What is Claude
Claude is an AI built by Anthropic. You give it data, ask questions in plain English, get answers. No formulas, no pivot tables, no report builder. It reads your entire file and remembers context, so you can ask follow-ups and dig deeper.
What You Can Do With Ad Data
- •Find wasted spend in seconds
- •Spot patterns across campaigns, ad sets, audiences
- •Get negative keyword lists or exclusion lists ready to import
- •Compare any time period to any other
- •Ask "why did CPA spike" and get an actual answer
- •Generate bulk changes for Google Ads Editor or Meta bulk upload
No more digging through the UI. No more building the same reports every week.
What This Means for Your Workflow
You already know how to analyze your accounts. Claude just makes it faster. Same questions you'd ask yourself, answered without the manual work.
Two Ways to Connect Claude to Your Google/Meta Ads Data
Option 1
CSV Export
Simplest approach
- Go to Google/Meta Ads
- Export any report as CSV
- Upload to claude.ai
- Ask questions
Works immediately, no setup required.
Read-only analysis
Option 2
Direct Connection via Ryze AI
Full automation
- Go to get-ryze.ai
- Connect your Google/Meta Ads account
- Claude analyzes and applies changes
From $40/month. Free trial available.
Execute changes in your account
Prompts Library
Copy-paste these prompts after uploading your data. Each category has versions for both Google Ads and Meta Ads.
Campaign Performance
Google AdsAnalyze this campaign data and find all campaigns with ROAS below 2
Include for each campaign:
- Campaign name
- Campaign type (Search, Performance Max, Display, Shopping)
- Total spend
- Conversions
- Conversion value
- ROAS
- Impression share
- Click share
Sort by spend highest to lowest
For Search campaigns:
- Note the match type mix (broad, phrase, exact)
- Flag if broad match is eating most of the budget
For Performance Max:
- Flag if asset groups have low asset performance ratings
- Note if search themes are missing
For each underperformer, tell me:
- Should I pause it completely
- Should I reduce budget and optimize
- What specific thing to fix first
Calculate total spend going to campaigns with ROAS below 2
Meta AdsAnalyze this campaign data and find all campaigns with ROAS below 2
Include for each campaign:
- Campaign name
- Objective (sales, leads, traffic, awareness)
- Total spend
- Purchases or leads
- Purchase value
- ROAS
- CPM
- Frequency
Sort by spend highest to lowest
Break down by:
- Placement performance (Feed vs Stories vs Reels vs Audience Network)
- Device (mobile vs desktop)
For Advantage+ campaigns:
- Compare performance to manual campaigns with similar audiences
- Flag if Advantage+ is spending on low-quality placements
For each underperformer, tell me:
- Kill it entirely
- Restructure with different audience
- Test new creative
- Adjust budget allocation
Calculate total spend going to campaigns with ROAS below 2
Wasted Spend
Google AdsFind all wasted spend in this search terms report
Criteria for wasted:
- Cost above $30 and zero conversions
- Cost above $100 and ROAS below 0.5
- Irrelevant terms that don't match our product/service
For each wasted term, show:
- Search term
- Matched keyword
- Match type
- Impressions
- Clicks
- Cost
- Conversions
Group wasted terms by theme:
- Competitor terms
- Informational queries (how to, what is, etc)
- Wrong intent (jobs, careers, reviews, free)
- Geographic mismatches
- Misspellings and typos
For each theme, give me:
- Negative keywords in phrase match format
- Negative keywords in exact match format
- Recommended negative list name
Calculate:
- Total wasted spend in this date range
- Projected monthly savings if I add these negatives
- Percentage of total spend that's wasted
Meta AdsFind all wasted spend in this ad set data
Criteria for wasted:
- Spend above $50 and zero purchases
- Spend above $200 and ROAS below 0.5
- Frequency above 4 with declining CTR
- CPM above $30 with no conversions
For each wasted ad set, show:
- Ad set name
- Campaign name
- Audience type (interest, lookalike, broad, retargeting)
- Spend
- Reach
- Frequency
- Purchases
