We Fired Our Best Performing Ad Because It Was Attracting the Wrong Customers
The ad was performing beautifully.
4.2x ROAS. Lowest CPA in the account. The platform loved it. Every optimization algorithm screamed “scale this.”
So we killed it.
Revenue immediately dropped. CPA went up 30%. The marketing team panicked.
Six months later, LTV was up 85% and the business was profitable for the first time in two years.
Here’s what happened.
The Ad That Looked Perfect
Heavy discount offer: “50% off your first 3 months.”
The metrics were incredible:
- 3.2% CTR (account average: 1.8%)
- $42 CPA (account average: $67)
- 4.2x platform ROAS (account average: 2.9x)
We were acquiring customers for 38% less than any other campaign. The executive team wanted to pour budget into it.
Then we looked at what happened after acquisition.
The Cohort Analysis That Changed Everything
Month 1 retention by acquisition source:
- Discount ad: 88% (great!)
- Standard messaging: 82% (fine)
Month 3 retention:
- Discount ad: 58% (hmm)
- Standard messaging: 79% (better)
Month 6 retention:
- Discount ad: 22% (disaster)
- Standard messaging: 68% (solid)
The discount ad was attracting customers who were price-sensitive, not value-aligned. They’d stick around for the discounted period, then churn the moment they hit full price.
The Unit Economics Nobody Was Watching
Let’s do the math on 100 customers:
Discount Ad Economics:
- Acquisition cost: $4,200 (100 customers × $42 CPA)
- Month 1-3 revenue: $2,500 (50% off $50/month × 100 customers × 3 months)
- Month 4-6 revenue: $3,300 ($50/month × 66 remaining × 3 months)
- Month 7-12 revenue: $1,320 ($50/month × 22 remaining × 6 months)
- Total 12-month revenue: $7,120
- Gross profit (40% margin): $2,848
- Net: -$1,352 (loss)
Standard Messaging Economics:
- Acquisition cost: $6,700 (100 customers × $67 CPA)
- Month 1-12 revenue: $40,800 ($50/month × weighted retention)
- Gross profit (40% margin): $16,320
- Net: +$9,620 (profit)
The “winning” ad was losing $13.50 per customer. The “expensive” ad was making $96 per customer.
Platform metrics told us to scale the discount ad. Business metrics told us to kill it.
Why This Happens More Than You Think
Attribution windows are too short.
Platforms optimize for 7-day or 30-day conversions. They have no idea what happens at month 3, 6, or 12.
An ad that drives high-intent, qualified customers might have:
- Higher CPA (they’re harder to acquire)
- Lower CTR (the messaging is specific, not broadly appealing)
- Lower short-term ROAS
But those customers stick around. They refer friends. They upgrade. They’re worth 3-5x what the “efficient” discount customers are worth.
Creative attracts intent, not just attention.
The discount ad got clicks because everyone likes a deal. But it attracted deal-seekers, not customers who valued the product.
The standard messaging (“Solve [specific problem] in 10 minutes”) got fewer clicks but attracted people with an actual need. Those people didn’t care about the price. They cared about the solution.
Platforms optimize for what they can measure.
Meta doesn’t see:
- Whether customers actually use your product
- Whether they renew after month 1
- Whether they expand or upgrade
- Whether they refer other customers
It sees: Did they click? Did they convert? Did they spend money in the attribution window?
That’s it.
If your success metrics and the platform’s success metrics aren’t aligned, you’re optimizing for the wrong thing.
How to Audit Your Own Ads
1. Segment Customers by Acquisition Creative
Most platforms let you pass creative IDs or campaign names through to your CRM. If you’re not doing this, start.
