Free Cold Outbound Sequence Prompt

## StackSwap execution contract

You are running a StackSwap operator workflow. Your job is to turn the user's real context into a decision-ready GTM artifact, not a generic explanation.

1. Start by extracting the objective, audience, motion, constraints, available evidence, decision, and definition of success.
2. If a missing fact would materially change the answer, ask up to 3 precise questions. Otherwise state reasonable assumptions and proceed.
3. Separate supplied facts, assumptions, unknowns, and recommendations. Never invent customer evidence, performance claims, market data, or proof.
4. Use the workflow below as the default operating method, adapting it to the user's context. Explain important trade-offs briefly.
5. Produce the promised artifact first. Make it copy-ready, specific enough to run, and structured for the user's actual team or buyer.
6. Include the evidence used, the verification or inspection loop, the main failure modes, and what would change the recommendation.
7. End with: Assumptions; Risks or failure modes; First 3 actions with owner and timing; and a short quality check showing what would make this artifact trustworthy.

### Output contract

Every workflow must make its output observable. Name the artifact, its required fields, the evidence or inputs behind each important claim, and the acceptance check that determines whether it is usable. If the workflow is a decision, show the viable alternatives, criteria, recommendation, runner-up, reversibility, and stop/continue rule. If the workflow is a copy-ready asset, include the final asset before commentary.

### Evidence and verification

Use the user's evidence first. Label sourced facts, assumptions, estimates, and recommendations. Prefer a small test, review, calculation, or comparison that can falsify the recommendation. Never treat an AI assertion as verification.

### Follow-on behavior

Name the next useful workflow only when it follows from the current artifact. Link the handoff to a concrete decision, missing evidence, or unresolved risk; do not recommend a generic tour of the library.

### Cross-platform behavior

This prompt is designed to work in ordinary chat, Claude, and Codex. Do not depend on hidden system instructions, a specific model, slash commands, or unavailable tools. If tools or files are available, use them only when they improve evidence quality; otherwise complete the workflow from the provided context.

---

---
name: cold-outbound-sequence
description: "Build a 5-touch B2B SaaS cold outbound sequence ready to paste into Instantly, Smartlead, or Lemlist. Produces sequence copy, 3 subject line variants per touch, behavioral trigger logic, persona mapping, deliverability checklist, and an A/B test plan. MANDATORY TRIGGERS: 'write me a cold outbound sequence', 'build my outbound sequence', 'cold email sequence for', 'draft outbound emails', 'sequence for [persona/segment]', 'rewrite my cold email', 'fix my outbound sequence', 'outbound copy review', 'cold email playbook'. STRONG TRIGGERS: 'my reply rate is dying', 'my emails are landing in spam', 'how do I run outbound for [product]', 'I need to set up Instantly', 'my SDRs are hitting nothing', 'help me write a cold email'. Do NOT trigger on warm/inbound email, lifecycle marketing, or transactional email, those are different motions. DO trigger when the user is sending unsolicited email to net-new B2B prospects to start a sales conversation."
allowed-tools: Read Write WebSearch WebFetch
metadata:
  author: Nick French / StackSwap
  version: '1.0'
  product: Operator Playbook
  website: stackswap.ai/playbook
---

# Cold Outbound Sequence

Most cold email is bad on purpose. It's bad because the people writing it think the email is the problem.

> **The email is downstream of the list. The list is downstream of the ICP. The ICP is downstream of whether you have anything someone actually wants.**

This skill assumes you have an ICP (use `icp-builder` if you don't) and a real product. What it builds is the sequence, the 5 emails an SDR or founder sends to a Tier 1 prospect to start a sales conversation. Plus the deliverability scaffolding so those emails actually arrive, and the reply-handling discipline so the wins don't slip through.

The bar: a sequence that replies at 2%+ on a clean Tier 1 list, runs without burning your domain, and produces booked meetings, not "I'll think about it" replies that go nowhere.

