Sequence Performance
Email campaign/sequence performance review composite. Pulls campaign data (sends, opens, replies, bounces), reads actual email copy and subject lines, analyzes reply content (objections, positive interest, questions), and produces a diagnostic report covering quantitative metrics, copy quality, lead quality, and actionable recommendations. Tool-agnostic — works with Smartlead (MCP), Instantly, Outreach, Lemlist, Apollo, or CSV data.
by Athina AInpx gooseworks install --all # then, in Claude Code, Cursor, or Codex: /gooseworks use the sequence-performance skill
Sequence Performance
Goes beyond vanity metrics. Most campaign reports tell you open rate and reply rate. This skill reads the actual emails you sent, reads every reply you received, classifies the responses, evaluates your copy, evaluates your lead quality, and tells you specifically what's working, what's not, and what to do about it.
Three layers of analysis:
- Quantitative: The numbers — sends, opens, replies, bounces, conversions, by touch and by variant
- Qualitative (Copy): Are the subject lines, email bodies, CTAs, and personalization actually good?
- Qualitative (Replies): What are people actually saying? What objections keep coming up?
When to Use
Use this skill when:
- User says "how's my campaign doing", "sequence performance", "campaign review", "email analytics"
- User says "analyze my outreach", "why isn't my campaign working", "review my email results"
- A campaign has been running for 7+ days and has meaningful data
Phase 0: Intake
Outreach Tool
- What outreach tool do you use? (Smartlead / Instantly / Outreach.io / Lemlist / Apollo / Other)
- How do we access campaign data? (MCP tools / API / CSV export / paste metrics)
Campaign Selection
- Which campaign? (name or ID)
- Date range? (or "all data")
Your Company Context (for copy evaluation)
- What does your company do? (one-liner)
- Who is your ICP? (titles, industries, company size)
- What problem do you solve?
- What's your CTA goal? (book meeting, get reply, drive to page)
Benchmark Context
- Is this cold outreach or warm/nurture?
- What segment are you selling to? (SMB, mid-market, enterprise)
Step 1: Pull Campaign Data
Pull three categories of data from the user's outreach tool:
A) Campaign Metrics
| Data Point | What We Need |
|---|---|
| Total emails sent | By touch (Touch 1, Touch 2, Touch 3, etc.) |
| Total unique recipients | Deduplicated count |
| Opens | By touch, unique opens vs. total opens |
| Replies | By touch, total reply count |
| Bounces | Hard bounces + soft bounces |
| Unsubscribes | Count |
| Clicks | If link tracking is on |
| Positive replies | If categorized in the tool |
| Meetings booked | If tracked |
How to pull by tool:
| Tool | Method |
|---|---|
| Smartlead (MCP) | mcp__smartlead__get_campaign_stats, mcp__smartlead__get_campaign_sequence_analytics, mcp__smartlead__get_campaign_variant_statistics |
| Instantly / Outreach / Lemlist / Apollo | Ask user for CSV export or paste metrics |
| Other | User provides CSV with columns: email, status, opened, replied, bounced |
B) Email Copy (Sequence Content)
Pull the actual templates for every touch:
| Tool | Method |
|---|---|
| Smartlead (MCP) | mcp__smartlead__get_campaign_sequences |
| Others | User pastes the copy or provides CSV export |
C) Reply Content
Pull the actual text of every reply:
| Tool | Method |
|---|---|
| Smartlead (MCP) | mcp__smartlead__get_campaign_leads_history, mcp__smartlead__fetch_master_inbox_replies |
| Others | User provides reply dump or CSV export |
Human Checkpoint
Campaign: [name]
Status: [active/paused/completed]
Sent: X emails to Y recipients
Replies: Z (full text pulled for analysis)
Touches: N touches, M variants
Data looks complete? (Y/n)Step 2: Quantitative Analysis
Benchmarks
| Metric | Cold (SMB) | Cold (Mid-Market) | Cold (Enterprise) | Warm/Nurture |
|---|---|---|---|---|
| Open rate | 40-60% | 30-50% | 25-40% | 50-70% |
| Reply rate | 3-8% | 2-5% | 1-3% | 10-20% |
| Positive reply rate | 1-3% | 0.5-2% | 0.3-1% | 5-10% |
| Bounce rate | <3% | <3% | <2% | <1% |
| Unsubscribe rate | <1% | <1% | <0.5% | <0.5% |
Calculate
Overall metrics: open rate, reply rate, positive reply rate, bounce rate, unsubscribe rate, deliverability rate. Compare each to the benchmark.
