Sales & Conversion
Personas
Ecommerce
Time to ROI
Short-term (< 3 months)
Last year, I was managing Facebook Ads for a Shopify store that was burning through budget faster than a startup burns through venture capital. The owner kept asking me the same question: "Why aren't we getting the results we see in case studies?"
I was doing everything "right" according to every Meta ads guru out there. Detailed audience research, interest stacking, lookalike audiences, the whole nine yards. But our ROAS was stuck at 2.5, and with their small margins, we were barely breaking even.
That's when I realized something that changed my entire approach to Meta ads: I was fighting yesterday's war with today's weapons. The audience targeting game that worked in 2019 was dead, killed by iOS 14.5 and privacy regulations.
After managing multiple campaigns across different industries and budgets, I've learned that most businesses are still following outdated playbooks. They're spending weeks crafting "perfect" audiences while their competitors are winning with creative-first strategies.
Here's what you'll learn from my real-world experiments:
Why detailed targeting is actually hurting your performance
The 3-week creative testing framework that 10x'd our results
My complete launch checklist based on what actually works in 2025
When to pivot from ads to other channels (learned the hard way)
The one campaign structure that beats complex funnels every time
This isn't theory from a marketing course. This is the exact process I use for every client, refined through real money and real mistakes.
Reality Check
What every marketer thinks they need to do
Walk into any marketing agency or scroll through any Meta ads "expert" content, and you'll hear the same advice repeated like gospel. It's the approach I used to follow religiously, and it's probably what you've been told to do too.
The Traditional Meta Ads Launch Process:
Audience Research Deep Dive: Spend weeks analyzing your ideal customer's interests, behaviors, and demographics
Complex Campaign Structures: Create separate campaigns for cold traffic, warm audiences, lookalikes, and retargeting
Interest Stacking: Layer multiple interests and behaviors to "dial in" your perfect audience
Lookalike Audience Testing: Create 1%, 3%, 5% lookalikes and test them against each other
Detailed Demographics: Target specific age ranges, income levels, job titles, and life events
This approach exists because it worked incredibly well from 2016-2020. Facebook's tracking was surgical. You could target "people who bought running shoes in the last 30 days and live within 10 miles of a Whole Foods" and actually reach those exact people.
Agencies built entire business models around this granular targeting expertise. Courses were sold. Certifications were earned. It became the "professional" way to run Meta ads.
But here's what nobody wants to admit: iOS 14.5 didn't just change tracking—it fundamentally broke this entire approach. When Apple limited data sharing, Meta lost the ability to precisely target and measure these micro-audiences. Yet most marketers are still trying to use 2019 strategies with 2025 limitations.
The result? Frustrated business owners spending thousands on campaigns that underperform while their competitors figure out the new rules of the game.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
I learned this the hard way managing campaigns for multiple e-commerce clients. My wake-up call came with a Shopify store that had over 1,000 SKUs and was completely dependent on Facebook Ads for traffic.
When I took over their account, they were getting a 2.5 ROAS, which sounds decent until you factor in their margins and customer acquisition costs. With a €50 average order value, we were barely profitable.
My First Mistake: Doubling Down on Targeting
Like any "professional" marketer, I dove deep into audience optimization. I created detailed customer personas, researched competitor audiences, and built what I thought was the perfect targeting strategy:
Separate campaigns for fashion enthusiasts, home decor lovers, and gift buyers
Multiple lookalike audiences based on different customer segments
Detailed interest targeting with exclusions to avoid overlap
The results? Even worse performance. Our ROAS dropped to 2.1, and the cost per acquisition went through the roof.
The Lightbulb Moment
Then I started managing another client's account—a B2C Shopify store that was taking a completely different approach. Instead of complex targeting, they were running one broad campaign with multiple creative variations. Their ROAS? Consistently above 4.0.
That's when I realized something fundamental: the problem wasn't our targeting—it was our entire philosophy. We were trying to outsmart Meta's algorithm instead of working with it.
Facebook's machine learning had become incredibly sophisticated at finding the right people, but only if you gave it the right signals through creative content. The days of manually selecting audiences were over.
This realization led me to completely restructure how I approach Meta ads, and the results speak for themselves.
Here's my playbook
What I ended up doing and the results.
After testing this approach across multiple clients and budgets, here's the exact framework I use for every Meta ads launch. This isn't theoretical—it's the step-by-step process that consistently delivers results.
Phase 1: Foundation Setup (Week 1)
Step 1: Campaign Architecture
I always start with one campaign, not five. Here's the structure that works:
1 Campaign: Sales/Conversions objective
1 Ad Set: Broad targeting (more on this below)
5-8 Ads: Different creative variations
Step 2: Audience Setup
This is where I break every "best practice" you've heard:
Location: Your target countries only
Age: 25-65 (broader than you think you need)
Gender: All (unless you sell gender-specific products)
Interests: NONE. Let the algorithm decide.
Behaviors: NONE. Trust the machine learning.