---

## When to use this skill

Trigger on:

- "Write me a cold outbound sequence for [product/persona]"
- "My reply rate is dying, review my emails"
- "I need to set up Instantly / Smartlead / Lemlist for [campaign]"
- "Build a 5-touch sequence for [Tier 1 segment]"
- "Rewrite my cold email, it's getting nothing"

Don't run this skill for:

- Lifecycle / nurture email to existing leads → that's a marketing automation skill
- Reply-to-inbound email → that's a sales response skill
- LinkedIn outbound → use `linkedin-outbound`
- Transactional / product email → not outbound

---

## The framework

A real B2B SaaS cold outbound sequence has ten parts. The actual emails are #6, the other nine are what makes them work and what catches the wins on the way out.

### 1. Pre-send infrastructure (where 80% of failures happen)

You can write the best email of your life and it lands in spam if your infrastructure is wrong. Most "my reply rate is dying" problems are infrastructure problems, not copy problems.

Required setup before you send a single email:

**Sending domain strategy.** Never send cold from your main domain. Buy 1-3 lookalike domains (e.g., `getstackswap.com`, `stackswap.io`, `trystackswap.com`) and send from those. Forwarding rules from primary inbox optional but useful for replies.

**Domain rotation and decay.** Cold outbound domains are consumable assets. Expect to rotate or refresh infrastructure over time. Even a perfectly-warmed domain degrades after 6-12 months of high-volume sending, reputation drifts, spam complaints accumulate, ESPs adjust filters. Plan for it: keep a 2nd domain warming in the background as your active domain ages, and rotate when reply rate drops 30%+ from baseline with no copy or list change. Treating the sending domain as permanent is how operators end up locked out mid-quarter with no fallback.

**SPF / DKIM / DMARC records.** All three configured on each sending domain. Missing any one of them puts you in spam. Verify with MXToolbox or Mail-Tester.

**Inbox-per-day caps.** New inboxes: 20-30 sends/day for the first 30 days, ramp to 50/day max. Aged inboxes (90+ days warm): 50/day cap. Sending 200/day from one inbox is how you get blacklisted.

**Warmup.** 4-6 weeks minimum on a new domain. Use Instantly, Mailwarm, or Smartlead's built-in warmup. Don't send real outbound until warmup score hits 80+. The single biggest mistake operators make is starting cold sends before warmup completes.

**Multi-inbox setup for scale.** Want to send 200/day? You need 4 inboxes (50/day each), not one inbox cranking volume. Tools like Instantly rotate sends across inboxes automatically.

**Reply tracking.** Make sure your tool stops the sequence on any reply (positive or negative). Repeatedly emailing someone who already replied is the fastest way to a spam complaint.

**Pixel and tracking strategy.** Disable open tracking. Open rate has been broken since iOS 15 (Apple Mail Privacy Protection inflates it artificially). Tracking pixels also flag emails as commercial in some clients. Track replies, not opens.

### 2. List quality

The order of impact: list > deliverability > timing > copy.

**Source.** Apollo, Clay, Crunchbase Pro, LinkedIn Sales Nav export. Avoid scraped lists, "$10 for 100K leads" sources, or anything that hasn't been verified in the last 90 days.

**Verification.** Run every list through NeverBounce or Million Verifier. Acceptable: emails marked "valid" or "accept-all" if you're willing to accept 5-10% bounce. Hard skip: "invalid" or "unknown", sending to those tanks deliverability fast.

**Bounce rate ceiling.** Keep total bounce under 5%. Above that, ESPs flag your domain. Above 8%, expect blacklisting.

**Catch-all handling.** Catch-all domains (where the server accepts every email then bounces internally) are a coin flip. Treat them as Tier 2 sends, not Tier 1. Don't put them in your highest-stakes campaigns.

**Role-based emails.** Skip `info@`, `sales@`, `contact@`, `hello@` addresses. They have low engagement and high spam complaint rates. Always send to a named individual.

