Media Buying Mentorship

Profit Lab Agency - Media Buying Mentorship

Profit Lab Agency — Mentorship Program

MEDIA BUYING

PERFORMANCE MARKETING

A structured framework for testing, optimizing, and scaling on Meta Ads

Contents

Session 1 — Finance for Ecommerce

Welcome • Why Finance First • Primary Metrics

Breakeven ROAS • Profitability Scenarios • Profit Per Order

Pricing Strategy • Offers • Cash Flow • Homework

Quick Rules — Golden rules before you touch anything

Phase 1 — Testing

1.1 Creative Testing — ABO

1.2 Creative Testing — CBO with Min Spend

1.3 Creative Testing — Single Ad per Adset

1.4 Flexible Format Testing

1.5 Andromeda Media Diversity (CBO Single Adset)

1.6 Audience Testing — 4-Adset Framework

1.7 Audience Breakdown Testing

1.8 Bidding Strategy Testing (CPG & Min ROAS)

1.9 Catalog Testing

1.10 Breakdown Campaigns (Age, Gender, Platform)

Phase 2 — Optimization

2.1 Creative Trimming & Thumbnails

2.2 When to Kill vs. When to Wait

2.3 Common Mistakes

Phase 3 — Scaling

3.1–3.6 Scaling Methods & Budget Strategies

Phase 4 — Creative & Media Directions

4.1 Why Creative Beats Media Buying

4.2 Format × Stage Freedom

4.3 Creative Formats (Static, Carousel, Video, GIF, DPA)

4.4 Creative Angles

Before You Start

Quick Rules

Memorize these before you touch a single campaign

🔒

Rule #1 — Never Kill a Winner. Duplicate It.

When an ad or adset is performing well, NEVER edit the original. Don't change its budget, creative, audience, or anything. Instead, duplicate it and make your changes on the copy. Leave the original running untouched.

"متلعبش فالجايب" — Don't mess with what's already working.

⚖️

Rule #2 — Don't Pump or Punish Without Volume

Aggressive budget changes (pumping up or cutting down) only make sense when you have very high order volume — at least ~50 orders/day (not a fixed number, depends on your business). Without volume, budget swings destabilize Meta's learning and give you unreliable data.

💡	Rule #3 — Test ONE Variable at a Time When testing creatives — keep targeting and product the same. When testing audiences — keep creatives the same. Breaking this rule means you'll never know what actually caused the result.

⏳	Rule #4 — Don't Kill Before It Spends Your Avg CPP Never pause an ad or adset until it has spent at least your average cost per purchase. Killing too early = killing potential winners before they've had a fair test.

🎥	Rule #5 — Preferably Separate Videos and Images It's better to keep videos and images in separate adsets. Mixing is possible but not ideal — Meta tends to favor one format over the other and you lose clarity on what's actually performing.

💰	Rule #6 — Budget Per Media Asset = 0.75× Target CPP Each media asset in an adset needs at least 0.75× your target CPP in daily budget. 3 assets + 200 EGP CPP = ~450 EGP/adset. 3 assets + 300 EGP CPP = ~675 EGP/adset. Range: 450-700 EGP depending on your CPP target.

🏆	What is a "Winner"? A creative, audience, or product with the highest ROAS and a good number of purchases. Not just low cost — it must have volume too. A winner at 1 purchase is not a winner.

Phase 1

Testing

Find winning creatives, audiences, products, and bidding strategies

The testing phase is mainly to find winners. Test creatives first, then audiences, then bidding, then catalog. These are guidelines, not absolute rules — adapt based on your budget, product, and market.

Creative Testing Strategies

1.1

ABO Campaign — Budget Per Adset

Budget is set at the adset level. Each adset gets an equal, fixed daily budget. All adsets use broad targeting. You're only testing creatives.

◆ Diagram — ABO Creative Testing

📌 ABO CAMPAIGN Budget set per adset • Broad targeting
▼
ADSET 1 🎥 Videos	ADSET 2 🎥 Videos	ADSET 3 📸 Images	ADSET 4 📸 Images

1.2

CBO with Minimum Spend Limit

Budget at the campaign level. Meta distributes it, but you set a minimum daily spend per adset to guarantee each batch gets enough data. Same broad targeting, same separation of videos/images.

◆ Diagram — CBO with Minimum Spend

📌 CBO CAMPAIGN Meta distributes • Min spend = safety net
▼
ADSET 1 🎥 Videos Min: spend/day	ADSET 2 🎥 Videos Min: spend/day	ADSET 3 📸 Images Min: spend/day	ADSET 4 📸 Images Min: spend/day

1.3

Creative Testing — Single Ad per Adset

Each creative gets its own adset. This guarantees equal spend per creative — the fairest comparison. Requires higher budget but gives you the most accurate data.

