The 7 Questions Every Advertiser Should Ask Their Ads Every Week
Based on real usage data — the most common and most useful questions that surface actionable insights when you connect your Facebook Ads to AI. Not theory — these actually work.
The 7 Questions Every Advertiser Should Ask Their Ads Every Week
There are thousands of "prompt libraries" for Facebook Ads. Most of them are useless — generic questions that produce generic answers.
These 7 questions are different. They're based on what real advertisers actually ask when their AI can see their live ad data. Not theory. Not hypothetical. These are the questions that consistently surface actionable insights.
For each question, I'll show you the exact prompt, what a good answer looks like, and what to do based on what you learn.
Question 1: "What's wasting my money right now?"
The prompt
"Look at all my active ads. Which ones have spent more than $20 in the last 7 days with zero conversions or a CPA more than double my average? List them with their spend, impressions, and CTR."
What a good answer looks like
The AI should return a specific list with numbers — not vague advice. Something like:
"Ad 'Summer Collection v3' in campaign 'Retargeting - Website Visitors' has spent $87 in the last 7 days with 0 conversions. CTR is 0.3%, well below your account average of 1.2%. Recommend pausing immediately."
What a bad answer looks like
"You should check your low-performing ads and consider pausing them."
If you get a vague answer, your AI doesn't have access to your actual data. That's the difference MCP makes.
What to do
Pause the identified ads immediately. Don't wait. Dead ads don't come back to life — they just drain budget that could go to your winners.
Question 2: "Which ad should I scale and why?"
The prompt
"Which of my active ads has the best combination of low CPA, high volume, and stable performance over the last 7 days? I want to increase budget — tell me which one and by how much."
What a good answer looks like
"Your 'Testimonial Video - Sarah' ad has a CPA of $12.40 (your average is $18.90), with 23 conversions in the last 7 days. Performance has been stable — CPA was $11.80 last week. It's currently getting 15% of your total budget. I'd recommend increasing its ad set budget by 20-30% to test scaling."
Key insight
Good scaling recommendations consider more than just CPA. A low-CPA ad that only got 2 conversions isn't a scaling candidate — that's noise, not signal. You need low CPA + sufficient volume + stability over time.
What to do
Increase the budget on the recommended ad set by 20-30%. Don't double it overnight — Facebook's algorithm doesn't react well to dramatic budget jumps. Check back in 48 hours to see if performance holds.
Question 3: "Are any of my ads showing fatigue?"
The prompt
"Check my top-performing ads for creative fatigue. Look at frequency, CTR trend over the last 14 days, and CPA trend. Flag any ad where frequency is above 2.5 and CTR is declining."
What a good answer looks like
"Your 'Before/After Results' ad has a frequency of 3.8 and CTR has dropped from 1.8% to 0.9% over the last 14 days. CPA has risen from $14 to $22 in the same period. This ad is fatigued — your audience has seen it too many times."
Why this matters
Most advertisers run their ads until performance completely crashes. By catching fatigue early — when CTR starts declining but before CPA doubles — you save money and maintain momentum.
What to do
Prepare replacement creative. Don't just pause the fatigued ad — have the next version ready to go. Ask the AI:
"Based on the messaging in my best-performing ads, give me 5 new angle variations I can test as replacements."
Question 4: "How does this week compare to last week?"
The prompt
"Compare my overall ad account performance this week vs. last week. Show me total spend, conversions, CPA, ROAS, and CTR. Highlight anything that changed significantly."
What a good answer looks like
A clean comparison table with the percentage change for each metric, plus commentary on what's driving the changes.
"Total spend is up 12% ($1,240 vs $1,107), conversions are up 18% (67 vs 57), so CPA actually improved from $19.42 to $18.51 (-5%). The improvement is driven by your new 'Social Proof' campaign which launched Tuesday and is outperforming your other campaigns."