- CPA
- ROAS
Identify the reason for waste:
- Audience too narrow (reach issues)
- Audience too broad (relevance issues)
- Creative fatigue (high frequency, declining CTR)
- Wrong objective for the audience
- Placement bleed (Audience Network eating budget)
For each wasted ad set, recommend:
- Kill immediately
- Pause and test new creative
- Change audience
- Adjust placement settings
Calculate:
- Total wasted spend
- Percentage of total budget wasted
- Projected monthly savings
Creative Analysis
Google AdsAnalyze responsive search ad performance across all ad groups
For each RSA, show:
- Ad group name
- Campaign name
- Impressions
- Clicks
- CTR
- Conversions
- Conversion rate
- Ad strength rating
Analyze headline performance:
- List top 10 headlines by CTR
- List top 10 headlines by conversion rate
- Flag headlines where high CTR does not mean high conversions
- Identify headlines that Google marked as "low" performing
Find patterns in winning headlines:
- Average character length
- Use of numbers or stats
- Use of CTAs (get, buy, try, start)
- Questions vs statements
- Brand mention vs no brand
- Price or offer mention
Find patterns in losing headlines:
- What they have in common
- Why they might be underperforming
Identify ad groups with only 1-2 active RSAs
- List by spend
- These need more ad variants
Give me:
- 10 new headline ideas based on winning patterns
- 5 description ideas based on what's working
- Priority list of ad groups to fix first
Meta AdsAnalyze ad creative performance across all campaigns
For each ad, show:
- Ad name
- Ad set name
- Campaign name
- Format (static image, video, carousel, collection)
- Spend
- Impressions
- Reach
- CTR
- Purchases
- CPA
- ROAS
Break down performance by format:
- Static image average ROAS
- Video average ROAS
- Carousel average ROAS
- Which format wins for this account
For video ads specifically:
- 3-second view rate
- Average watch time
- Does higher watch time correlate with more conversions
- Identify videos where people drop off early
Analyze creative patterns in top performers:
- Hook style (question, statement, shock, benefit)
- Visual style (lifestyle, product, UGC, graphic)
- Text overlay or no text
- CTA button type
- Length for videos
Flag creative fatigue:
- Any ad running longer than 14 days
- CTR declining week over week
- Frequency above 3
Give me:
- Top 5 creatives to scale (increase budget)
- Top 5 creatives to refresh (same concept, new execution)
- Creatives to kill immediately
- 5 new creative concepts based on what's working
Audience Analysis
Google AdsBreak down performance by audience segment
Include all audience types:
- In-market audiences
- Affinity audiences
- Remarketing lists
- Customer match
- Similar audiences
- Combined audiences
For each audience, show:
- Audience name
- Audience type
- Impressions
- Clicks
- CTR
- Conversions
- Conversion rate
- CPA
- ROAS
- Bid adjustment currently applied
Analysis:
- Flag any audience with CPA 50% higher than account average
- Flag any audience with ROAS 50% lower than account average
- Identify audiences worth increasing bid adjustment
- Identify audiences to exclude entirely
For remarketing lists:
- Check list sizes
- Flag any list under 1000 users
- Note if membership duration makes sense
For in-market and affinity:
- Which ones actually convert
- Which ones just add reach without conversions
Give me:
- Audiences to increase bids on (with recommended %)
- Audiences to decrease bids on (with recommended %)
- Audiences to exclude
- New audience ideas based on what's working
Meta AdsBreak down performance by audience type
Include all targeting types:
- Interest targeting (list each interest)
- Lookalike audiences (note percentage: 1%, 2%, 5%, 10%)
- Broad targeting (no detailed targeting)
- Custom audiences (website, customer list, engagement)
- Retargeting audiences
For each audience, show:
- Audience name/description
- Audience size
- Spend
- Reach
- Frequency
- CPM
- Purchases
- CPA
- ROAS
Analysis:
- Compare lookalike performance by percentage (do 1% lookalikes beat 5%?)
- Compare interest targeting vs broad (which wins?)
