 82% of attributed conversions would have happened anyway. Last-click attributed 100% of those sales to Meta.
How to run incrementality tests with Claude's help: "We want to test the incrementality of our Instagram budget. Design a geo holdout test — which regions should be our control (no Instagram ads) vs test (normal Instagram ads), and for how long do we need to run it to detect a 15% lift with 80% power?"
Claude can design the test, calculate required sample sizes, and analyse the results.
The only true measure of a channel's value is what happens when you turn it off. Use Claude to design incrementality tests before making major budget decisions.
Module 5 — Attribution
Google & Meta Ads Analysis with Claude
Google and Meta are the two largest paid advertising platforms for most businesses. Claude helps you analyse performance, identify inefficiencies, and optimise spend across both platforms simultaneously.
Cross-Platform Ads Analysis Prompt
Compare Google Ads vs Meta Ads performance (last 30 days). Business: D2C supplement brand | Target ROAS: 3.5x | Gross Margin: 58% GOOGLE ADS: Spend: ₹3,20,000 | Revenue: ₹12,80,000 | ROAS: 4.0x New Customers: 185 | Returning: 142 | CAC: ₹1,730 META ADS: Spend: ₹2,80,000 | Revenue: ₹5,88,000 | ROAS: 2.1x New Customers: 312 | Returning: 48 | CAC: ₹898 Meta-attributed ROAS is 7-day click, Google is last-click 1. Which platform is actually more efficient given attribution differences? 2. Meta has lower CAC but lower ROAS — is this a problem? 3. Google has more returning customers — what does this tell us? 4. How should we reallocate the combined ₹6L budget for next month? 5. What one test should we run to resolve the attribution confusion?
Module 5 — Attribution
ROAS Optimisation Framework
Optimising ROAS is not just about pausing underperforming campaigns. It requires a systematic framework that considers marginal returns, audience saturation, creative fatigue, and channel interaction effects.
ROAS Optimisation Prompt
Our blended ROAS has declined from 4.2x to 3.1x over 8 weeks. Weekly data: Week | Spend | Revenue | ROAS | New Cust | CPM | CTR | CVR W1 | ₹4.2L | ₹17.6L | 4.2x | 220 | ₹180 | 2.8% | 3.1% W2 | ₹4.5L | ₹18.0L | 4.0x | 228 | ₹185 | 2.7% | 3.0% W3 | ₹5.0L | ₹19.0L | 3.8x | 235 | ₹195 | 2.5% | 2.8% W4 | ₹5.5L | ₹19.8L | 3.6x | 240 | ₹210 | 2.3% | 2.7% W5 | ₹6.0L | ₹20.4L | 3.4x | 242 | ₹225 | 2.1% | 2.6% W6 | ₹6.5L | ₹20.8L | 3.2x | 238 | ₹245 | 1.9% | 2.5% W7 | ₹7.0L | ₹21.7L | 3.1x | 231 | ₹268 | 1.7% | 2.5% W8 | ₹7.5L | ₹23.3L | 3.1x | 228 | ₹285 | 1.6% | 2.4% 1. Diagnose the ROAS decline — is it CPM inflation, creative fatigue, CVR drop? 2. What is the "optimal spend" based on marginal ROAS (incremental return)? 3. At week 5, should we have reduced budget? What signal indicates this? 4. What 3 actions would stabilise ROAS back to 3.8x+?
Module 5 — Lab
Lab: Attribution Model Comparison
Lab Exercise
Attribution Analysis for Your Business
This lab helps you understand how different attribution models change your view of channel performance — and budget decisions.
1
Pull your GA4 conversion paths report (Advertising → Attribution → Conversion paths). Export the top 20 paths. Paste into Claude and ask: "Under last-click vs linear attribution, which channels gain and which lose credit?"
2
Ask Claude to calculate: "If we shifted to linear attribution, how would our budget allocation change? Recalculate channel ROAS under linear model."
3
Design an incrementality test: "We want to test the true value of our email retargeting. Design a 2-week holdout test — what % of our email list should be the control group, and what metrics should we track?"
Module 5 — Quiz
Quiz — Attribution & Paid Ads (6 Questions)
Q1. Which attribution model gives the most credit to channels that initiate the customer journey?