Why this matters
Week-over-week comparison is the single most useful regular check. It catches trends early — both positive and negative — before they become obvious in monthly reports.
What to do
If things are improving, identify what's driving it and double down. If things are degrading, dig into which specific campaigns or ads are responsible and address them.
Question 5: "What's my real cost per customer?"
The prompt
"What's my blended CPA across all campaigns? Break it down by campaign so I can see which campaigns are efficient and which are expensive. Include conversion volume so I can see if cheap campaigns are also meaningful in volume."
What a good answer looks like
A breakdown showing each campaign's CPA alongside its conversion volume. The critical insight is the relationship between efficiency and scale.
"Campaign 'Prospecting - Lookalike' has a CPA of $22 with 34 conversions. Campaign 'Retargeting' has a CPA of $8 with 12 conversions. Campaign 'Brand Awareness' has a CPA of $45 with 3 conversions. Blended CPA across all campaigns: $18.90."
Why this matters
Many advertisers only look at individual campaign CPA and miss the bigger picture. Your retargeting might look incredible, but it only converts people your prospecting campaigns already found. The blended view tells you what a customer actually costs.
What to do
Compare your blended CPA to your customer lifetime value. If blended CPA is well below LTV, you have room to scale. If it's approaching LTV, tighten up — focus on improving the campaigns with the highest CPA.
Question 6: "Which audience is performing best?"
The prompt
"Break down my performance by ad set. Which audiences are converting best? Show me CPA, conversion rate, and spend for each ad set so I can see which audiences I should invest more in."
What a good answer looks like
"Your 'Lookalike - Purchasers 1%' ad set has the best CPA at $14.20 with 28 conversions. 'Interest - Digital Marketing' is at $19.80 with 15 conversions. 'Broad Targeting' is at $24.50 with 22 conversions but the highest volume. 'Lookalike - Website Visitors 3%' is at $32.00 with only 4 conversions — consider pausing."
Why this matters
Audience performance shifts over time as audiences saturate and ad fatigue sets in. Checking weekly ensures you're allocating budget to the audiences that are currently working, not the ones that worked last month.
What to do
Shift budget toward your best-performing audiences. If a high-performing audience is getting a small share of budget, increase it. If an audience is consistently underperforming, reduce or pause it.
Question 7: "Write me 5 new headlines based on what's working"
The prompt
"Look at my top 3 performing ads by CPA. Analyze what makes their messaging effective — the angle, the hook, the value proposition. Then write 5 new headline variations that use similar patterns but with fresh angles."
What a good answer looks like
The AI should reference specific ads and explain WHY they work before generating new copy:
"Your top performers share a pattern: they lead with a specific result or number ('Cut my ad spend by 40%'), use second-person address ('Your ads are...'), and create urgency without being clickbaity. Here are 5 new variations..."
Why this matters
Most people use AI to generate ad copy in a vacuum — "write me headlines for my product." That produces generic copy. When the AI can see what's actually converting in your account, it generates copy rooted in proven patterns.
What to do
Take the best 2-3 headlines and create new ad variations. Keep everything else the same (same image, same audience) so you're isolating the headline as the variable. Run for at least 72 hours before judging.
The Meta-Lesson
Notice the pattern in these questions: they're all specific and data-driven. Not "how can I improve my ads?" (too vague), but "which ads have spent over $20 with zero conversions in the last 7 days?" (specific, actionable).
The quality of your AI's analysis is directly proportional to the specificity of your question. Vague input = vague output. Specific input = actionable output.
Get Answers in 30 Seconds
All 7 of these questions work right out of the box when your ads are connected to AI via MCP.
Connect your Facebook Ads to AdChat → and try Question 1 right now. The amount of wasted spend you'll find in the first 30 seconds will pay for the subscription.
Try this with your own ads
AdChat connects your Facebook Ads to Claude or ChatGPT. See your live ad data, get AI-powered analysis, and make better decisions in seconds.
Get AdChat