Agent Skills: Outbound Optimizer

Outbound Optimizer

Diagnose current outbound performance, identify root causes, apply the outbound-sequences skill for improvements, validate output, and recommend A/B tests.

Step 1: Gather Metrics

Collect current performance data:

1. Metrics: Open rate? Reply rate? Meeting book rate?
2. Volume: How many sequences/week? Total contacts?
3. Target: Who are you reaching? (Title, industry, company size)
4. Channels: Email only? Multi-channel? (Email, phone, LinkedIn)
5. Sample: Best-performing email or script
6. Challenge: Low opens? Low replies? No-shows? Wrong prospects?

Step 2: Diagnose Against Benchmarks

| Metric | Benchmark | If Below → Problem | |--------|-----------|-------------------| | Open rate | 40-60% | Subject line / deliverability | | Reply rate | 5-15% | Copy / relevance | | Positive reply rate | 2-5% | Targeting / offer | | Meeting book rate | 1-3% | CTA / friction | | Show rate | 70-80% | Confirmation / timing | | Connect rate (calls) | >5% | Timing / list quality |

Step 3: Identify Root Cause

| Symptom | Likely Cause | Investigation | |---------|--------------|---------------| | Opens low, replies low | Subject line problem | Test new subject patterns | | Opens high, replies low | Copy doesn't resonate | Review first line, value prop | | Replies high, meetings low | CTA too aggressive | Lower friction ask | | Meetings high, shows low | Weak confirmation | Add reminder sequence | | All metrics low | Wrong ICP | Review targeting criteria |

Step 4: Apply Outbound Sequences Skill

Apply outbound-sequences with structured context:

context:
  current_metrics: [open rate, reply rate, meeting rate]
  diagnosed_problem: [subject lines | copy | targeting | CTA]
  target_persona: [title, industry, company size]
  channels: [email | multi-channel]
  value_prop: [what you solve]
  social_proof: [notable customers, results]
  current_best_performer: [paste sample]
  tool: [Apollo, Outreach, Lemlist, etc.]
request:
  - New sequence addressing diagnosed problem
  - Include A/B test variants for problem area

Step 5: Validate Output

Quality Checklist

• [ ] Subject lines are personalized and specific (not generic)
• [ ] First line shows research (references something specific)
• [ ] Emails are under 100 words
• [ ] CTA is clear and low-friction
• [ ] Sequence has 5-7 touches across channels
• [ ] Follow-up emails add new value (not just "checking in")
• [ ] Response handling covers all scenarios
• [ ] Templates are ready to load into tool

Red Flags

| Issue | Problem | Fix | |-------|---------|-----| | Generic opener | "Hope you're well" | Specific observation | | Long emails | Won't be read | Under 75 words | | Multiple CTAs | Confusion | Single clear ask | | No personalization vars | Can't scale | Add {{variables}} | | Same value in each email | No reason to reply | New angle per email |

Step 6: Recommend A/B Tests

Based on diagnosed problem, recommend:

| Problem Area | Test | Measure | |--------------|------|---------| | Subject lines | 2-3 variations | Open rate | | First line | Personalized vs. direct | Reply rate | | CTA | Meeting vs. question | Response rate | | Send time | Morning vs. afternoon | Open rate | | Sequence length | 5 vs. 7 touches | Total reply rate |

Metrics Tracking Template

Weekly Review:
Sequences Sent: [X]
Open Rate: [X%] (benchmark: 50%)
Reply Rate: [X%] (benchmark: 10%)
Positive Rate: [X%] (benchmark: 3%)
Meetings Booked: [X]
Pipeline Generated: $[X]

Top Performer: [sequence/template name]
Underperformer: [sequence/template name]

This Week's Test: [what we're testing]
Result: [outcome]

Next Week Actions:
1. [specific optimization]
2. [specific optimization]

Handling Edge Cases

| Situation | Action | |-----------|--------| | No metrics data | Ask for approximate open rate (estimate is fine) | | No sample copy | Ask for current best email | | Multiple problems | Focus on earliest funnel stage first (opens before replies) | | Output too generic | Re-apply skill with more specific persona details | | Tool constraints | Adjust templates for tool limitations |

Deliverables

1. Diagnosis of current performance
2. Root cause analysis
3. Optimized sequence from outbound-sequences skill
4. A/B testing plan
5. Metrics tracking template
6. Weekly review cadence