A. First-touch attribution — assigns 100% credit to the first channel a customer interacted with, making it best for understanding brand awareness and top-of-funnel channels
B. Last-touch attribution
C. Linear attribution
D. Time-decay attribution
Q2. Incrementality testing measures:
A. How much revenue each channel generates
B. How many additional conversions occurred because of a campaign — the true causal impact, isolated by comparing a group that saw the ad vs a matched control group that didn't
C. The total number of touchpoints before conversion
D. The statistical significance of ROAS differences
Q3. ROAS declines from 4.2x to 3.1x over 8 weeks while spend increases. CTR falls from 2.8% to 1.6%. The primary cause is:
A. Landing page conversion rate dropped
B. Creative fatigue combined with CPM inflation — rising CPM means ads cost more per impression, falling CTR means fewer of those impressions convert to clicks. Both signal audience saturation and the need for fresh creatives.
C. The product is no longer in demand
D. Competitor budget increased
Q4. Meta reports 2.1x ROAS using 7-day click attribution, Google reports 4.0x using last-click. To compare them fairly, you should:
A. Average the two figures
B. Align attribution windows — request Meta data with same window as Google (or vice versa), or use GA4 as the single source of truth applying the same model to both channels
C. Trust Meta's number — it covers more days
D. Trust Google's number — search intent is higher
Q5. A channel with 312 new customers and lower ROAS vs a channel with 142 returning customers and higher ROAS. Which is more valuable long-term?
A. Higher ROAS channel always — profitability is the only metric that matters
B. Depends on LTV — if new customer LTV × retention rate exceeds returning customer incremental value, the new customer channel wins. Ask Claude to calculate 12-month LTV for each group before deciding.
C. Lower ROAS channel — volume is always the priority
D. They are equivalent since total revenue is similar
Q6. Marginal ROAS at current spend level is 1.8x even though average ROAS is 4.0x. This means:
A. Increase budget — average ROAS is healthy
B. Consider reducing budget — the next ₹1 of spend only returns ₹1.80. The campaign is showing diminishing returns and may be approaching audience saturation. Redirecting budget to fresher channels is likely more efficient.
C. Switch to a different attribution model
D. Average and marginal ROAS are the same thing
Slide 1 of 7
Slide 1 of 5
Use Next to navigate
Module 6 — Growth Analytics
Social Media Analytics with Claude
Social media generates enormous amounts of data — but most brands track the wrong metrics and make the wrong decisions. Claude helps you focus on the metrics that actually predict business outcomes, not just likes and follows.
📊
Engagement Rate
Interactions / Reach. More meaningful than raw likes — it tells you what % of people who saw your content actually responded to it.
🎬
Video Metrics
View rate, completion rate, saves — these predict algorithmic distribution more than likes. Claude helps you interpret what they mean.
💰
Social Commerce
Revenue attributed to organic social. Often underreported — Claude helps you estimate true organic social contribution.
Real World — How SUGAR Cosmetics Uses Social Analytics
SUGAR Cosmetics, India's fastest-growing cosmetics brand, built their business largely on Instagram. They track content-level performance obsessively — which posts generate DMs, link clicks, and story swipes that lead to purchases. Claude can replicate this analytical process for any brand.
Module 6 — Growth Analytics
Social Media Performance Analysis
Analysing social media performance requires moving beyond vanity metrics. Claude helps you connect social performance to business outcomes — awareness, website traffic, and ultimately revenue.
Social Analytics Prompt
Analyse Instagram performance for our fashion brand (last 30 days). Followers: 125,000 | Monthly Active Audience: 85,000 Post Type | Posts | Avg Reach | Eng Rate | Link Clicks | Saves | Content Theme Reels-Product | 8 | 22,000 | 4.2% | 380 | 1,250 | New arrivals Reels-Lifestyle| 6 | 45,000 | 6.8% | 125 | 2,800 | Outfit ideas Carousels | 5 | 12,000 | 5.1% | 420 | 980 | Styling tips Static posts | 10 | 8,000 | 2.1% | 95 | 180 | Product shots Stories | 45 | 18,000 | - | 850 | - | Mix 1. Which content type drives most website traffic vs most engagement? 2. Lifestyle Reels have 3x reach vs Product Reels — does this mean we should post more lifestyle content? 3. Stories have highest link clicks — what does this tell us? 4. Build a content mix recommendation for next month
Module 6 — Growth Analytics
Growth Funnel Analysis
Every business has a growth funnel — the stages from first awareness to loyal customer. Claude helps you quantify each stage, identify the biggest drop-off, and prioritise where to invest for the biggest impact.