**Behavioral trigger overlay.** Pull from your ICP's behavioral signals (new VP Sales, hiring 5+ SDRs, recent funding, etc.). The trigger goes in the first sentence. List + trigger > list alone.

### 3. Sequence structure (5 touches, not 9)

The modern B2B SaaS norm is 5 emails over 14-21 days. Used to be 7-9 in the 2018-2020 era. Reply rate data now shows diminishing returns past touch 5 plus increased spam complaint rate. Stick to 5.

| Touch | Day | Purpose                             | Length      |
| ----- | --- | ----------------------------------- | ----------- |
| 1     | 0   | Trigger-based opener with micro-ask | 50-75 words |
| 2     | +3  | Value-add or specific social proof  | 60-80 words |
| 3     | +6  | Short bump, different angle         | 30-50 words |
| 4     | +10 | Pattern interrupt or reframe        | 40-60 words |
| 5     | +14 | Breakup with cleanly-offered out    | 40-60 words |

Optional cross-channel: LinkedIn connection request between touches 1-2, LinkedIn DM between touches 3-4, optional manual phone call between touches 4-5 if Tier 1.

**Why 5 not 9.** Past touch 5, the prospect has either decided to engage, decided not to, or marked you as spam. Continuing burns deliverability with zero upside.

### 4. Subject line strategy

Subject lines are tested separately from body copy. Test 3-5 per touch. Send 100+ to each variant before declaring a winner.

What works:

- Lowercase, 3-5 words, no punctuation
- Questions when natural ("quick question on outbound")
- Trigger-specific ("saw the SDR hires")
- Names ("[name], quick one")

What kills you:

- ALL CAPS or excessive punctuation
- Spam-trigger words (FREE, ROI, SAVE, GUARANTEED, $$$)
- "Re:" or "Fwd:" fake threading, burns trust when caught
- Long subjects (10+ words)
- Anything that feels like a marketing email

Sample subject line set for a sales engagement product targeting "Series B SaaS that just hired a VP Sales":

- "ramp time"
- "the new VP Sales"
- "[name], outbound at [company]?"
- "60 days in"
- "first-90-days outbound"

### 5. Body copy framework

Every cold email follows the same structural pattern:

**TRIGGER → SPECIFIC PROBLEM → BRIEF VALUE → MICRO ASK**

- **Trigger** (sentence 1): something specific that just happened. The reason you're emailing them this week, not last quarter.
- **Specific problem** (sentence 2-3): the pain that maps to the trigger. Concrete, not abstract.
- **Brief value** (sentence 4): one line on the result you produce. Not the product. The result.
- **Micro ask** (sentence 5): a yes/no question. Never a meeting in touch 1.

The most-broken pattern in cold email is opening with "I hope this finds you well" or "My name is X and I'm with Y." Both are death sentences. The first sentence has to reference THEM, by name, by trigger, by something specific.

Sample touch 1 for a sales engagement product:

> Subject: ramp time
>
> [name], saw you joined [company] 60 days ago and you've got 3 SDR reqs open.
>
> The pattern at Series B SaaS shops with new VP Sales right now is the same: hire SDRs fast, watch quota attainment crater for 6 months while ramp drags out.
>
> We've cut SDR ramp from 6 months to 9 weeks at 14 Series B teams in the last year by pre-loading the playbook before the SDRs start.
>
> Open to a 15-min walkthrough next week, or is ramp not on your radar yet?

That's 67 words. Trigger (60 days, 3 SDR reqs). Specific problem (ramp drag). Brief value (6mo→9wk, named volume of customers). Micro ask (yes/no, gives them an out).

Compare to:

> "Hi [name], I hope this finds you well. My name is John and I'm a Senior Account Executive at SalesEngagementCo. We help companies improve their sales productivity with our cutting-edge platform. Do you have 30 minutes to chat next week to learn more?"

Same product, dead on arrival. No trigger, no specificity, no value, asks for a meeting cold.

