Required Credentials
| Credential | Where to Get | Used For | OAuth Scope |
|---|
| ----------- | ------------- | --------- | ------------- |
| META_ACCESS_TOKEN | Meta Developer Console → Graph API Explorer → Generate Token | Fetching campaign and form integration data | ads_read (read-only) |
| META_AD_ACCOUNT_ID | Ads Manager URL: adsmanager.facebook.com/act_XXXXXXXXX | Identifying which account to query | — |
When to Use
When running lead generation campaigns and standard e-commerce analysis logic doesn't apply. Specifically for:
- CPL is higher than expected or target
- Lead volume is healthy but conversion rate to closed deals is low (quality problem)
- LPV rate looks fine but form submit rate is unknown
- CAPI may not be sending offline lead data back to Meta
- Not sure if the problem is the ad, the audience, or the form itself
Lead gen has fundamentally different metrics than e-commerce — standard CVR and LPV benchmarks don't apply. Use this skill instead of applying e-commerce logic to lead gen campaigns.
Key Differences from E-commerce Analysis
| Metric | E-commerce Benchmark | Lead Gen Benchmark | Why It Differs |
|---|
| --- | --- | --- | --- |
| LPV Rate | > 70% is healthy | > 50% is healthy | Many leads browse without intent to submit |
| Form Submit Rate | N/A | > 20% is healthy | Form friction is the primary drop-off point |
| CPL | Varies by industry | The primary metric | What you're actually paying for |
| CVR (post-click) | Purchase rate | Lead-to-qualified rate | Meta can't see what happens after form submit |
| Attribution | 1-day or 7-day click | 7-day click usually | Longer window for consideration |
Step 1: Gather Lead Gen Specific Data
Pull these metrics at the campaign and adset level:
- Spend, impressions, CTR
- LPV Rate (Landing Page View rate = landing page views / link clicks)
- Form Submit Rate (if available — requires CAPI or form integration)
- CPL (cost per lead)
- Frequency
- Lead volume vs. prior period
Also ask the user:
- What does a "good lead" look like for their business?
- What's their lead-to-close rate historically?
- Are they using CAPI to send offline lead data back to Meta?
Step 2: CAPI Verification — Critical for Lead Gen
Without CAPI sending offline lead data back to Meta:
- Meta is optimizing for form submissions, not qualified leads
- CPL shown may be "form submit cost" not "actual lead cost"
- The algorithm can't learn which leads actually convert
How to verify CAPI is connected:
- Go to Meta Events Manager → select your pixel
- Check "About CAPI" — status should be green/active
- Ask: are offline leads (calls, CRM-qualified leads) being sent back via CAPI?
If CAPI is NOT connected for offline leads:
- This is a priority fix — it will lower CPL over 2–3 weeks
- Recommend a CRM integration or Zapier/Make workflow
- Or use Meta's Lead Forms (form submission = conversion automatically tracked)
Step 3: Audience Targeting Analysis for Lead Gen
Cold Audiences (Interest-Based Targeting)
- Interest targeting works for cold audiences
- Frequency builds fast — limit budget to avoid rapid fatigue
- Test narrower interest layers (2–3 stacked interests, not broad categories)
Custom Audiences (Retargeting and Lookalikes)
- Website visitors (pixel): typically highest conversion rate
- Email lists (matched audiences): strong for existing customer re-engagement
- Lookalikes (LAL): based on converter audiences, quality depends on seed list quality
- Test LAL layers 1–3% first — wider LAL = lower quality but more volume
Step 4: Format Output
Lead Gen Health Summary:
- Spend / CPL / Lead Volume
- LPV Rate vs. benchmark (> 50% healthy for lead gen)
- Form Submit Rate (if available)
- CAPI Status (connected / not connected)
SECTION 1: CAPI Gap (if applicable)
State whether offline lead data is being sent to Meta. Impact: algorithm is optimizing for quantity, not quality.
SECTION 2: Audience Diagnosis
Diagnose targeting based on frequency and CPL — is the audience too broad or well-matched?
SECTION 3: Funnel Stage
Identify the CPL bottleneck stage:
- LPV < 50% → Ad-to-Form disconnect
- LPV ≥ 50% but low leads → Form friction (check form fields)
- LPV and CVR OK but CPL high → Audience mismatch or CAPI not connected
> LP Landing Page diagnosis: The full LP disconnect diagnostic (Step 4a–4d) lives in meta-ads-recommendation → Step 4. This skill identifies the funnel stage; recommendation prescribes the fix.
Session Context — What This Skill Writes
After completing analysis, store the following in session context:
| Key | Description | Example |
|---|
| ----- | ------------- | --------- |
| lp_diagnosis | Funnel stage causing the CPL problem | "Form friction — submit rate 12%, below 20% benchmark" |
| capi_status | Whether offline leads are being sent to Meta | "Not connected — offline lead data not flowing" |
| cpl_breakdown | Which stage is the bottleneck | "LPV 65% OK; Form submit 11% is the bottleneck" |
| recommended_fix_priority | Ranked fix order | "1. CAPI 2. Form fields 3. Audience" |
> meta-ads-recommendation reads these keys to produce the lead gen action plan.
Key Collision Resolution
> ⚠️ Key priority: meta-ads-analysis writes lp_diagnosis_general; this skill writes lp_diagnosis (the lead-gen-specific version). meta-ads-recommendation is configured to prefer lp_diagnosis when available — so running both skills in sequence does NOT cause overwriting. The two keys coexist.
Rules
- Apply lead gen benchmarks, not e-commerce benchmarks
- Always verify CAPI status before diagnosing CPL issues — missing CAPI is the most common cause of misleading CPL data
- Do not recommend creative changes if form friction is the bottleneck
- Do not recommend audience changes if LPV rate indicates a page problem
- This is analysis only — recommendations route to meta-ads-recommendation