Growth Funnel Prompt
Analyse our growth funnel and identify the biggest opportunity. Business: SaaS productivity tool | Monthly data Stage | Users | Conv to Next | Drop-off Website Visits | 85,000 | 18.0% | 82.0% Free Trial Start| 15,300 | 32.0% | 68.0% Feature Activated| 4,896 | 58.0% | 42.0% Day 7 Retained | 2,840 | 45.0% | 55.0% Converted Paid | 1,278 | 72.0% | 28.0% Month 2 Retained| 920 | - | - 1. At which stage is the biggest absolute loss of users? 2. If we improve Trial Start → Feature Activated from 32% to 45%, how many more paid subscribers do we get monthly? 3. The Day 7 retention at 58% — is this good or bad for SaaS? 4. Which single stage improvement has the highest revenue leverage? 5. Design one experiment for each of the top 3 drop-off points
Module 6 — Growth Analytics
Course Summary & Your Next Steps
You now have a complete marketing analytics toolkit powered by Claude. Here is what you have built — and what to do with it.
✓
Foundation
You know how to brief Claude effectively, feed data correctly, and get outputs in the exact format you need for any marketing context.
✓
Campaign Performance
You can diagnose campaign issues, analyse A/B tests for significance, allocate budget by marginal ROAS, and write CMO-ready performance reports — all with Claude.
✓
Customer Segmentation
You can build RFM models, analyse cohort retention, model LTV by acquisition channel, and design differentiated strategies for each customer segment.
✓
SEO & Content Analytics
You can audit content performance, find keyword opportunities, identify competitor gaps, and prioritise content by revenue per session — not just traffic.
✓
Attribution & Paid Ads
You understand attribution models, can run multi-touch analysis, design incrementality tests, and optimise ROAS systematically.
✓
Growth Analytics
You can analyse social performance, build growth funnels, identify drop-off points, and quantify the revenue impact of improving each stage.
Your Next Step
Take the final exam. Pass with 70%+ to earn your Marketing Analytics Using Claude certificate, valid for 1 year. Then pick one real campaign, run one full analysis using Claude, and share the insight with your team. That first real insight is worth more than any certificate.
Module 6 — Quiz
Quiz — Social Media & Growth (6 Questions)
Q1. Engagement Rate is calculated as:
A. Likes / Followers
B. Total Interactions (likes + comments + shares + saves) / Reach — measuring what % of people who saw your content actually engaged with it
C. Comments / Posts
D. Shares / Impressions
Q2. Lifestyle Reels get 3x more reach than Product Reels but fewer link clicks. The correct conclusion is:
A. Post only lifestyle content — reach is all that matters
B. Use both strategically — Lifestyle Reels build awareness and drive saves/follows; Product Reels drive purchase intent and clicks. Optimise each for its role in the funnel.
C. Product Reels are failing — pause them
D. Reach is a vanity metric — ignore it entirely
Q3. In a SaaS funnel, "Feature Activated" to "Day 7 Retained" conversion of 58% means:
A. 58% of all website visitors are retained
B. Of users who activated a key feature, 58% are still using the product after 7 days — this is a critical milestone because users who complete activation are dramatically more likely to convert to paid
C. 58% of paid customers activated features
D. The product has a 42% satisfaction rate
Q4. The highest revenue leverage point in a funnel is typically:
A. The stage with the highest number of users
B. The stage with the highest absolute drop-off that still has high commercial intent downstream — improving this stage cascades through all subsequent conversion steps
C. The final conversion stage always
D. The first touchpoint — awareness is everything
Q5. Saves on Instagram posts are valuable because:
A. They count as purchases in Meta Ads Manager
B. Saves indicate high-intent engagement — the user wants to return to this content later (e.g., to shop, to try a recipe, to reference an idea). They also signal to the algorithm that content is valuable, improving distribution.
C. Instagram pays brands for every save received
D. Saves are counted in reach metrics
Q6. If improving Trial Start → Feature Activated from 32% to 45% generates 414 additional activations monthly, and 58% of activations convert to paid at ₹2,999/month, the monthly revenue impact is:
A. ₹12,39,586
B. ₹7,19,760 — 414 additional activations × 58% paid conversion = 240 additional paid customers × ₹2,999 = ₹7,19,760 additional monthly recurring revenue
C. ₹4,80,000
D. ₹1,24,000
Slide 1 of 5
Slide 1 of 14
Use Next → to navigate
Module 7 — Reporting & Advanced Analytics
Building CMO Dashboards with Claude
The CMO dashboard is the single most important marketing deliverable you create each week. It must tell a clear story, surface the right insights, and drive one or two clear decisions. Claude helps you build it faster and make it more insightful.
