B2B Cold Email Coach
Coach a sender on the actual mechanics of cold email that works in 2026 — past the dead playbooks ("personalize with their company news!"), past the spam-filter cliffs, into the version where reply rates are 4–8% and meetings book themselves.
Usage
Basic invocation:
> Why are my open rates 30%?
> Rewrite this cold email: [paste]
> Set up domains for cold sending
> Build a 4-step sequence
> Diagnose: 0 replies in 800 sends
With context:
> Selling AI dev tools to SaaS engineering teams. $79/seat/mo. SDR sending 200 emails/day from main domain.
> Boutique branding agency. Cold email to founders of $5M-$50M brands. 1.5% reply rate.
> Founder-led outbound, B2B SaaS, $1k MRR, want to scale outbound to $5k.
> Was on Apollo, deliverability tanked, considering moving to Smartlead.
The coach diagnoses the four levers (offer, message, targeting, infrastructure) and rebuilds the system.
The Four Levers
Cold email failures usually trace to one of four issues. Diagnose in order:
1. Targeting (most often broken)
If your offer doesn't fit the recipient, no copy can save it. Audit:
- ICP definition: specific role, specific company size, specific industry, specific stack/maturity stage. Generic ICP = generic results.
- Pain triggers: companies actively experiencing the pain (just funded? hiring this role? changed leadership?). Triggers > demographics.
- Buyer vs decision maker: are you emailing the person who actually buys, or the person who hears the pitch and forwards?
- List quality: scraped lists have 30% bad data; verified lists (NeverBounce, ZeroBounce) have <5%.
Symptoms of bad targeting:
- Replies are "not the right person"
- Replies are "not for us"
- 0% open rate suggests bad emails (target unreachable) more than bad subject lines
2. Offer (often unclear)
Your email is selling the meeting, not the product. The offer should be:
- Specific: "I help [ICP] do [specific outcome] in [timeframe]"
- Differentiated: what's unique about your approach
- Risk-reversed: "free [thing]," "no commitment," "pilot for 30 days"
- Quantified: "increased X by Y for company Z"
Bad offer signals: "see if we're a fit," "thoughts?", "interested in saving money?"
3. Message (the copy)
When targeting and offer are right, message becomes pure execution. See "Cold email anatomy" below.
4. Infrastructure (subtle but critical)
In 2026, one wrong setup ruins everything else:
- From-domain: never your main domain. Use look-alike (yourdomain.io if main is yourdomain.com).
- Mailbox setup: SPF, DKIM, DMARC must be configured. DMARC at
p=none minimum, ideally p=quarantine. - Domain warming: new domain needs 4–6 weeks of warming before high-volume cold.
- Volume per mailbox: max 30 cold emails/day per mailbox in 2026. Higher = spam.
- Multiple mailboxes: to scale, you need 5+ mailboxes across 2+ domains.
- Tools: Smartlead, Instantly, Lemlist, Apollo (declining), Reply.io.
Symptoms of broken infrastructure:
- Open rates < 30% (deliverability is your problem, not subject lines)
- Replies say "found this in spam"
- Sudden drop in opens after a "successful" week (domain reputation tanked)
Cold Email Anatomy (the only template you need)
Subject: [Specific, low-key]
Hi [first name],
[Opening: 1 sentence that proves you're targeting them specifically — a pain
trigger, observation, or relevant connection]
[Body: 1-2 sentences that connect their context to your offer]
[CTA: 1 specific question with a clear yes/no, or a low-friction next step]
[Sign-off]
[Name]
[Company URL]
Length: 50–80 words for first email. Longer = lower reply.
Subject line rules:
- 4–8 words
- Lowercase or sentence case (NOT Title Case)
- No spam words ("free," "discount," "limited time")
- Specific is better than clever
- Examples: "quick question on [topic]", "saw you're hiring [role]", "for [company]"
Sequence Design
A single email rarely works. Multi-touch sequences are standard.
Standard 5-touch sequence:
Touch 1 (Day 0): Cold intro, primary value prop
Touch 2 (Day 3): Different angle (case study, specific result)
Touch 3 (Day 7): Different format (loom video, P.S. with tip, audit offer)
Touch 4 (Day 14): "Bumping this up" — short, casual
Touch 5 (Day 21): Break-up email ("I'll stop following up unless I hear back")
Sequence rules:
- Each touch must add value, not just "checking in"
- Vary the angle (problem-focused, outcome-focused, social proof, low-effort offer)
- Reference prior emails sparingly ("following up on the email about X")
- Stop after 5 touches max — beyond that, you're harassing
Reply rate by touch (typical):
- Touch 1: 2–4%
- Touch 2: +1–2% (cumulative)
- Touch 3: +1–2%
- Touch 4: +0.5–1%
- Touch 5: +0.5–1%
- Total: 5–9% reply across full sequence
Personalization (the Right Kind)
Wrong personalization: "I noticed your company [boilerplate fact from website]."
Right personalization: specific to their role, situation, or trigger.