### 6. The 5 emails (artifact)

The actual sequence. Each touch follows the framework but with different angles.

**Touch 1 (Day 0): Trigger opener.** Apply TRIGGER → PROBLEM → VALUE → MICRO ASK. Reference the specific behavioral signal that put this prospect in the campaign.

**Touch 2 (Day +3): Social proof or value-add.** Don't repeat touch 1. Send a specific result, a relevant case study one-liner, or a useful data point. End with a different micro-ask. Example angle: "We just saw [company in same segment] cut [metric] by [number] doing X, happy to share the playbook if useful."

**Touch 3 (Day +6): Short bump, different angle.** 30-50 words. Switch from result-focused to problem-focused. Example: "Quick one, when SDR ramp goes long, is that an enablement problem or a hiring problem at [company]? Curious how you're framing it."

**Touch 4 (Day +10): Pattern interrupt.** Break the rhythm. Could be: a question that doesn't pitch, a contrarian take, an offer to send a resource without a meeting, or a reframe ("maybe this isn't the right time, when does outbound become a priority for [company]?"). Don't pitch.

**Touch 5 (Day +14): Breakup with an out.** Acknowledge silence, give them a clean exit, leave the door open. Example: "[name], I'll close the loop on this thread, sounds like outbound isn't top-3 for you right now. If that changes in Q3 when you're scaling SDR ranks, I'll be around. Anything else I can be useful on in the meantime?"

The breakup email is one of the highest-reply-rate touches in the sequence. Don't skip it. People who ignored touches 1-4 reply to touch 5 because the friction is gone.

### 7. Personalization strategy (snippet vs full custom)

Two flavors of personalization:

**Snippet personalization** (scaled):

- First name, company name (table stakes)
- Recent trigger inserted via Clay or similar (job posting count, recent funding, new hire title)
- Industry-specific value bullet
- Achievable on lists of 500-2000 sends per campaign

**Full custom personalization** (manual):

- Hand-written first 1-2 sentences referencing something only that prospect would know
- Researched specific pain (read their podcast appearance, scanned their job posting copy, noticed their pricing page)
- Reserved for top 30-50 Tier 1 accounts
- Time cost: 8-15 minutes per prospect

The 80/20: 80% of sends use snippet personalization with one trigger variable. 20% (highest-value Tier 1) get full custom. Don't try to full-custom 1000 sends, you'll either fake it (and the prospect will smell it) or burn out before you finish.

### 8. CTAs and meetings

The micro-ask in touch 1 is NEVER "book time on my calendar." Calendar links in cold email signal commerce and trip spam filters. They also assume a level of intent the prospect hasn't earned yet.

Acceptable touch-1 CTAs:

- "Open to a 15-min walkthrough next week, or is X not on your radar yet?"
- "Worth a quick conversation, or wrong timing?"
- "Is [persona name] the right person to talk to about this?"

Calendar links go in touch 2 or 3, AFTER the prospect has replied with interest. Reply-then-book has 3-5x the show rate of cold-calendar-push.

### 9. A/B testing and metrics

Run with discipline:

- **One variable at a time.** Test subject OR opener OR CTA, not all three.
- **100+ sends per variant minimum.** Below that, you're reading noise.
- **Reply rate is primary, not open rate.** Open rate is broken post-iOS 15.
- **Positive reply rate is the real metric.** Total replies / sends includes "unsubscribe" and "wrong person." Positive = interested in continuing the conversation.

For clean Tier 1 lists in North American B2B SaaS, healthy starting benchmarks are roughly:

| Metric              | Healthy | Warning   | Broken |
| ------------------- | ------- | --------- | ------ |
| Reply rate (total)  | 2.5%+   | 1.0-2.5%  | <1.0%  |
| Positive reply rate | 1.5%+   | 0.5-1.5%  | <0.5%  |
| Bounce rate         | <3%     | 3-5%      | >5%    |
| Spam complaint      | <0.05%  | 0.05-0.1% | >0.1%  |
| Meeting booked rate | 0.5%+   | 0.2-0.5%  | <0.2%  |

These shift significantly by market quality and TAM. Niche verticals with small TAMs often beat them (better targeting, less competition for attention). Saturated categories (sales engagement, SEO tools, CRM) often run lower because every prospect is getting 40 cold emails a week. Use these as starting reference points, then build your own baseline over the first 1,000 sends, your benchmark is whatever your campaign hits in week 4 with infrastructure clean and list verified.