📊
Executive Summary
3-5 bullets: performance vs target, top win, biggest risk, key decision needed. Claude writes this from raw data in seconds.
📈
Performance vs Target
Traffic, leads, revenue vs plan. Claude calculates variance and identifies which channels are ahead/behind vs forecast.
🎯
Channel Performance
Blended ROAS, CAC by channel, best/worst performers. Claude ranks and flags anomalies automatically.
💡
Key Insight
The one thing the CMO should know that isn't obvious from the numbers. Claude surfaces the non-obvious pattern.
✅
Recommended Action
The one budget or strategy change that should happen this week. Specific, not vague.
🔮
Next Week Preview
What to watch: upcoming campaigns, seasonal shifts, competitor activity. Claude helps you anticipate, not just report.
🌍 Real World — How Zepto Uses Weekly Dashboards
Zepto's marketing team runs a 30-minute Monday morning review — all channel performance pasted into a structured Claude prompt that outputs a 5-bullet summary, top performer, biggest underperformer, and one budget reallocation recommendation. This meeting replaced a 2-hour weekly deck building session.
Module 7 — Reporting & Advanced Analytics
CMO Dashboard Prompt
This is the most valuable prompt in this entire course — the weekly CMO dashboard generator. Save this, customise it for your business, and use it every Monday morning.
The Weekly CMO Dashboard Prompt
Generate our weekly marketing dashboard. Business: [your brand] Revenue target this month: ₹[X]. Days elapsed: [Y]. Revenue to date: ₹[Z]. CHANNEL PERFORMANCE (WoW): [paste your channel data table] TOP 3 CAMPAIGNS: [paste best performing campaigns] BOTTOM 3 CAMPAIGNS: [paste worst performing campaigns] FUNNEL METRICS: Sessions: [X] | Leads/Signups: [X] | Paid Conversions: [X] | CAC: ₹[X] ANOMALIES THIS WEEK: [any unusual events: competitor activity, platform issues, creative launches] Generate: 1. EXECUTIVE SUMMARY (5 bullets, each max 15 words) 2. PERFORMANCE vs TARGET (on track / at risk / behind — with specific number) 3. KEY INSIGHT (the non-obvious pattern in this data) 4. BUDGET RECOMMENDATION (specific reallocation for next week) 5. ONE QUESTION THIS DATA CANNOT ANSWER (but the CMO will ask)
The Dashboard Philosophy
A good CMO dashboard answers 3 questions: Where are we vs where we should be? Why? What should we do about it? If your dashboard cannot answer all three, it is a data dump, not an insight document. Claude's job is to transform the data dump into the insight document.
Module 7 — Reporting & Advanced Analytics
Influencer Analytics with Claude
Influencer marketing is one of the fastest-growing channels — but also one of the hardest to measure. Claude helps you evaluate influencer performance, calculate true ROI, and identify which influencers actually drive revenue vs just awareness.
Influencer Performance Analysis Prompt
Analyse influencer campaign performance for our skincare brand (last quarter): Campaign fee paid + product cost included in "Investment" Influencer | Followers | Views | Eng% | Investment | Clicks | Conv | Revenue | ROAS @beauty_priya | 850K | 420,000 | 4.8% | ₹1,80,000 | 8,400 | 142 | ₹7,10,000| 3.9x @skincare_meera | 320K | 280,000 | 8.2% | ₹80,000 | 5,600 | 98 | ₹4,90,000| 6.1x @glam_rohan | 2.1M | 95,000 | 1.2% | ₹3,50,000 | 1,900 | 18 | ₹90,000 | 0.3x @naturals_kavya | 48K | 62,000 | 12.4% | ₹25,000 | 3,100 | 68 | ₹3,40,000| 13.6x @lifestyle_arjun| 680K | 180,000 | 3.8% | ₹1,20,000 | 3,600 | 52 | ₹2,60,000| 2.2x 1. Rank influencers by ROAS. Who should we book again? 2. @glam_rohan has 2.1M followers but 0.3x ROAS — what went wrong? 3. @naturals_kavya has only 48K followers but 13.6x ROAS — why? 4. What follower size range delivers best ROI for our brand? 5. Design a Q3 influencer strategy based on these learnings
🌍 Real World — The Micro-Influencer Insight
Mamaearth built a ₹1,000+ Cr brand largely on micro-influencers (10K-100K followers) rather than mega-celebrities. Their data showed micro-influencers delivered 6-8x better conversion rates at 80% lower cost per creator. Claude can run this same analysis on your influencer data — the insight changes budget allocation dramatically.