Tag every customer with:
- Campaign name
- Ad set name
- Creative ID (if possible)
2. Track Cohort Retention by Creative
Build a simple retention table:
Creative A:
Month 0: 100 customers
Month 1: 88 customers (88% retention)
Month 3: 79 customers (79% retention)
Month 6: 68 customers (68% retention)
Month 12: 58 customers (58% retention)
Creative B:
Month 0: 100 customers
Month 1: 94 customers (94% retention)
Month 3: 52 customers (55% retention)
Month 6: 22 customers (23% retention)
Month 12: 8 customers (8% retention)
If you’re a SaaS company, track trial → paid conversion by creative. If you’re e-commerce, track 30-day, 60-day, 90-day repeat purchase rates. If you’re lead gen, track lead → opportunity → close rates.
3. Calculate Actual LTV by Creative
Don’t use averages. Calculate actual observed LTV for each creative cohort.
Example for a SaaS company:
Discount Creative (6-month cohort):
- Average customer lifespan: 4.2 months
- Average revenue per customer: $174
- Gross margin: 70%
- Customer LTV: $122
Standard Creative (6-month cohort):
- Average customer lifespan: 11.8 months
- Average revenue per customer: $518
- Gross margin: 70%
- Customer LTV: $363
Now you know: Standard creative customers are worth 3x more than discount creative customers, even though discount creative has a “better” CPA.
4. Calculate CAC Payback Time by Creative
How long does it take for a customer acquired through each creative to become profitable?
Creative A:
- CAC: $67
- Monthly revenue: $50
- Gross margin: 70% ($35 margin per month)
- Payback: 2 months
Creative B:
- CAC: $42
- Monthly revenue: $25 (discounted first 3 months)
- Gross margin: 70% ($17.50 margin per month)
- Payback: Never (churns at month 4)
If a creative never pays back CAC before customers churn, kill it no matter what the platform metrics say.
5. Test Discount vs Value Messaging
Run this as an A/B test:
- Creative A: Value-based messaging (solve a problem, highlight differentiation)
- Creative B: Discount/promotional messaging (50% off, limited time deal)
Track both platform metrics AND cohort quality:
- Which has better CPA? (Platform cares)
- Which has better 6-month LTV? (You should care)
In my experience, value-based messaging usually has:
- 30-50% higher CPA
- 40-80% better LTV
- 2-3x better CAC payback economics
When to Kill a “Winning” Ad
Kill it if:
- Cohort LTV < CAC: Customers never become profitable
- Payback > average lifespan: Customers churn before paying back acquisition cost
- Retention significantly worse than baseline: More than 20% worse retention at 3-6 months
- Negative selection bias: The ad attracts the exact customers you don’t want (price-sensitive, low-intent, wrong segment)
What We Did Instead
We killed the discount ad. Reallocated that budget to three new creatives:
Creative 1: Problem-Solution “Tired of [specific pain point]? [Product] solves it in 10 minutes.”
- CPA: $71
- 6-month retention: 71%
- LTV: $380
Creative 2: Social Proof “Join 10,000+ [customer segment] using [Product] to [outcome]”
- CPA: $63
- 6-month retention: 66%
- LTV: $340
Creative 3: Differentiation “Unlike [competitor], [Product] gives you [unique benefit]”
- CPA: $78
- 6-month retention: 73%
- LTV: $410
All three had “worse” CPAs than the discount creative. All three had massively better LTV.
Total result after 6 months:
- Customers acquired: Down 22%
- Total acquisition spend: Down 18%
- 6-month gross profit: Up 64%
- Blended LTV: Up 85%
We acquired fewer customers. We made way more money.
The Lesson
Your best performing ad is only your best performing ad if the customers it brings in are worth keeping.
Platform metrics measure the top of the funnel. Business metrics measure the entire customer journey.
If those two things aren’t aligned, trust the business metrics.
A $42 CPA that brings in customers who churn in 4 months is worse than a $78 CPA that brings in customers who stay for 2 years.
Don’t optimize for metrics that lie.
Noah Manion is a fractional growth consultant specializing in marketing infrastructure, paid acquisition, and analytics. He’s spent 13+ years managing paid spend from $1K to $1M monthly and building cohort retention models that actually inform creative strategy. Find him at softpath.co.