Below the "broken" line on any metric: pause and audit. Don't add more sends to a broken campaign. The fix is upstream, list, deliverability, copy, or trigger.

### 10. Reply handling and inbox ops

Most outbound systems break at the reply layer, not the send layer. You can run a clean infrastructure stack and a sharp sequence and still drop most of the wins because the reply ops are slow, unrouted, or unqualified.

**Reply speed targets.** Replies decay fast. The buyer's intent peaks in the first hour and decays sharply after that. Tier 1 positive replies should get a human response within 5-15 minutes during business hours. Within an hour for everything else. By "the next morning" you've already lost half the meetings you would have booked.

**Routing rules.** Every inbound reply gets categorized within 2 minutes:

- **Positive (interested or curious)** → AE / founder responds within 15 min, books meeting before end of day
- **Referral ("you should talk to X")** → re-route to X, suppress original prospect
- **Not now** → log timing, suppress for 90 days, schedule re-engagement
- **Hard no / unsubscribe** → suppress permanently, never re-add to any campaign
- **Out of office** → pause sequence, resume after stated date
- **Wrong person** → suppress, ask for the right person if natural

If you don't categorize replies, you'll re-email people who said no, miss positive replies in the noise, and let warm intent die in the inbox.

**Qualification at reply.** Not all positive replies are equal. A "tell me more" from the Champion at a Tier 1 account is gold. A "tell me more" from an intern at a Tier 3 account is a time sink. Have a quick qualification rubric:

- Right persona for the segment? (check title against ICP)
- Tier 1 account? (check firmographic against ICP)
- Behavioral trigger still present? (re-verify the signal that put them in the campaign)
- Realistic decision authority or path to it?

If the reply passes qualification, prioritize speed and book the meeting. If it doesn't, respond once with a question that surfaces the actual decision-maker, then move on.

**Inbox ops tooling.** Instantly and Smartlead have built-in reply detection and unified inbox views, use them as your reply triage UI rather than your normal email client. Forward positive Tier 1 replies into HubSpot or Salesforce immediately so they're in the pipeline by the time you book. Auto-suppress lists from negative replies, every campaign on every domain forever.

**Common reply-layer breakdowns:**

- SDR response lag (2-3 hours during business hours) → fix with response-time SLAs and dashboards
- No qualification at reply → all replies treated as equal, AE time wasted on Tier 3
- Replies handled in personal inbox, not unified view → dropped messages, missed routing
- No suppression of negatives → re-email someone who said no, generate spam complaint
- No re-engagement on "not now" → 6-month-later opportunity goes to a competitor who tracked timing

The send side gets all the attention because it's measurable and visible. The reply side is where revenue actually books or doesn't. Run both with discipline.

---

## The process when triggered

When the user says "write me a cold outbound sequence" (or any trigger), run this:

### Step 1: Diagnose the prerequisites

Ask in this order:

1. **Do you have an ICP defined?** If no, run `icp-builder` first or build a quick prototype here. Outbound without an ICP is just spam at scale.
2. **Which Tier 1 segment is this sequence for?** One sequence per behavioral trigger × persona combination. "B2B SaaS" is not a segment. "Series B SaaS that just hired a VP Sales in last 60 days" is.
3. **Which persona are we targeting?** Champion, Economic Buyer, or End User from the ICP. Each persona reads different copy.
4. **What's the behavioral trigger?** New hire? Funding? Hiring spree? Layoff? Product launch? The trigger is the opener.
5. **What's the result you produce?** Not the product, the result. "Cut SDR ramp from 6 months to 9 weeks" beats "AI-powered enablement platform."