Module 7 — Reporting & Advanced Analytics
Referral Programme Analytics
Referral programmes are the most capital-efficient acquisition channel when they work — CAC is essentially the referral reward cost. Claude helps you analyse referral performance, optimise incentive structures, and identify your most valuable referrers.
Referral Programme Analysis Prompt
Analyse our referral programme performance (last 6 months): Reward structure: Referrer gets ₹200 credit, Referee gets 20% off first order Metric | Value Total referral links shared | 8,400 Referral links clicked | 3,360 (40% click rate) First orders from referrals | 672 (20% conversion of clicks) Referral CAC | ₹280 (reward cost / acquired customer) Referral customer AOV | ₹2,840 Referral customer M3 retention| 52% (vs 38% organic average) Top referrer cohort (>5 refs) | 124 customers | 41% of all referrals Referral revenue contribution | 8.4% of total revenue 1. How does referral CAC compare to our paid channels (Meta CAC ₹842)? 2. Referral customers have 52% M3 retention vs 38% organic — why? 3. 124 customers drive 41% of referrals — who are these "super-referrers"? 4. If we increase referrer reward from ₹200 to ₹400, model the economics 5. Design a "super-referrer" programme to activate the top 5% of referrers
Module 7 — Reporting & Advanced Analytics
App Analytics — DAU, MAU & Stickiness
For app-based businesses, DAU/MAU ratio (stickiness) is the single most important engagement metric. It tells you whether users have built a habit around your product. Claude helps you interpret app analytics and identify growth levers.
App Analytics Prompt
Analyse our fitness app analytics (last 3 months): Month | Downloads | DAU | MAU | Stickiness | Session/Day | Avg Duration | Uninstalls Jun | 24,200 | 8,400 | 38,000 | 22% | 1.4 | 8.2 min | 3,200 Jul | 31,400 | 10,200 | 46,000 | 22% | 1.5 | 8.8 min | 4,100 Aug | 28,800 | 11,800 | 52,000 | 23% | 1.6 | 9.4 min | 5,200 Feature Usage (% of DAU): Workout Tracker: 68% | Nutrition Log: 34% | Community: 18% | Live Classes: 12% 1. Our stickiness is 22-23% — is this good for a fitness app? (Benchmark: Duolingo 47%, Instagram 62%, average apps 13%) 2. Uninstalls rising 63% (Jun-Aug) while MAU grows — what causes this? 3. Nutrition Log at 34% DAU but Workout Tracker at 68% — what does this mean? 4. Live Classes at only 12% — high investment, low usage. Pivot or persist? 5. What 3 product changes would most improve stickiness to 30%+?
Stickiness Benchmarks
DAU/MAU = Stickiness. Good: 20%+ (social/utility). Great: 40%+ (habit-forming). Best-in-class: 60%+ (Instagram, WhatsApp). For fitness apps, 25-35% is strong. Below 15% suggests users have not formed a habit — the biggest product risk. Always include benchmark context when presenting stickiness to stakeholders.
Module 7 — Reporting & Advanced Analytics
Viral Coefficient & Growth Loops
The viral coefficient (K-factor) measures how many new users each existing user brings. If K > 1, your product grows without any paid acquisition. Claude helps you calculate your K-factor and identify which growth loops to invest in.
The Viral Coefficient Formula
K = Average Invites Sent per User × Conversion Rate of Invites
If K = 1.2: Each user brings 1.2 new users → exponential growth
If K = 0.8: Each user brings 0.8 new users → you need paid acquisition to compensate
If K = 0.3: Heavy acquisition dependence — viral loop is broken
If K = 1.2: Each user brings 1.2 new users → exponential growth
If K = 0.8: Each user brings 0.8 new users → you need paid acquisition to compensate
If K = 0.3: Heavy acquisition dependence — viral loop is broken
Viral Coefficient Prompt
Calculate our viral coefficient and model growth scenarios. Collaboration SaaS tool | Current users: 12,000 Sharing data (last 30 days): - Users who invited teammates: 28% (3,360 users) - Average invites sent per active inviter: 3.2 - Invite acceptance rate: 34% - New users from invites: 3,652 - New users from paid acquisition: 1,200 - Total new users: 4,852 1. Calculate our K-factor. Are we in viral growth territory? 2. If we improve invite acceptance from 34% to 50%, what happens to K? 3. What is the "viral cycle time" — how long does one cycle take? 4. Model growth at current K vs K=1.0 vs K=1.2 over 6 months 5. What 3 product changes would most increase % of users who invite?