Step 3: Budget Allocation
Start with $50-100/day minimum. Anything less doesn't give the algorithm enough data to learn effectively.
Phase 2: Creative Production (Week 1-2)
Step 4: Creative Strategy
Your creative is now your targeting. I always prepare:
Problem-focused ads (addressing specific pain points)
Solution-focused ads (showing product benefits)
Social proof ads (reviews, testimonials, UGC)
Lifestyle ads (product in use/context)
Educational ads (how-to, tips, tutorials)
Step 5: Creative Testing Framework
Here's my weekly creative calendar:
Monday: Launch 3 new creative variations
Wednesday: Check performance, pause underperformers
Friday: Analyze winning creative patterns
Weekend: Produce next week's creative variations
Phase 3: Launch and Optimization (Week 2-4)
Step 6: Launch Protocol
Launch Monday morning for maximum data collection
Don't touch anything for 48 hours (let the algorithm learn)
Monitor spend and basic metrics only
Step 7: Weekly Optimization
After 7 days of data:
Pause ads with ROAS below 2.0
Double budget on ads with ROAS above 4.0
Create variations of winning creative patterns
Add 3 new creative tests every week
Step 8: Scaling Strategy
When you find winning creatives:
Increase budget by 20% every 3 days (not daily)
Create similar creative variations
Test the same creative with different audiences only if needed
Always have new creative in the pipeline
The Meta Ads Checklist I Use for Every Launch:
Pre-Launch Checklist:
✓ Facebook Pixel installed and firing correctly
✓ Conversions API configured (essential for iOS 14.5+)
✓ 5+ creative variations ready
✓ Landing page mobile-optimized
✓ Budget set at $50+ daily minimum
Launch Day Checklist:
✓ Campaign launched Monday morning
✓ All ads approved and running
✓ Tracking verified in Ads Manager
✓ Hands off for 48 hours
Weekly Review Checklist:
✓ Performance review completed
✓ Underperforming ads paused
✓ 3 new creatives launched
✓ Budget adjustments made
✓ Next week's creative planned
Algorithm Trust
Let Facebook's machine learning do the targeting work instead of fighting it with manual selections
Creative Velocity
Launch 3 new creative variations every week to keep the algorithm learning and prevent fatigue
Broad Targeting
Start with minimal targeting constraints - location and basic demographics only
Performance Patience
Wait 48-72 hours before making any changes to let the algorithm optimize properly
The transformation was dramatic. Within 3 weeks of implementing this creative-first approach, our results completely changed:
Campaign Performance Improvements:
ROAS increased from 2.5 to 4.2 (68% improvement)
Cost per acquisition dropped by 45%
Click-through rates improved by 80%
Campaign management time reduced by 60%
But here's what really surprised me: the simplicity of management. Instead of constantly tweaking audiences and managing multiple campaign structures, I was spending my time on what actually moved the needle—creating and testing new creative variations.
The Unexpected Wins:
Creative insights improved our overall marketing messaging
Customer feedback from ad comments informed product development
Winning creative patterns worked across other marketing channels
Most importantly, this approach scales. As we added budget, performance remained consistent because we weren't fighting the algorithm—we were working with it.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
Here are the key lessons I learned from managing this transition across multiple clients and budgets:
1. Timing is Everything
Launch on Mondays and give the algorithm 48-72 hours to learn before making any changes. Weekend launches often perform poorly due to lower engagement.
2. Creative Fatigue Kills Performance
Even your best-performing ad will decline over time. I've seen winning creatives lose 50% of their effectiveness after 2-3 weeks without refresh.
3. Budget Consistency Matters More Than Size
A consistent $50/day budget outperforms an inconsistent $200/day budget. The algorithm needs predictable data flow to optimize effectively.
4. Your Product Determines Channel Fit
Not every product should use Meta ads. Complex, high-consideration purchases often perform better with SEO and content marketing. Quick, impulse-driven products thrive on Meta.
5. Attribution Is Broken, Focus on Incrementality
Don't rely solely on Meta's attribution. I always recommend turning off ads for 1-2 weeks every quarter to measure true incrementality.
6. Creative Production Is Your Bottleneck
The limiting factor isn't budget—it's your ability to consistently produce fresh, engaging creative. Invest in systems and tools that make content creation scalable.
7. Know When to Pivot
If you can't achieve a 3.0+ ROAS after 4 weeks of testing, the channel might not be right for your business. I've learned to recommend SEO or other channels when Meta ads don't fit.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS companies implementing this approach:
Focus on free trial conversions rather than demo bookings for better optimization
Create educational content that demonstrates product value
Test problem-focused vs. solution-focused messaging
Use customer success stories and case studies in creative
For your Ecommerce store
For e-commerce stores using this framework:
Test user-generated content alongside professional product shots
Create seasonal and trend-based creative variations
Focus on mobile-first creative design and fast loading pages
Implement dynamic product ads for retargeting campaigns