Meta Ads Management for B2B

Orchestrator for all Meta Ads (Facebook/Instagram) tasks. Meta Ads for B2B can deliver half the cost per lead and lower cost per qualified opportunity compared to LinkedIn - but only when audience data quality is right and creative does the targeting work.

Methodology

This skill implements the Ivan Falco B2B demand generation methodology for Meta - a data-first, creative-as-targeting approach that leverages Meta's algorithm (Andromeda + Gem) with high-quality audience inputs and systematic creative testing.

Core Philosophy

On Meta, your data is everything. Your creative is your targeting.

Meta's native targeting can't match LinkedIn's B2B precision (no job title, company, seniority filters). The way to make Meta work for B2B is to bring your own high-quality audience data (CRM, third-party providers) and use creative specificity to filter for ICP. The algorithm (Andromeda + Gem) does the rest.

Routing Logic

Always Load First

Any Meta work starts here:

Intent	File	Priority
Scaling qualified B2B pipeline end to end (the goal is SQLs / qualified pipeline, not lead volume)	scale-b2b-qualified-pipeline.md	LOAD FIRST when the goal is qualified pipeline. The 6-step spine (map market -> audiences -> funnel events/CAPI -> creative -> segments -> SQL reporting) that ties the audience, CAPI, creative, and reporting files together.
ANY operational decision (pause, scale, graduate, budget, creative count)	meta-ads-operating-system.md	ALWAYS LOAD FIRST. This is THE decision framework. All formulas, thresholds, and actions live here.
Auditing an account (pull data, classify ads, produce recommendations)	Load the OS file above + run scripts/get_active_ads_copy.py	Pull active ads, classify by OS rules, produce recommendations.
Creative production decisions (what to build, when to iterate, cadence, formats)	creative-cadence-operating-system.md	Iteration hierarchy, concept sourcing, format playbook, testing cadence, fatigue detection, quality scoring.

Deeper Context (Load When Needed)

These files provide detailed methodology on specific topics. The OS drives decisions; these explain the deeper why and how.

User Intent	Knowledge Base File	When to Use
Full Meta B2B overview, algorithm, strategy	meta-b2b-overview.md	Andromeda + Gem, strategy, "what works" - overview and quick reference
Pixel, tracking, first campaign setup	meta-setup-and-tracking.md	Installing the pixel, Events Manager, domain verification, CAPI, pre-launch
CAPI, conversion events, HubSpot vs n8n	meta-capi-and-events.md	Conversions API, event hierarchy, CRM to CAPI, deduplication, Event Match Quality
Webinar/event, third-party conversion	meta-third-party-conversion-tracking.md	Off-domain conversion (Luma, etc.), pixel in platform vs thank-you page
Audience targeting, data strategy, lookalikes	audience-strategy.md	Building audiences, data sources, audience validation, third-party tools
Detailed campaign structure, phases, roadmap	campaign-structure.md	Phase 1/2/3 deep dive, naming conventions, campaign architecture, month-by-month roadmap
Creative concepts, copy, formats	creative-strategy.md	Creative development, concept testing, placement optimization, creative-as-targeting
Creative fatigue detection, rotation system	creative-fatigue-detection.md	Full fatigue workflow, rotation cadence, pipeline management, format-specific notes
Advantage+ setup and details	advantage-plus.md	When to use Advantage+, setup steps, budget requirements, ABM considerations
Optimization playbook, benchmarks	optimization-playbook.md	Decision trees, B2B benchmarks, seasonal patterns, weekly cadence, scaling protocol
Ad quality scoring, message validation	message-validation.md	Scoring ads against revenue quality, winner scaling pattern, validation process
Lead forms, conversion optimization	lead-form-optimization.md	Lead form setup, work email validation, custom questions, social amnesia
ABM on Meta	abm-on-meta.md	ABM targeting, Advantage+ conflicts, manual vs hybrid approach
Offer strategy by funnel stage	offer-strategy.md	What offers to run at each funnel stage
Creating campaigns, ads, audiences (the full build chain)	creating-campaigns-ads-audiences.md	End-to-end: audience -> campaign -> ad set -> creative -> ad. Endpoints, scripts, PAUSED-by-default rules.
Reporting / dashboards	use the meta-reporting skill	Performance analysis + a branded HTML dashboard with your logo.

Scripts (execution layer)

Python scripts live in scripts/ and hit the Meta Marketing API (v22.0) through the shared scripts/client.py (credentials from .env). Full index, key args, and examples: scripts/README.md. Every create defaults to PAUSED. Budgets are in cents. Any script accepts --account-id act_XXXXXXXXX to override the default account.