Module 7 — Reporting & Advanced Analytics
Automated Reporting Workflows
The best marketing analytics teams don't spend time on reporting — they spend time on insight. By building standardised Claude prompt templates for recurring reports, you can produce weekly, monthly, and quarterly reports in minutes instead of days.
1
The Weekly Report Template
Create a Google Sheet where your team pastes the week's numbers into fixed cells. Copy the populated table into your Claude "Weekly Dashboard" prompt. Output: 5-bullet CMO summary, channel ranking, one recommendation. Time: 8 minutes.
2
The Monthly Performance Report
End of month: paste 4 weeks of data into Claude. Prompt: "Compare this month vs last month vs same month last year. Identify trend changes. Write a 300-word performance narrative suitable for board reporting." Time: 15 minutes.
3
The Campaign Post-Mortem
After every major campaign (sale, product launch, seasonal): "Analyse this campaign's performance vs forecast. What worked, what didn't, what would we do differently? Format as a lessons-learned document for the team." Time: 10 minutes.
4
The Quarterly Business Review
Every quarter: "Given this 13-week performance data, identify the 3 most significant shifts in our marketing performance and their likely causes. Then recommend 3 strategic changes for next quarter." Time: 20 minutes.
🌍 Real World — How Swiggy Automates Analytics
Swiggy's performance team has a "daily digest" prompt that runs every morning — yesterday's data from all channels pasted in, Claude outputs a 3-minute morning briefing. Total time from data export to team communication: 12 minutes. The same system can work for any team size.
Module 7 — Reporting & Advanced Analytics
Presenting Data to Stakeholders
The best analysis in the world is worthless if you can't communicate it effectively. Claude helps you tailor your data presentation to different audiences — and anticipate the questions they'll ask before they ask them.
Stakeholder Communication Prompt
I need to present our Q2 marketing performance to 3 different audiences. Key data: Blended ROAS 3.2x (target 3.5x), CAC ₹1,840 (+12% QoQ), Revenue ₹4.8Cr (vs ₹5.2Cr target), New customers +18% QoQ Write 3 versions of the Q2 performance summary: VERSION 1 — For the CEO (60 seconds max, business outcomes only): Focus on: revenue vs target, growth trajectory, one key risk, one key opportunity VERSION 2 — For the CFO (financial focus, numbers-led): Focus on: efficiency metrics, CAC trends, ROAS vs target, unit economics health VERSION 3 — For the Marketing Team (tactical, motivating): Focus on: what worked, what we learned, what we're changing, team wins Also: What is the one question each audience will ask that this data doesn't clearly answer? Prepare a response for each.
Module 7 — Reporting & Advanced Analytics
Building Your Marketing Analytics Stack
The tools you use determine the questions you can answer. Claude helps you design the right analytics stack for your business size and maturity — and identify the gaps that are costing you insight.
| Business Stage | Essential Tools | Claude Use Case |
|---|---|---|
| Early stage (0-₹1Cr/mo) | GA4 + Meta Ads Manager + Google Ads + Google Search Console | Weekly performance analysis, campaign diagnosis, content audit |
| Growth stage (₹1-10Cr/mo) | + Klaviyo/Mailchimp + Hotjar + SEMrush + a CRM | Email performance, user behaviour analysis, SEO opportunity identification |
| Scale stage (₹10Cr+/mo) | + Segment/mParticle + Amplitude/Mixpanel + Looker + Data warehouse | Cohort analysis, funnel optimisation, multi-touch attribution, executive dashboards |
Analytics Stack Audit Prompt
Audit our current analytics stack and identify gaps: We currently use: GA4, Meta Ads Manager, Google Ads, Mailchimp, basic Excel Monthly revenue: ₹3.5Cr | Team size: 8 people | 3 in marketing 1. What analytics questions can we NOT currently answer with this stack? 2. What is the highest-impact tool addition for our stage? 3. What data are we collecting but not using? 4. Build a 12-month analytics maturity roadmap for our business
Module 7 — Reporting & Advanced Analytics
Marketing Forecasting with Claude
Marketing forecasting — predicting next month's revenue from planned spend — is one of the most valuable skills for any senior marketer. Claude helps you build data-driven forecasts and communicate confidence intervals to stakeholders.