Per-touch breakdown:
- Touch-level open/reply rates
- Marginal reply rate (replies from THIS touch / people who received this touch but hadn't replied yet)
- Touch contribution (what % of total replies came from each touch)
Variant analysis (if A/B testing):
- Open rate and reply rate per variant
- Statistical confidence: <50 sends = "insufficient data", 50-100 = "directional", 100-250 = "likely winner", 250+ = "statistically significant"
- Winner recommendation: scale, keep testing, or kill
Step 3: Reply Analysis
Read every reply, classify it, and extract patterns.
Reply Categories
| Category | Definition |
|---|---|
| Positive interest | Wants to learn more, open to a conversation |
| Meeting request | Explicitly asks to meet or provides availability |
| Warm / Curious | Interested but non-committal, asks questions |
| Objection — Timing | Not now, but potentially later |
| Objection — Budget | Can't afford or not a priority |
| Objection — Competitor | Already using a competing solution |
| Objection — Relevance | Doesn't see the fit |
| Objection — Authority | Not the right person |
| Not interested | Flat no |
| Auto-reply / OOO | Automated response |
| Referral | Redirects to someone else |
| Question | Asks about product/offering |
Objection Patterns
- Which objection appears most? (reveals systemic issues)
- Do objections cluster at Touch 1 (bad targeting) vs. Touch 3 (fatigue)?
- Which are handleable (timing, authority) vs. terminal (relevance)?
- What exact language do people use?
Positive Signal Patterns
- Which touch/variant generated positive replies?
- What do positive responders have in common? (title, industry, company size)
- What questions do warm leads ask? (reveals what's missing from the email)
Reply Quality Score
| Score | Criteria |
|---|---|
| Strong | >50% positive/warm. Objections are handleable. |
| Mixed | 30-50% positive. Mix of handleable and terminal. |
| Weak | <30% positive. Dominated by "not interested" and "not relevant." |
| Toxic | High unsubscribe + angry replies. Something is fundamentally wrong. |
Step 4: Copy Quality Assessment
Evaluate the actual email copy against best practices and reply data.
Subject Lines
| Criterion | Red Flags |
|---|---|
| Length | >60 chars gets truncated on mobile |
| Specificity | Generic "Quick question" or "Checking in" |
| Spam triggers | "Free", "Limited time", ALL CAPS |
| Open rate correlation | Low open rate = subject line problem |
Email Body
| Criterion | Red Flags |
|---|---|
| Hook (first line) | "I'm reaching out because..." or "We are a company that..." |
| Length | Over 150 words |
| Value prop clarity | Jargon, vague language, buzzwords |
| Proof points | No proof = no credibility |
| Personalization | Only {first_name} merge field |
| CTA | Multiple CTAs, high-friction asks, or no CTA |
| Filler language | "Hope this finds you well", "just checking in" |
| Sequence progression | Touch 2 is just a "bump" of Touch 1 |
Grades
Grade each touch A through F on: hook quality, value prop clarity, proof usage, personalization level, CTA quality.
Step 5: Lead Quality Assessment
Evaluate whether we're sending to the right people.
Targeting Check
- Do lead titles match ICP buyer/champion/user personas?
- Are leads in target industries?
- Right seniority level for the ask?
- Company size in target range?