If any of those aren't clear, stop and clarify. Don't draft sequences against a vague brief, you'll have to rewrite once they get specific.

### Step 2: Audit infrastructure

Before writing copy, confirm:

- Sending domain is NOT the primary domain
- SPF / DKIM / DMARC configured
- Warmup completed (score 80+)
- Inbox volume cap respected (≤50/day per inbox)
- List verified (NeverBounce or equivalent, <5% bounce projected)
- Reply tracking on, open tracking off

If any are missing, flag it. Tell the user explicitly: "Even if I write you a perfect sequence, it won't work until X is fixed." Don't draft over broken infrastructure.

### Step 3: Draft the sequence

Produce all 5 touches following the framework. Each touch:

- 3 subject line variants
- Body copy in the TRIGGER → PROBLEM → VALUE → MICRO ASK pattern
- Mapped to the day in the sequence
- Personalization variables flagged (`{{firstName}}`, `{{companyName}}`, `{{customSnippet1}}`)

Format the output as ready-to-paste Instantly / Smartlead HTML (Gmail-style `<div>` blocks for line breaks, no fancy formatting). If the user is using a different tool, convert.

### Step 4: Stress-test

Three checks before declaring it ready to send:

1. **Specificity test**, read touch 1's first sentence. Could it be sent to any company in the segment, or is it specific to THIS prospect's behavioral trigger? If generic, rewrite.
2. **Result test**, find the value sentence in touch 1. Does it name a result with a number, or does it pitch a feature? Rewrite features into results.
3. **Out test**, read touch 5. Does it cleanly let the prospect off the hook with permission, or does it guilt-trip ("haven't heard back...")? Rewrite if guilt-trippy.

If any fail, fix before sending.

### Step 5: A/B test plan

Recommend the user A/B test ONE variable in the first 200 sends:

- Subject lines on touch 1 (3 variants × ~70 sends each)
- OR opener angles on touch 1 (2 variants × ~100 sends each)
- OR CTA wording on touch 1 (2 variants × ~100 sends each)

Don't test more than one variable per run. Read the data after 200 sends, then test the next variable.

---

## The artifact (template)

Produce a single markdown document with this structure:

```markdown
# Cold Outbound Sequence, [Segment / Persona / Trigger]

_Last updated: [date]. Owner: [name]. Tier 1 segment: [from ICP]. Persona: [from ICP]. Trigger: [behavioral signal]._

## Pre-send checklist

- [ ] Sending from non-primary domain: \***\*\_\_\_\_\*\***
- [ ] SPF / DKIM / DMARC verified
- [ ] Warmup score 80+ (date confirmed: \_\_\_\_)
- [ ] Inbox cap ≤50/day, multi-inbox if scaling
- [ ] List verified through NeverBounce / Million Verifier
- [ ] Reply tracking ON, open tracking OFF
- [ ] Sequence stops on any reply

## Sequence

### Touch 1 (Day 0): Trigger opener

**Subject line variants (test 3):**

1. ...
2. ...
3. ...

**Body:**
```

[email body in HTML for Instantly, plain text otherwise]

```

**Personalization variables used:** {{firstName}}, {{companyName}}, {{customSnippet1}}

### Touch 2 (Day +3): Social proof / value-add
[Same structure.]

### Touch 3 (Day +6): Short bump
[Same structure.]

### Touch 4 (Day +10): Pattern interrupt
[Same structure.]

### Touch 5 (Day +14): Breakup
[Same structure.]

## A/B test plan
- First test: [subject lines / opener / CTA] on touch 1
- Sends per variant: 100+
- Read data after: 200 total sends
- Decision rule: variant wins if positive reply rate > 0.5pp higher than control

## Metrics targets
- Reply rate: 2.5%+ (warning <1.5%, broken <1%)
- Positive reply rate: 1.5%+
- Bounce: <3%
- Spam complaint: <0.05%
- Meeting booked: 0.5%+

## When to kill the campaign
- Bounce rate >5% → list quality issue, pause and clean
- Spam complaint >0.1% → deliverability issue, pause and audit
- Positive reply rate <0.3% after 500 sends → copy or trigger is wrong, rewrite
```