Revenue Forecasting Prompt
Build a marketing revenue forecast for Q4 (Oct-Dec). Historical data and planned changes: HISTORICAL (same period last year): Oct: ₹82L spend → ₹3.8Cr revenue (4.6x ROAS) Nov: ₹95L spend → ₹4.4Cr revenue (4.6x ROAS) Dec: ₹68L spend → ₹3.1Cr revenue (4.6x ROAS) PLANNED CHANGES THIS YEAR: - Budget increased 25% vs last year - New product line launching Oct 15 (estimated +15% AOV) - New influencer partnership starting Nov 1 - Competitor known to increase spend in Nov (expect +10% CPMs) - Website redesign improving CVR by est. 12% (Sep deployment) Build forecast for Oct, Nov, Dec showing: 1. Base case (last year × budget increase × CVR improvement) 2. Bull case (new product line performs, influencer works) 3. Bear case (competitor CPM inflation erodes efficiency) 4. Most likely Q4 revenue and confidence range
Module 7 — Reporting & Advanced Analytics
💡 The Analytics Mindset
💡 Insight Card — Analytics Mindset
"The Best Marketers Are Scientists, Not Artists — And Now They Have Claude"
There is a persistent myth in marketing that creativity is the primary driver of success and data is a constraint. The most effective marketers know this is wrong.
The best campaigns are hypotheses. The best marketing decisions are experiments. The best teams are those who learn faster from data than their competitors.
What separates good from great marketing teams:
• Good teams know their ROAS. Great teams know their marginal ROAS by audience and creative.
• Good teams track conversions. Great teams track conversion by cohort, acquisition channel, and device.
• Good teams report on what happened. Great teams use Claude to model what will happen.
The Claude advantage: Before AI, building a rigorous weekly marketing analysis required a data analyst, 6-8 hours of work, and usually happened too late to act on the insight. Claude compresses this to 15 minutes, available to every marketer regardless of technical skill.
The marketers who will lead teams in 5 years are those who combine creative instinct with analytical rigour — and use Claude to move at the speed of both simultaneously.
You now have the analytical toolkit. The creative instinct is yours. The combination is unstoppable.
The best campaigns are hypotheses. The best marketing decisions are experiments. The best teams are those who learn faster from data than their competitors.
What separates good from great marketing teams:
• Good teams know their ROAS. Great teams know their marginal ROAS by audience and creative.
• Good teams track conversions. Great teams track conversion by cohort, acquisition channel, and device.
• Good teams report on what happened. Great teams use Claude to model what will happen.
The Claude advantage: Before AI, building a rigorous weekly marketing analysis required a data analyst, 6-8 hours of work, and usually happened too late to act on the insight. Claude compresses this to 15 minutes, available to every marketer regardless of technical skill.
The marketers who will lead teams in 5 years are those who combine creative instinct with analytical rigour — and use Claude to move at the speed of both simultaneously.
You now have the analytical toolkit. The creative instinct is yours. The combination is unstoppable.
Marketing success is not about spending more — it is about learning faster. Claude makes your learning loop 10x faster. Use it every week, on every campaign, with every dataset.
Module 7 — Lab
Lab: Build Your First CMO Dashboard
🔬 Lab Exercise
Complete Marketing Dashboard Build
Using your real data (or the sample provided), build a complete CMO-ready marketing dashboard using Claude. This is the capstone exercise of the course.
1
Export last 4 weeks of data from your top 3 channels. Organise into a table: Channel | Week 1-4 Spend | Revenue | ROAS | New Customers | Notes. Paste into Claude with the Weekly CMO Dashboard prompt from Slide 2.
2
Ask Claude to write 3 versions of the performance summary: one for CEO (60 seconds), one for CFO (financial focus), one for your team (tactical and motivating). Compare how the same data is framed differently for each audience.
3
Ask Claude: "Based on this performance data, what are the 3 most important questions we should be trying to answer in our analytics next month? What data would we need to answer each?"
4
Build a 90-day marketing analytics roadmap: "Given our current data maturity and team size, what analytics capabilities should we build over the next 90 days? Prioritise by business impact."
Module 7 — Key Distinctions
Key Distinctions — Advanced Analytics
Reporting vs Analysis vs Insight
Reporting = what happened (ROAS was 3.2x). Analysis = why it happened (ROAS fell because CPM rose 18% in week 3). Insight = what to do about it (reduce Meta prospecting budget by 20% and reallocate to Google Shopping which is undersaturated). Claude can do all three — but only if you ask for all three.