- Check if retargeting audiences are profitable at current frequency
- Identify audience overlap between ad sets
For overlapping audiences:
- List ad sets with more than 30% overlap
- Recommend which to consolidate
- Calculate budget waste from overlap
Give me:
- Audiences to scale
- Audiences to kill
- Audiences to consolidate
- New audience ideas based on what converts
- Recommended lookalike sources based on performance
Budget and Scaling
Google AdsAnalyze budget efficiency across all campaigns
For each campaign, show:
- Campaign name
- Campaign type
- Daily budget
- Average daily spend
- Impression share
- Search impression share lost to budget
- Search impression share lost to rank
- Conversions
- CPA
- ROAS
Identify limited campaigns:
- Campaigns losing more than 20% impression share to budget
- Calculate estimated clicks lost per day
- Calculate estimated conversions lost per day using current CVR
Scaling analysis for top performers (ROAS above 2):
- Current spend
- Current conversions
- Current CPA
- Projected spend at +30% budget
- Projected conversions (factor 15% efficiency loss)
- Projected CPA after increase
For underperformers:
- Recommend budget to reallocate from poor campaigns to good ones
- Calculate impact of reallocation
Give me:
- Priority list of campaigns to increase budget
- Specific new daily budget for each
- Campaigns to decrease budget
- Total budget reallocation plan
Meta AdsAnalyze scaling opportunities across all ad sets
For each ad set, show:
- Ad set name
- Campaign name
- Daily budget
- Lifetime spend
- Learning phase status
- Purchases
- CPA
- ROAS
- Frequency
- Audience size
Identify scaling candidates:
- Ad sets that exited learning with CPA below target
- Ad sets with ROAS above 2 and frequency below 2
- Ad sets with room to grow (audience size vs reach)
For each scaling candidate:
- Current daily budget
- Recommended new daily budget
- Expected impact on CPA (factor in efficiency loss)
- Risk assessment (audience saturation, creative fatigue)
Identify ad sets stuck in learning:
- List all ad sets still in learning phase
- Time in learning
- Reason likely stuck (budget too low, audience too small, not enough conversions)
- Recommendation to fix or kill
Identify ad sets to reduce:
- High frequency with declining performance
- CPA creeping up week over week
- Audience saturation signs
Give me:
- Scaling plan with specific budget numbers
- Ad sets to pause
- Budget to reallocate and where
Time and Placement
Google AdsAnalyze performance by time and device
Hour of day breakdown:
- Conversions by hour
- CPA by hour
- Conversion rate by hour
- Spend by hour
Day of week breakdown:
- Conversions by day
- CPA by day
- Conversion rate by day
- Spend by day
Identify:
- Top 5 performing hours (lowest CPA, highest CVR)
- Bottom 5 performing hours (highest CPA, lowest CVR)
- Best performing days
- Worst performing days
For bottom performers:
- Calculate wasted spend during those hours/days
- Recommend bid adjustments (specific percentages)
- Or recommend excluding entirely
Device breakdown:
- Desktop vs mobile vs tablet
- Conversions, CPA, ROAS for each
- Flag any device 50% worse than average
- Recommend device bid adjustments
Give me:
- Ad schedule recommendations (specific hours and bid adjustments)
- Device bid adjustment recommendations
- Projected savings from schedule optimization
Meta AdsAnalyze performance by placement
Break down by all placements:
- Facebook Feed
- Facebook Marketplace
- Facebook Video Feeds
- Facebook Right Column
- Instagram Feed
- Instagram Stories
- Instagram Reels
- Instagram Explore
- Audience Network
- Messenger
For each placement, show:
- Spend
- Impressions
- Reach
- CPM
- Clicks
- CTR
- Purchases
- CPA
- ROAS
Analysis:
- Rank placements by ROAS
- Rank placements by CPA
- Flag placements with CPA 2x higher than average
- Flag Audience Network if it's underperforming
- Compare Feed vs Stories vs Reels performance
For Advantage+ placement campaigns:
- Is the algorithm allocating budget well
- Or is it wasting money on bad placements
Give me:
- Placements to exclude
- Placements to focus budget on
- Projected savings from excluding bad placements
- Recommendation: Advantage+ vs manual placements for this account
What Claude Can and Cannot Do
Can do:
- Analyze any data you upload
- Find patterns across thousands of rows
- Calculate metrics and build reports
- Write scripts for Google Ads Editor
- Generate negative keyword lists
- Draft ad copy variations
Cannot do (without Ryze):
- Pull data from your account directly
- Make changes to campaigns
- Pause or enable anything
- Adjust bids or budgets
With Ryze connected, Claude can execute all of the above.
Tips for Better Results
Be specific about thresholds
Bad: "Find bad keywords"
Good: "Find keywords with CTR below 1.5% and spend above $100 in the last 30 days"
Include time frames
Bad: "How are my campaigns doing"
Good: "Compare last 7 days to previous 7 days, flag changes above 15%"
Ask for the format you need
- "Output as a table I can paste into Sheets"
- "Format as a Google Ads Editor bulk upload"
- "List as comma-separated values for negative keyword import"
Chain your analysis
Start broad: "What are the biggest issues in this account"
Then drill down: "Now analyze the top 3 problem areas in detail"
Connect Claude to Your Account
For CSV analysis: Just upload to claude.ai
For direct execution: Use get-ryze.ai (free trial)
With Ryze, Claude can go into your account, pull data, and apply changes—pause campaigns, adjust bids, add negatives—all from a single conversation.