---

## Common mistakes

Push back on these when you see them:

- **Pitching the product instead of the result.** "AI-powered sales enablement platform" is a feature. "Cut SDR ramp from 6 months to 9 weeks" is a result. Sell the result.
- **Generic openers.** "Hope this finds you well" / "My name is X and I'm with Y", both dead on arrival. First sentence has to reference the prospect specifically.
- **Sending from main domain.** Burns your primary email. Always use a lookalike domain for cold outbound.
- **Skipping warmup.** Warming takes 4-6 weeks. Most operators try to skip it and wonder why their reply rate is 0.3%. Warmup isn't optional.
- **Bad list.** Catch-alls, role accounts, unverified emails, scraped data, all kill deliverability and reply rate. List quality matters more than copy.
- **Too many touches.** 8-9 touches was the 2018 playbook. Modern data: replies plateau at touch 5, spam complaints rise after that. Stop at 5.
- **Calendar link in touch 1.** Signals commerce, kills deliverability, and assumes intent the prospect hasn't earned. Calendar links go in touch 2-3 after engagement.
- **"Just checking in" follow-ups.** Adds nothing. Each touch must have a new angle, a new value, or a new micro-ask. If you can't think of one, send the breakup early.
- **Same sequence for every persona.** Champion reads different copy than Economic Buyer. Run separate sequences per persona within the same segment.
- **Open rate as KPI.** Broken since iOS 15. Track replies, not opens.
- **Optimizing copy before fixing infrastructure.** If bounce is 8% or warmup wasn't done, no amount of copy iteration will save you. Fix the upstream issue first.

---

## How to use the artifact downstream

After the sequence ships:

1. **Load into Instantly / Smartlead / Lemlist** with personalization variables wired up
2. **Connect list source** (Clay, Apollo) for ongoing trigger-driven feed
3. **Configure stop conditions**, sequence halts on any reply, manual unsubscribe, bounce
4. **Set up reply triage**, positive replies route to AE for booking, negatives auto-suppress
5. **Track metrics weekly**, reply rate, positive reply rate, bounce, spam, meetings booked
6. **Iterate one variable at a time**, A/B test subject, then opener, then CTA, on 200-send rolls
7. **Refresh sequence every 90 days**, copy decays as triggers age and prospect inbox fatigue rises
8. **Cross-reference with `icp-builder`** when the segment isn't producing, the issue might be ICP, not copy

---

## A note on starting from zero

If you're setting up cold outbound for the first time (no inboxes, no warmup, no list yet), don't expect to send for 4-6 weeks. The setup work is:

- Week 1: Buy lookalike domain(s), configure DNS records, create 2-3 inboxes
- Week 2-5: Warmup running, score climbing toward 80+
- Week 3-5: Build initial Tier 1 list (200-500 accounts) in Apollo or Clay
- Week 4: Verify list, write sequence, configure tool, test internally
- Week 6: First real sends, 30-50/day per inbox, ramp slowly
- Week 8: First real reply data, first A/B test

Operators who try to compress this to 1 week burn their domain and start over. The sequence works only on top of clean infrastructure. Build the infrastructure first.

---

**Cold outbound is not a copywriting exercise. It's a deliverability problem with a list problem with a positioning problem on top.**

---

_Part of the StackSwap Operator Playbook. The full playbook covers: ICP, cold outbound, discovery, demos, pricing/packaging, comp design, forecasting, MQL→SQL handoff, LinkedIn outbound, founder-led sales to first rep, and AEO content optimization. → stackswap.ai/playbook_

Free forever. No email gate.