K-factor vs Referral Rate
Referral rate = % of customers who make at least one referral. K-factor = net new users generated per existing user through virality. K-factor accounts for both the referral rate AND the conversion rate of those referrals. K > 1 = viral growth. K < 1 = acquisition dependence.
DAU vs MAU vs Stickiness
DAU = users active today. MAU = users active this month. Stickiness = DAU/MAU = habit formation rate. A product with 100K MAU but 10K DAU (10% stickiness) has weak habit formation. A product with 100K MAU and 40K DAU (40% stickiness) has built strong daily habits — and is much more defensible.
Micro vs Macro Influencers
Macro (1M+) = reach and awareness, higher CPM, lower engagement rates, harder to fake. Micro (10K-100K) = niche authority, higher engagement, better conversion rates, easier to scale with budget. Neither is universally better — depends on your objective and the brand's stage.
Module 7 — Quiz
Quiz — Reporting & Advanced Analytics (8 Questions)
Q1. A CMO dashboard should answer which 3 questions?
A. How much did we spend, what did we earn, what was the ROAS?
B. Where are we vs target? Why? What should we do about it? — These three questions distinguish an insight document from a data dump.
C. Who is our best customer, worst campaign, and top channel?
D. What happened last week, last month, and last year?
Q2. A micro-influencer (45K followers) outperforms a macro-influencer (2M followers) in ROAS because:
A. Larger influencers are always overpriced
B. Micro-influencers typically have higher engagement rates, more trust with their niche audience, and their followers have stronger purchase intent alignment — leading to better conversion rates despite lower reach
C. Smaller audiences always convert better
D. The macro-influencer's content was poor quality
Q3. K-factor of 0.8 means:
A. 80% of users refer at least one friend
B. Each user generates 0.8 new users through virality — below 1, so the product cannot sustain growth without paid acquisition. The viral loop decays rather than compounds.
C. The product has 80% month-2 retention
D. 80% of revenue comes from referrals
Q4. App stickiness (DAU/MAU) of 22% compared to Instagram at 62% means:
A. The app is failing and should be shut down
B. Users access the app roughly 6-7 days per month vs Instagram's 19 days — for a fitness app this is below best-in-class (35%+) but not catastrophic. The goal is building stronger daily habits through product improvements.
C. The app is performing at industry standard
D. 22% of users are paying subscribers
Q5. UTM parameters "facebook" and "FACEBOOK" appearing as separate sources in GA4 causes:
A. GA4 automatically merges them
B. Duplicate channel reporting — performance appears split across two "channels", understating Facebook's true contribution and making benchmarking inaccurate. Always use lowercase and consistent naming.
C. Security alerts in GA4
D. The campaigns to pause automatically
Q6. Revenue "forecasting" with Claude requires which inputs?
A. Just last month's revenue
B. Historical performance by period, planned budget changes, known external factors (competitor activity, seasonality, product launches, platform changes), and clear assumptions for each scenario
C. Only the marketing budget for the forecast period
D. ROAS from the most recent campaign
Q7. The difference between "reporting" and "insight" is:
A. Reporting uses charts, insight uses numbers
B. Reporting states what happened (ROAS 3.2x). Insight explains why and what to do (ROAS fell because CPM rose 18% due to creative fatigue — recommend rotating creative and reducing Meta prospecting budget by 20%)
C. Reporting is for the CMO, insight is for the team
D. Insight requires a data science team; reporting can be done manually
Q8. Core Web Vitals (LCP, INP, CLS) matter for marketing analytics because:
A. Google AdWords requires Core Web Vitals scores
B. Poor Core Web Vitals directly reduce organic search rankings (SEO revenue) and increase bounce rates on paid landing pages (reducing conversion rates and ROAS). Every marketing channel's performance is affected by site speed.
C. They only matter for technical SEO teams, not marketing
D. Core Web Vitals only affect mobile traffic
Slide 1 of 14
📊
Final Certification Exam
Test your knowledge across all 6 modules. Pass 70%+ to receive your Marketing Analytics Using Claude certificate, valid for 1 year.
Questions
50 Questions
Pass Mark
70% (35/50)
Attempts
Maximum 2
Certificate
1-Year Validity
Questions
Answered
Flagged
Unanswered
60:00
Question 1 of 50