Signal Quality (from replies)
| Pattern | What It Tells You |
|---|---|
| High "not relevant" replies | Sending to people who don't have the problem |
| High "wrong person" replies | Right companies, wrong roles |
| High "already have a solution" | Right problem, late to the party |
| High "timing" objections | Right people, right problem, wrong moment — not a targeting issue |
| Low reply + high open rate | People open but don't find it relevant — copy/targeting mismatch |
| High bounce rate | List quality issue — bad emails, old data |
Step 6: Generate Report
Report Structure
# Sequence Performance Review: [Campaign Name]
**Period:** [date range] | **Status:** [active/paused/completed]
---
## Executive Summary
**Overall verdict:** [One sentence]
| Dimension | Grade | Assessment |
|-----------|-------|-----------|
| Metrics | [A-F] | [one-liner] |
| Copy Quality | [A-F] | [one-liner] |
| Lead Quality | [A-F] | [one-liner] |
| Reply Quality | [Strong/Mixed/Weak/Toxic] | [one-liner] |
### What's Working (Double Down)
- [Specific thing with data]
### What's Not Working (Fix or Kill)
- [Specific thing with data]
### Top 3 Actions
1. [Highest-impact action]
2. [Second]
3. [Third]
---
## Detailed Metrics
### Overall Performance
| Metric | Actual | Benchmark | Status |
|--------|--------|-----------|--------|
| Open rate | X% | Y% | [above/below] |
| Reply rate | X% | Y% | [above/below] |
| Bounce rate | X% | <3% | [flag] |
| ... | ... | ... | ... |
### Performance by Touch
| Touch | Sent | Open Rate | Reply Rate | Marginal Reply Rate | % of Total Replies |
|-------|------|-----------|------------|--------------------|--------------------|
| 1 | X | Y% | Z% | Z% | W% |
### Variant Performance (if A/B testing)
| Touch | Variant | Subject | Sent | Open Rate | Reply Rate | Confidence | Action |
|-------|---------|---------|------|-----------|------------|------------|--------|
---
## Reply Deep Dive
### Reply Classification
| Category | Count | % of Replies |
|----------|-------|-------------|
### Top Objections
| Objection | Count | Handleable? | Suggested Response |
|-----------|-------|------------|-------------------|
### Notable Replies
[5-10 most instructive replies with quotes]
---
## Copy Assessment
[Subject line verdicts, body grades, sequence architecture assessment]
---
## Lead Quality
[Targeting assessment, actual vs intended ICP]
---
## Recommendations (Prioritized)
### High Priority (Do This Week)
1. **[Action]** — [data point] → [expected impact]
### Medium Priority (Do This Month)
2. **[Action]** — [data point] → [expected impact]
### Kill List
- [Anything that should be stopped]Recommendation Logic
| Finding | Recommendation |
|---|---|
| Open rate below benchmark | Subject line rewrite — suggest 3 alternatives |
| Reply rate below + open rate fine | Body copy issue — focus on hook, proof, CTA |
| Both below benchmark | Full sequence rewrite |
| High "not relevant" objections | Targeting issue — tighten ICP filters |
| High "wrong person" referrals | Title targeting issue — shift to referred titles |
| High "already have solution" | Add competitive differentiation to copy |
| High "timing" objections | Not a problem — set up 90-day re-engagement |
| One variant clearly winning | Scale winner, test new idea in losing slot |
| Touch 2/3 near-zero marginal replies | Cut sequence short or rewrite with new angles |
| High bounce rate | List hygiene — verify emails, check data source |
| Deliverability <95% | Infrastructure — check SPF/DKIM/DMARC, reduce volume |
Human Checkpoint
Present the executive summary, then ask:
Full detailed report available. Want to see the full breakdown, or act on a specific recommendation?Adapting to Data Availability
| Missing Data | What Gets Skipped | Still Useful? |
|---|---|---|
| Reply text | Reply classification + objection patterns | Partially — metrics + copy still run |
| Variant data | Variant analysis | Yes — single-variant analysis still runs |
| Lead demographics | Targeting assessment | Yes — infers from reply patterns |
| Open tracking | Open rate analysis | Partially — reply rate + copy still run |
Minimum viable data: Emails sent + reply count + email copy text.
