AI Performance Marketing · Social Media Performance Advertising
Audience Strategy With AI
AI-driven audience strategies
Welcome back to TechFlow!
Your Google campaigns are optimized and performing, but at today's team meeting, your CMO assigned you a new challenge, "Google Search works, but we need to expand reach. Let's test Meta ads."
In this lesson, you'll discover who we should target, how Meta's audience targeting changed, why broad targeting often beats narrow segmentation, and how to design audience strategies based on testing rather than assumptions.
How Meta Targeting Changed
Before iOS 14.5, Meta tracked users across apps and websites using pixel data, enabling precise targeting like competitor site visitors.
After the update, most iOS users opted out of tracking. Meta now relies on inferred interests and on-platform behavior instead of detailed browsing history.
This means TechFlow can't micro-target "CMOs at $10M+ SaaS companies who read marketing blogs" anymore.
Instead, you target broader categories such as "marketing professionals" or "business software," and Meta's algorithm identifies who actually engages and converts within that pool.
Select all that apply
Why do you think broad targeting often works better than narrow targeting on Meta now?
Now that you understand why broad often beats narrow, let's first explore narrow-audience recommendations.
practice preview
Interactive practice
Fill in the blank
Ask ChatGPT to build a narrow audience for TechFlow.
Next, let's create a broad targeting approach and compare results.
practice preview
Interactive practice
Fill in the blank
Now, ask DeepSeek to build a broad audience for TechFlow.
Choose one
What is the key difference between these two approaches?
Yes! That's how they differ. Narrow targeting limits Meta's ability to learn. But broad targeting alone isn't enough — you need strategic hypotheses about why people buy.
Psychographic Hypotheses vs. Demographics
- Demographics tell you who people are.
- Psychographics tell you why they buy.
A 40-year-old marketing manager at a growing startup has different motivations than a 40-year-old marketing manager at an enterprise company — same demographics, completely different problems and buying triggers.
Let's practice building demographic versus psychographic strategies to see which yields clearer targeting.
practice preview
Interactive practice
Fill in the blank
Ask DeepSeek to create a demographic audience profile.
But how does a psychographic profile differ in practice? Let's ask about the same customer and compare with Claude.
practice preview
Interactive practice
Fill in the blank
Ask Claude to create a psychographic audience profile.
Choose one
Which profile, in your opinion, gives TechFlow better creative direction?
Great Work!
Yes, psychographic profiles "drowning in manual tasks, need to scale" create actionable creative strategies, while demographics "30-50, Marketing Manager" just define audience parameters.
Instead of one generic psychographic profile, create three distinct hypotheses representing different customer motivations. Then test which resonates.
practice preview
Interactive practice
Fill in the blank
Ask Claude to create testable audience hypotheses.
Perfect!
You've created three distinct hypotheses — each with a different pain point, trigger, and creative angle.
Now let's see how to structure these for Meta Advantage+ to test which hypothesis converts best.
Structuring for Advantage+ Discovery
Meta Advantage+ works like Google's Performance Max — you provide creative variations and broad audience signals, then the algorithm discovers which combinations convert.
The key: one broad campaign with multiple creative variations, one per hypothesis, not separate narrow campaigns per hypothesis.
practice preview
Interactive practice
Fill in the blank
Ask Claude to create a campaign structure for hypothesis testing.
Choose one
But why do you think you should test all three hypotheses in one broad campaign instead of three separate campaigns?
Excellent!
You understand how to structure TechFlow's Advantage+ campaign — broad audience (500K+), three creative variations, one per hypothesis, equal budget split initially, then shift to winners.
But how to read performance signals early and decide when to act? Let's figure it out!
After one week, TechFlow's Advantage+ campaign shows different performance across the three creative hypotheses you built:
- "Scaling Strugglers": 2.8% CTR, 9 conversions, $45 CPA
- "ROI Provers": 1.1% CTR, 3 conversions, $120 CPA
- "Efficiency Seekers": 0.6% CTR, 0 conversions, $95 spend
Most marketers stop low performers. But one-week data is directional, not conclusive. Let's see how to interpret these signals.
practice preview
Interactive practice
Fill in the blank
Ask DeepSeek to analyze early Advantage+ results.
Choose one
Why do you think DeepSeek suggested to pause "Efficiency Seekers" after zero conversions in 7 days, not stop it?
Great Work!
You've learned to read early signals with patience — shift budget to clear winners, but don't kill potential winners prematurely.
But when you double the budget, Meta is forced to show the same creative to less-interested prospects — people who saw it once and didn't click, or audiences further from your ideal profile.
So the result: your CPA climbs from $45 to $75+ because you're buying lower-quality impressions with the same message. Instead, you need creative iteration — expressing the same "Scaling Strugglers" insight in different ways to expand reach while maintaining message fit.
practice preview
Interactive practice
Fill in the blank
Ask Claude to generate variations of the winning hypothesis.
Perfect!
You've generated five variations of "Scaling Strugglers" — each expressing "automate before hiring" differently with specific hooks.
Scale by creative iteration, not just budget. Generate multiple variations of winning hypotheses to reach new audience segments without creative fatigue. And use Advantage+ to automatically allocate budget to top-performing variations.
- Broad targeting wins post-iOS: Meta needs 500K+ audience to find patterns. Narrow limits learning.
- Psychographics drive creative: "Drowning in manual tasks" makes a strong headline. "30-50, Marketing Manager" doesn't.
- Test hypotheses together: One campaign, three creatives, let's Advantage+ optimize fast. Separate campaigns take weeks.
- Scale via creative, not just budget: Generate 5+ variations of winning angle to expand reach without fatigue.
Congratulations!
You've practiced building narrow vs. broad strategies, demographic vs. psychographic profiles, and learned to interpret early performance signals with patience.
You can now answer CMO's challenge, "Who should we target on Meta?" Your answer: Start broad with Advantage+ targeting marketing professionals interested in B2B software, letting Meta's algorithm find your best customers. Layer retargeting for website visitors. Broad wins initially — narrower ones later, based on evidence, not assumptions.
In the next lesson, you'll discover how format and placement affect Meta's delivery algorithm, why automatic placements outperform manual selection, and how to create format-native creative for Feed, Stories, and Reels. See you at the next assignment!