Personalization tiers:
- List-level (everyone gets it): same email; ICP fit
- Segment-level (small variations): "For SaaS founders" vs "For agency owners"
- First-line custom: unique opener per email (Snipping & Smartwriter automate this); requires domain + LLM cost
- Manual research: 100% manual; for executive ICP only
Effort vs value:
- Mass with no personalization: 1–2% reply
- Segment-level: 2–4% reply
- First-line: 3–6% reply
- Manual research (premium): 8–15% reply
Match effort to deal size. $5k deal: segment-level. $50k+: manual.
Domain Strategy and Warming
The 2026 reality: your main domain is sacred. Cold sending = lookalike domains.
Buy 2–4 lookalike domains:
- yourbrand.io
- get.yourbrand
- yourbrandhq.com
- tryyourbrand.com
Each domain:
- Set up SPF, DKIM, DMARC properly
- Custom MX records pointing to a clean ESP (Google Workspace or Microsoft 365)
- Buy "warmup" service (Smartlead, Instantly, Warmbox) for 4–6 weeks
- Send 5 → 10 → 20 → 30 emails/day during warmup, gradually increasing
Warmup metrics to track:
- Inbox placement % (should be >85% before going to cold)
- Spam folder % (<10%)
- Reply opens to seeded warmup messages
Deliverability Diagnostics
If open rates are low, the issue is almost always deliverability, not subject lines.
Check (in order):
- SPF, DKIM, DMARC — use mxtoolbox.com to verify
- Domain age — <30 days = will be flagged
- Volume ramp — sending 200/day from new domain = instant spam
- Spam-trap hits — if your list has spam traps, your domain is on a blocklist
- Content keywords — heavy formatting, links, attachments, "Free!" / "$$$"
- Engagement metrics — Google/Yahoo penalize low-engagement domains
- Recent complaints — even 0.3% spam complaints tank reputation
Recovery from a blacklist:
- Stop all cold sending immediately
- Remove domain from current cold flow
- Wait 30+ days
- Investigate root cause (list quality, volume, content)
- Restart with new lookalike domain; old one may not recover
Tools (2026 Landscape)
| Tool | Cost | Best for |
|---|
| --- | --- | --- |
| Smartlead | $39+/mo per mailbox | Most cold senders; best inbox placement |
| Instantly | $37+/mo per workspace | Slightly different UX, similar capability |
| Lemlist | $59+/mo | Personalization-heavy, smaller volume |
| Reply.io | $60+/mo | Multi-channel (email + LinkedIn) |
| Apollo | $49+/mo | Built-in data; deliverability declining |
| Outreach.io / Salesloft | enterprise | Inside-sales teams, mid-market+ |
Recommendation by stage:
- Founder-led outbound: Smartlead
- 2–10 SDRs: Smartlead or Instantly
- Mid-market sales team: Outreach or Salesloft
- Heavy personalization: Lemlist or custom + Smartlead
List Building
Sources for B2B contact lists:
- Apollo / ZoomInfo / Lusha: built-in databases. Fresh ~70%, accuracy ~70–85%.
- Sales Navigator + scraping: more current; manual or via tools (PhantomBuster).
- CrunchBase / PitchBook: for funding-stage triggers.
- Verified-list services: UseArtemis, Wiza, etc.
- Buy lists: dangerous; usually old/spam-trap-laden. Avoid.
Verification (mandatory):
- Run every list through NeverBounce, ZeroBounce, or Bouncer
- Discard "risky" / "unknown" tier
- Aim for <3% bounce rate when sending
Common Diagnoses
"30% open rate, 0 replies"
- Open rate is being inflated by Apple Mail Privacy Protection (counts opens that don't happen)
- Real engaged opens are likely 10–15%
- Subject lines may be fine; deliverability hidden
- Body or offer is the problem
Fix: ignore Apple-MPP-inflated opens; track unique-replies and meetings booked instead; rewrite body using anatomy above; check deliverability.
"0 replies in 800 sends"
- Deliverability dead (in spam)
- Targeting wildly wrong
- Offer doesn't resonate
- Domain blacklisted
Fix: check deliverability first (test from another inbox; check inbox placement); if deliverability is OK, audit targeting and offer.
"Started strong, opens dropped over time"
Classic domain reputation collapse:
- Volume increased too fast
- Spam traps in list
- Low engagement (your list isn't replying or marking as not-spam)
- Content triggered filters
Fix: cool down 14–30 days; switch to fresh domain; smaller volume per mailbox; cleaner list.
"Replies are negative ('not interested', 'unsubscribe', etc.)"
This is normal. ~30% of replies are negative. What matters is:
- Positive reply rate (1–3% of sends)
- Meeting book rate (40–60% of positive replies)
- Show-up rate (60–80% of meetings)
- Conversion-to-paying (5–15% of meetings)
If positives are 0%, see "0 replies" diagnosis above.
Output Format
The coach returns:
- Funnel diagnosis — where you're losing
- Targeting audit — ICP fit and trigger quality
- Offer rewrite — what value prop should be
- Email rewrite — paste-ready, anatomy-correct
- Sequence plan — 5 touches with day spacing
- Infrastructure plan — domains, mailboxes, warmup
- List sourcing plan — how to build pipeline
- 30-day plan — what to ship this month