Cost
Free. Pure reasoning + data from user's outreach tool.
Tips
- Run at Day 7 and Day 14. Day 7 catches deliverability and subject line problems. Day 14 gives enough replies for objection analysis.
- Reply analysis is where the gold is. Metrics tell you WHAT. Replies tell you WHY.
- High open + low reply = copy problem. The subject gets them to open but the email doesn't deliver.
- Low open + decent reply rate = subject line problem. The email works, people just aren't seeing it.
- "Not relevant" is the most important objection. If >20% say "this isn't for me," it's targeting, not copy.
- Don't kill a variant too early. Need 100+ sends per variant for directional data.
- Touch 2/3 should contribute 30-40% of replies. If Touch 1 is 90%+, your follow-ups aren't adding value.
What's included
Render VO Anchored Motion Listicle
Assemble an expert/educator motion-graphic LISTICLE video ad from a config — a spoken authoritative voiceover carries a numbered listicle while N web-animated hyperframe beats (HTML plus the Web Animations API, one branded design system of alternating tiles, big hero numerals, and glass-pill callouts) are rendered frame-by-frame via Playwright and anchored to the VO's word-level timestamps, periodic color-graded B-roll windows give visual breath, and captions burn ONLY inside those B-roll windows (2-word chunks, ASS Format header carrying a Name field so none drop) with the VO mixed under a low music bed. This is the FREE deterministic assembly stage (Playwright beat render plus ffmpeg concat plus window-masked caption burn plus VO-and-music mix plus final composite) — the VO, the music bed, and the stock B-roll come from create-vo-elevenlabs, create-music-elevenlabs, and media-proxy. Use for the vo-anchored-motion-listicle format.
Render Stopmotion Hand Swatch Cycle
Assemble a stop-motion hand-swatch-cycle product-demo ad from a config — a sequence of still PLATES (one hand swiping a single-barrel cosmetic across a cream skin-patch, the barrel + swatch changing per plate while the hand, background, crop, and lighting stay locked) is PNG→mp4 loop-encoded at each plate's own stop-motion hold (fast motion frames 150–250ms, per-shade ~380ms, hero beats 1100–1800ms), concat-demuxed with HARD cuts into a silent master, closed on a Playwright HTML-rendered branded end card (serif tagline + sans subtitle + real logo SVG over a hero BG, never AI-rendered text), and muxed with a pre-sourced music track playing under the end card with a fade tail (no VO). This is the FREE deterministic assembly stage (loop-encode + concat-demux + end-card render + music mux); the master-anchor plate, shade plates, and end-card BG come from create-image-gpt-image-fal and the track from create-music-elevenlabs. Use for the stopmotion-hand-swatch-cycle format.
Render Split Screen Creator
Assemble a split-screen creator ad from a config — a two-zone vertical composite where a supplied AI-creator lip-sync take fills the BOTTOM ~48% while real 16:9 product/demo clips run uncropped in the TOP ~52%, each top clip contain-fit with a darkened blurred cover-scale fill of the same clip (never black bars), a 3px brand-color divider between the zones, the creator slice cover-fit per the per-scene VO timing, scenes hard-concatenated with the body audio being the concatenated creator VO slices, an end card held on the last sharp frame ~3s, then the ASSEMBLED cut transcribed with local Whisper (not the raw VO — concat drops inter-scene silence) and word-level captions burned in the chosen style. This is the FREE deterministic assembly + caption stage (two-zone composite + blurred fill + divider + hard-concat + end card + captions); the VO comes from create-vo-elevenlabs, the anchor from create-image-gpt-image-fal, and the whole-VO lip-sync from a paid VEED Fabric 1.0 take (a no-atom upstream input). Use for the split-screen-creator format.