✅ Pro:	Equal spend per creative = fair judgment
❌ Con:	Requires higher budget (more adsets = more budget needed)

1.4

Flexible Format Testing

Selected at the ad level. Upload multiple media into one ad — Meta picks the best media + placement combo. Best used for winner creatives after initial testing. The drawback: no per-creative breakdowns, so you won't know which specific creative won.

◆ Diagram — Flexible Format

✅ Good Example All videos — same product, different angles 🎥 Unboxing 🎥 Demo 🎥 Review	✅ Good Example All images — same product, color variants 📸 Pink 📸 Blue 📸 Black
⚠ Not Preferred Mixing 🎥 + 📸 in one flexible ad	⚠ Not Preferred Unrelated products (T-shirt + Jeans)

1.5

Andromeda Media Diversity (CBO Single Adset)

A great starting point. Single CBO campaign, one adset, all media loaded. Andromeda's ad ranking engine gives every new ad initial delivery to assess its quality signal. The more diverse creatives you feed it, the faster it identifies winners. Run 3-5 days, then extract winners for structured tests.

◆ Diagram — Andromeda Diversity Method

📌 CBO CAMPAIGN

▼

SINGLE ADSET — Broad

▼

ALL ADS (each = 1 media asset)

🎥 V1 🎥 V2 🎥 V3 📸 I1 📸 I2

Andromeda gives all ads initial delivery. Run 3-5 days, extract winners.

Audience Testing Strategies

⚠ Prerequisite: Use ONLY winning media from creative testing. Same rules apply. Each adset = 1 audience segment.

1.6

The 4-Adset Audience Framework

#	Adset	Audience	Purpose
1	Retargeting	All custom audiences. Exclude purchasers. Use personalized copy + special offer.	Re-engage warm traffic
2	Lookalike	LAL from best custom audiences	Find similar new people
3	Broad	No targeting (can select age/gender if essential)	Let Meta find buyers
4	Exclusion	Broad minus all custom audiences	Pure cold traffic only

Key: Broad can still reach warm audience. Exclusion forces truly cold prospecting. They are NOT the same.

◆ Diagram — Audience Testing

Same winning media in all adsets
🎯 RETARGET	👥 LOOKALIKE	🌎 BROAD	❌ EXCLUSION

1.7

Audience Breakdown Testing

Once a segment wins (e.g., Lookalike), break it down further. Each sub-segment gets its own adset with same winning media.

LAL Purchase

Highest quality

LAL Purchase

High quality

LAL Purchase

Good volume

LAL Purchase

Max volume

Also works for: LAL Add to Cart, LAL Website Visitors, or custom audiences by time window (7d vs 30d vs 90d).

Bidding Strategy Testing

1.8

CPG & Min ROAS Testing

Use winning media + winning audience. Only variable = bid level. Take your average from last 30 days and create 4 adsets at 0.5x, 1x, 1.5x, and 2x.

◆ CPG Test (Example: Avg CPP = 200 EGP)

0.5×

100 EGP

Best ROAS
Low delivery

1×

200 EGP

Normal ROAS
Baseline

1.5×

300 EGP

Higher volume
Lower ROAS

2×

400 EGP

Max volume
Lowest ROAS

◆ Min ROAS Test (Example: Avg ROAS = 7.0)

3.5

Min ROAS

Max delivery

7.0

Min ROAS

Baseline

10.5

Min ROAS

Efficient

14.0

Min ROAS

Very strict

Catalog & Product Testing

1.9

Catalog Testing Campaign

For e-commerce brands with multiple products. The goal is to test product sets / collections to find your best sellers. Use winner creatives and audiences — only variable is the product set. ABO or CBO, at least 2 adsets.

#	Adset	Product Set	Ads Inside
1	Best Sellers	Your top-performing products only	Carousel + Single Image + Video Collection Storefront
2	All Products	Your full catalog	Carousel + Single Image + Video Collection Storefront

◆ Diagram — Catalog Testing

📌 CATALOG CAMPAIGN (ABO or CBO)
▼
⭐ BEST SELLERS Top products only Carousel Single Img Video Storefront	📦 ALL PRODUCTS Full catalog Carousel Single Img Video Storefront

Find the best-selling products, then consolidate into a single adset to maximize ROAS.

1.10

Breakdown Campaigns

Use Meta's breakdown data to test age groups, genders, and platforms separately. Create dedicated adsets for top segments.

Breakdown	What You Learn
Age	Which age group buys most / has lowest CPP
Gender	Male vs Female performance
Platform	Facebook vs Instagram vs Audience Network

Phase 2

Optimization

Trim the losers, protect the winners, improve efficiency

2.1

Creative Trimming & Thumbnails

Regularly review at the individual video/image level. Kill underperformers (above CPP threshold after sufficient spend). Test different video thumbnails — the thumbnail is the first thing people see, swapping it can dramatically change CTR without new video content.

