Sales & Conversion
Personas
Ecommerce
Time to ROI
Short-term (< 3 months)
When I started managing Facebook Ads for a B2C Shopify store, I was convinced I'd cracked the code. I spent weeks meticulously crafting different audience segments, targeting specific demographics, interests, and behaviors like a master chess player positioning pieces on the board.
The results? Mediocre at best. We were burning through budget testing different audience combinations, and our ROAS wasn't improving. Sound familiar?
Then I discovered something that completely flipped my Meta ads strategy upside down: creatives are the new targeting. Instead of trying to outsmart Facebook's algorithm by manually selecting audiences, I learned to trust the platform's machine learning capabilities—but only after feeding it the right creative signals.
Here's what you'll learn from my journey:
Why audience targeting is largely dead in 2025 (and what replaced it)
The simple framework that turned my underperforming campaigns around
How to test 3 new creatives weekly without breaking your budget
The meta ads optimization strategy that works for both B2B SaaS and ecommerce
Why I now run one big campaign instead of multiple audience-segmented campaigns
Industry Reality
What every marketer thinks they need to do
Walk into any Meta ads course or agency pitch, and you'll hear the same advice: "Targeting is everything."
The conventional wisdom goes like this:
Create detailed audience personas - Map out your ideal customer's demographics, interests, and behaviors down to their favorite coffee brand
Build multiple campaign segments - Create separate campaigns for different audience groups, ages, and interests
Test audiences, not creatives - Run the same ad to different audience segments to find your "perfect" target
Lookalike audiences are gold - Upload your customer list and let Facebook find similar people
Exclude and include precisely - Fine-tune your targeting with detailed interest exclusions and inclusions
This approach made sense back when Facebook's targeting controls were more granular. Marketers could isolate specific audience segments and control exactly who saw their ads.
But here's the uncomfortable truth: Meta's algorithm has evolved beyond the need for manual audience targeting. The platform now optimizes for conversions so effectively that our "control" over targeting has become largely illusory. Most targeting inputs are now just suggestions that the algorithm can—and will—ignore if it finds better-performing audiences.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
When I took over Facebook Ads for a B2C Shopify store, I walked right into this targeting trap. The client sold consumer electronics with an average order value of around €50, and their previous marketer had set up what looked like a sophisticated targeting strategy.
We had separate campaigns for different age groups, interests like "tech enthusiasts" and "gadget lovers," and carefully crafted lookalike audiences based on their customer data. Each campaign had its own budget, and we were constantly shifting money between them based on performance.
The numbers looked decent on paper, but something felt off. Our ROAS was stuck around 2.5, and we were spending more time managing campaign budgets than actually improving performance.
I spent the first month doing what every marketer does—tweaking audiences, adjusting age ranges, testing new interests. I was convinced the right combination would unlock better performance. It didn't.
Then I read something that changed my perspective completely: 70-80% of Meta ad performance now stems from creative quality, not targeting. The algorithm had become so sophisticated at finding the right people that our main job was no longer audience selection—it was giving Facebook enough creative options to work with.
That's when I decided to completely restructure our approach and test what I now call "creative-first optimization."
Here's my playbook
What I ended up doing and the results.
I restructured our entire Meta ads approach around one core principle: give the algorithm multiple creative options within one broad campaign, instead of multiple audience options across different campaigns.
Here's the exact framework I implemented:
Step 1: Campaign Consolidation
Instead of running 5-6 campaigns targeting different audience segments, I created one main campaign with broad targeting (all countries we shipped to, ages 25-65, all genders). No detailed interests, no lookalike audiences, no manual restrictions.
Step 2: Creative Testing Rhythm
Every single week, without fail, I produced and launched 3 new creative variations. This wasn't random—each creative tested a different angle:
- Problem-solving focus ("Finally, a solution for...")
- Lifestyle-focused content (showing the product in use)
- Social proof and testimonials
Step 3: The 7-Day Testing Window
I let each new creative run for exactly 7 days alongside existing performers. The algorithm needed this time to optimize and find the right audience for each creative style. After 7 days, I'd analyze which creatives were getting the best ROAS and turn off the underperformers.
Step 4: Creative Lifecycle Management
Instead of campaign budgets, I managed creative lifecycles. High-performing creatives got more budget allocation (through Meta's automatic optimization), while fatigued creatives got replaced with fresh variations of the same angle.
Step 5: Advantage+ Integration
I used Advantage+ placements and creative optimization to let Meta automatically adjust our ads for different placements (Feed, Stories, Reels) and audiences. The algorithm was better at this than manual adjustments.
The breakthrough insight was this: each creative acts as a signal to the algorithm about who might be interested in your product. A lifestyle-focused creative attracts different people than a problem-solving creative, all within the same broad campaign structure.
Weekly Rhythm
Consistency beats perfection. Set a non-negotiable weekly creative production schedule and stick to it, even if some creatives aren't perfect.
Broad Targeting
Use one broad campaign instead of multiple audience-segmented campaigns. Let Meta's algorithm find your audience through creative signals.
Creative Angles
Test different creative approaches: problem-solving, lifestyle, social proof. Each angle attracts different customer segments naturally.
Performance Tracking
Focus on ROAS per creative, not per audience. Track which creative angles consistently perform best for your product category.
The results spoke for themselves. Within 30 days of implementing this creative-first approach:
ROAS improved from 2.5 to 3.8 - The same budget was generating significantly better returns because the algorithm was optimizing for actual conversions, not just cheap actions within restricted audiences.
Management time decreased by 60% - Instead of constantly adjusting budgets between audience campaigns, I spent my time on high-impact creative production.
Audience insights improved - Counter-intuitively, letting the algorithm choose audiences gave us better data about who was actually buying. We discovered customer segments we never would have targeted manually.
Most importantly, we found creative angles that consistently outperformed others. Problem-solving creatives delivered the highest ROAS, while lifestyle content drove higher engagement but lower conversions. This insight shaped our entire content strategy going forward.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
After implementing creative-first optimization across multiple campaigns, here are the key lessons that changed how I approach Meta ads:
The algorithm is better at targeting than you are - Stop fighting it with detailed restrictions. Broad targeting + good creatives beats narrow targeting + average creatives every time.
Creative fatigue happens faster than audience fatigue - Your audience doesn't get tired of seeing ads, they get tired of seeing the same creative. Fresh angles matter more than fresh audiences.
Quality beats quantity in creative testing - Three well-thought-out creatives per week outperform ten random variations. Each creative should test a distinct angle or approach.
Patience is required for algorithm learning - Give new creatives at least 7 days to optimize. The algorithm needs time to find the right audience for each creative style.
Campaign structure matters less than creative structure - Simplify your campaigns and complicate your creative strategy, not the other way around.
ROAS per creative is the metric that matters - Track performance by creative angle, not audience segment. This tells you what messaging resonates with your actual customers.
Manual optimization should focus on creative, not audience - Spend your time improving creative quality and testing new angles, not adjusting audience parameters.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS companies, apply this by creating broad campaigns and testing different creative angles weekly:
Problem-solving creatives highlighting specific pain points
Feature demonstration videos showing your product in action
Customer success stories and testimonials
Use case scenarios for different industries or roles
For your Ecommerce store
For ecommerce stores, focus on product-centric creative testing:
Product demonstration videos showing benefits in action
User-generated content and customer reviews
Lifestyle images showing products in real-world contexts
Before/after comparisons or problem/solution formats