Analyze / Report

Script	Use when you need to...
account_overview.py	Pull an account-level metrics dashboard (with optional period comparison)
list_campaigns.py / list_ad_sets.py / list_ads.py	List objects with status, budget, targeting, and 30-day metrics
get_campaign_performance.py	Get detailed campaign analytics with date presets / daily breakdown
get_active_ads_copy.py	Dump full creative and copy of every active ad (audits)
top_ads.py	Rank ads by spend or cost per lead to find winners

Create (default PAUSED)

Script	Use when you need to...
create_campaign.py	Create a campaign
create_ad_set.py	Create an ad set (targeting, optimization goal, pixel/page promoted object)
create_ad.py	Create a single-image or video ad (creative + ad)
create_dco_ad.py	Create a dynamic/flexible creative (DCO) ad from multiple assets

Manage

Script	Use when you need to...
update_campaign.py	Change a campaign's status, budget, or name
duplicate_ad.py	Duplicate an ad (optionally into another ad set, N copies)
ad_scheduler.py	Schedule automatic ad pauses (e.g. webinar ads), run from cron

Audiences

Script	Use when you need to...
create_custom_audience.py	Create a custom audience from a hashed CSV
update_custom_audience.py	Add or remove hashed members in an existing audience
upload_lookalike.py	Create lookalike audiences from a source audience (one per country)

Core Rules

1. Speak from this methodology directly and with conviction - it's built from managing real B2B spend, so present it as the operator playbook it is. Don't pad with generic "best practices."

2. Data quality determines Meta success for B2B. Always follow the data hierarchy: CRM lookalikes (Tier 1) -> Third-party data (Tier 2) -> Broad targeting (Tier 3). Never skip to broad without testing data-driven audiences first.

3. Validate audience quality before scaling creative. Use ABO (ad set budget) campaigns with one ad set per audience source and the same ads across all. Audit lead quality by job title and company. Only scale winning audiences with CBO.

4. Creative concept testing, not micro-variations. Test dramatically different concepts (UGC vs before/after vs meme vs product demo), not minor tweaks (blue vs green button). The algorithm needs creative diversity to learn.

5. Creative IS targeting on Meta. Your ad copy must explicitly call out who the ad is for. When running broader audiences, the creative does the filtering work - it should act as "mosquito repellent" for non-ICP.

6. Lead forms need intentional friction for B2B. Always require work email validation. Add 1-3 custom qualification questions. Use Higher Intent form type. This combats social amnesia and improves lead quality.

7. Build in this order: Remarketing -> Prospecting -> Acceleration. Start with remarketing (lowest risk, highest ROI, creates omnipresence). Layer prospecting once remarketing is running. Add acceleration for open pipeline if applicable.

8. Cross-channel remarketing is a superpower. Create Meta retargeting audiences from LinkedIn/Google traffic using UTM parameters. You retarget validated audiences on Meta at a fraction of the cost.

9. Broad targeting only works for large TAM. If you're doing ABM or targeting niche enterprise, broad will not work - stick to Tier 1 and Tier 2 data sources. Broad is for SMB/lower mid-market with massive addressable markets.

10. Meta's algorithm (Andromeda + Gem) needs creative volume. Maintain 4-6 unique creative concepts per campaign. The algorithm processes ads to learn what works - more diverse inputs = better optimization.

Key Differences from LinkedIn

Aspect	Meta	LinkedIn
Native B2B targeting	Weak - bring your own data	Strong - job title, company, seniority
Cost per lead	~50% of LinkedIn	Higher but more precise
Algorithm role	Heavy - drives targeting through creative signals	Light - advertiser controls targeting
Creative volume needed	High (4-6+ unique concepts)	Moderate (4-6 per campaign)
Remarketing strength	Excellent - cheap, cross-platform omnipresence	Good but more expensive
Lead form quality	Requires friction management (social amnesia risk)	Higher quality by default (work email auto-fill)
Best for	Large TAM, SMB/mid-market, remarketing, pipeline acceleration	Enterprise, ABM, precise ICP targeting

Output Standards

• Campaign plans -> Use docx skill for professional Word documents
• Performance reports -> Use xlsx skill for data-driven spreadsheets
• Creative briefs -> Structured concept descriptions with placement specs
• All documents -> Professional, client-ready quality. No source citations.

When to Combine with LinkedIn

Meta and LinkedIn are not competitors - they're complementary channels:

• LinkedIn for precision prospecting (right person, right company) + Meta for cheap remarketing (stay top of mind across platforms)
• LinkedIn for ABM (company-level targeting) + Meta for ABM extension (via third-party tools like Primer/Metadata.io)
• LinkedIn for enterprise (where precision justifies the cost) + Meta for mid-market/SMB (where volume and cost efficiency matter more)
• Use cross-channel UTM remarketing to bridge the two platforms