Remember Rule #1: never edit a winning ad directly. Duplicate first, then modify the copy.

2.2

When to Kill vs. When to Wait

❌

Kill It

Spent more than your avg CPP with zero or very poor results. Clear loser.

⏳

Wait

Hasn't spent your avg CPP yet. Too early to judge. Let it run.

2.3

Common Mistakes

❌

Multiple Variables

Changing creative AND audience at once

❌

Killing Too Early

Before spending avg CPP

❌

Letting Losers Run

High spend, no results, still active

Phase 3

Scaling

Grow profitably with proven winners

Scaling Methods

● scaling board — 6 plays

🏆 WINNING ADS & AUDIENCES

pick a play below

3.1 · CBO

All Winners, One Adset

CBO → 1 adset → every winning ad inside

[ CBO 2k ]

📌 Adset

🏆 🏆 🏆 🏆 🏆

Why: Meta auto-picks the best winner by user.

3.2 · ABO

One Winner, Own Adset

ABO → multi-adset → each winning creative = own adset

[ ABO ]

Adset 🏆

1 winner per adset

Why: Pure per-winner control — guaranteed spend each.

3.3 · PUMP

Daily 20% Vertical Scale

Raise the daily budget of an already-winning campaign by 20% each day

Day 1: 1000 → Day 2: 1200

Day 3: 1440 → Day 4: 1728

📈

Why: Safest way up — keeps Meta's algo stable.

│

3.4 · CBO FLEX

Flexible Winners Bundle

CBO → 1 adset → multiple Flexible ads, each holding several winners

[ CBO → Adset ]

Flex Ad A
🏆🏆🏆

Flex Ad B
🏆🏆🏆

Flex Ad C
🏆🏆

Why: Double algo: CBO picks best ad, Flex picks best placement + creative combo.

3.5 · ABO AUD

Audience-Split × All Winners

ABO → multi-adset → each adset = 1 winning audience + every winning ad inside

[ ABO ]

Adset 👥 LAL 🏆🏆🏆	Adset 🏡 RT 🏆🏆🏆
Adset 🌐 Broad 🏆🏆🏆	Adset 🔍 INT 🏆🏆🏆

Why: See which audience converts best — same ads, clean comparison.

3.6 · REMIX

Winner Variations Lab

CBO → 1 adset → same winner in multiple versions (trimmed differently or new thumbnails)

[ CBO → Adset ]

🏆 v1 Full 30s	🏆 v2 Trim 15s
🏆 v3 Thumb A	🏆 v4 Thumb B

Why: Refresh a fading winner without rebuilding — keep learnings, buy back CTR.

legend

3.1 CBO — all winners in one adset	3.2 ABO — one winner per adset	3.3 PUMP — 20%/day vertical scale
3.4 CBO FLEX — multi Flex ads per adset	3.5 ABO AUD — per-audience, all winners	3.6 REMIX — trim + thumbnail variants

3.1

CBO Single Adset — All Winners Inside

One CBO campaign, one adset, every winning ad inside. Meta routes budget to whichever winner is hot at the moment. Lowest operational overhead.

3.2

ABO Multi-Adset — One Winner Per Adset

ABO campaign with one adset per winning creative. Each winner gets a guaranteed daily budget — no cannibalisation. Ideal when you want clean per-creative data at scale.

3.3

20% Daily Vertical Pump

Don't touch structure — just raise daily budget of an already-winning campaign by up to 20% every day. Keeps Meta's algo stable and avoids CPP spikes from sudden budget jumps.

3.4

CBO Single Adset — Flexible Winners Bundle

CBO + one adset + multiple Flexible ads, each bundling a handful of winners. Meta picks the best ad (CBO layer) AND the best placement/creative combo (Flexible layer). Double-optimised.

3.5

ABO Multi-Adset — Audience Split, All Winners

ABO campaign. Each adset targets one proven audience (LAL, RT, Broad, INT). Inside every adset: all your winning ads. Shows you exactly which audience scales best with the same creative set.

3.6

CBO Single Adset — Winner Variations (Trim & Thumbnail)

One winning ad, multiple versions. Each version is either trimmed differently (15s, 10s, 6s) or has a different thumbnail. All variants live in one CBO adset — Meta picks whichever version still has attention left.

Trim variants: rebuild pacing, tighten hook
Thumbnail variants: swap to a new stop-scroll frame
Refreshes a fading winner without losing learnings

Budget Scaling Strategies

Pump & Punishment

When performing well, pump the budget. If performance drops after the pump, punish immediately by cutting back. Only use when you have high order volume (~50+/day).

📈

Performing ↑ PUMP

👀

Monitor 24-48h

🔴

CPP rising ↓ PUNISH

Increase Budget by 20% Per Day

The safest way to scale. No more than 20% per day. Keeps Meta's algorithm stable and avoids CPP spikes.

Day 1

500

➔

Day 2

600

+20%

➔

Day 3

720

+20%

➔

Day 4

864

+20%

Phase 4

Creative & Media Directions

Build creatives that work standalone — the funnel is dead

4.1 · INTRODUCTION

Creative > Media Buying

80% of performance comes from the creative. Tweaks to budgets, audiences, and bid strategies move the needle — but creatives are the needle.

Goal of this phase: pick the right format and angle for your product — and let each creative perform on its own.

4.2 · FORMAT × STAGE FREEDOM

Don't Lock Formats to Funnel Stages

The old TOFU / MOFU / BOFU segmentation is loose, not law. You don't need to dedicate a format or angle to one stage of the funnel.

🎯 Any format and any angle can work at any stage.

The only thing that changes is the content of the creative.

A static image can be a hook, a proof point, or a closer — depending on what's in it. Same with video, carousel, GIF, DPA. Pick the format that fits your product and the message you want to deliver, then write the content for the moment.

4.3 · Creative Formats

Five formats — pick the one that fits your product, not your funnel stage

Static Images

Quick · Scroll-stopping · Easy to test offers, USPs, hooks

Best for: testing hooks fast, offer-driven creatives, seasonal promos.

● Real-world example — Gre Leaves Detox Shampoo

Same product, same format, two different angles. Each one stands on its own.

EXAMPLE 1 · STORY HOOK

"She detoxed her scalp. Everything else followed."

→ Emotional hook + ingredient transparency (Charcoal, Vitamin E, Peppermint, Ginger). Builds curiosity & trust in one frame.

EXAMPLE 2 · BENEFIT STACK

"Your scalp needs a break too."

→ Pain-point hook + benefit list (cleanses, removes oil, soothes, fights dandruff). Direct, confident, conversion-focused.

💡 Lesson: Same static format. Same product. Two completely different vibes. Either could run alone — or both together to cover more buyer types.

B	Carousels Multiple slides — products, features, steps Showcase multiple products, features, or steps. Great for catalogs, bundles, and storytelling sequences. Best for: bundles, multi-feature products, comparison & storytelling.

Video Ads

The heaviest hitter — built to carry your funnel end-to-end

UGC / Testimonial — raw, relatable, trust-building. Pro tip: avoid making it look too professional. "Crazy good" production hurts trust.

⭐ TOP TIER FGC — Founder Generated Content — the founder on camera talking directly to the customer. One of the best media types — builds the highest level of trust because it puts a real face and a real promise behind the brand. Customers don't just see a product, they see the person who stands behind it.

Product Demo — benefits, how-to, unboxing.

Story / Emotional angle — lifestyle storytelling, brand positioning.

Best for: Trust building (UGC, FGC), demoing benefits, brand storytelling. Each video should hook + prove + close on its own.

D	GIFs / Motion Graphics Hybrid between static and video Showcase offers to drive direct sales. Perfect for quick features — works best for fashion (before/after outfits) or seasonal promotions. Best for: fashion before/after, quick feature highlights, seasonal offers.

E	Dynamic Product Ads (DPAs) Very, very good if you have a wide product catalog Pulls products directly from your catalog. Highly personalized — users see what they actually browsed. Best for: stores with many SKUs, personalized retargeting at scale.

4.4 · Creative Angles

The five angles — pick based on product and market. Mix multiple angles per campaign.

ANGLE 1

Problem / Solution

Pain points and fixes.

Best for cosmetics — before/after application to the skin.

ANGLE 2

Social Proof

Reviews, UGC, "people on the street."

Builds trust faster than any claim you make yourself.

ANGLE 3

Emotional / Lifestyle

Aspirational, identity-driven.

Works great for fashion.

ANGLE 4

Price / Offer-Driven

Discounts, bundles, urgency.

Perfect for seasonal sales (Black Friday) and direct-response campaigns.

ANGLE 5

Education / Value

Explain how it works — and which problem it indirectly solves.

Strong for technical / niche products.

● Mix angles —
never run only one.

🎯 Key Takeaway

Any format and any angle can work at any stage — the only thing that changes is the content of the creative. Stop building "TOFU ads" or "BOFU ads." Build creatives that hook, prove, and convert in a single watch. Test many. Let Meta find your buyers.

◆ Master Flow — Complete Pipeline

PHASE 1

Testing

Creative • Audience
Bidding • Catalog

➔ Winners

➔

PHASE 2

Optimize

Trim losers
Protect winners

➔ Lean machine

➔

PHASE 3