Cold email reply rates have collapsed industry-wide: generic sequences average 1–3%. Signal-based personalization — where the opening line references something specific that happened to the prospect's company — averages 15–25% reply rates when the signal is relevant.
The problem is research time. Writing a personalized first line for 200 prospects takes 8–10 hours manually. Clawdbot reduces it to 30 minutes: you provide the signal source, it generates the personalized openers and full email drafts at batch scale.
Why Signals Beat Templates
| Approach | Avg reply rate | Why |
|---|---|---|
| Generic template | 1–3% | Immediately recognizable as mass outreach |
| Persona-level personalization | 3–6% | Same message, slightly different framing |
| Account-level research | 8–12% | Specific to company; feels researched |
| Signal-based (trigger event) | 15–25% | Relevant to what's happening at the company right now |
Signal Types for Outreach
Fundraising announcement
'Congrats on the Series B — scaling [function] is usually the first thing that becomes a constraint at this stage.'
New executive hire
'Saw [Name] just joined as [Role] — new leaders usually re-evaluate the toolstack in their first 90 days.'
Job posting for role your product replaces
'Noticed you're hiring [Role] — we help teams skip that hire by automating [function].'
Company news or expansion
'Read about your expansion into [market] — [problem] tends to follow that move.'
Technology change (BuiltWith)
'You recently added [tool] — teams usually hit [pain] when doing that at scale.'
Email Structure
Keep it under 100 words. Long emails signal low confidence and get deleted. Structure:
Signal hook — reference the specific trigger
Bridge — connect the signal to the pain you solve
One-sentence product value prop — outcome, not features
Proof — one specific customer result or number
CTA — one question, not a meeting request
Subject: [Company] + [your category]
Noticed you're hiring a [Role] — that usually means [pain] is becoming a real drag on [function].
We help [similar companies] [specific outcome] without adding headcount.
[Customer name] went from [before] to [after] in [timeframe].
Worth a quick look, or not the right moment?
Sequence Logic
Signal-based personalized opener
Different angle — case study or data point. Not a 'just following up'.
Breakup email — 'Okay to close your file?' — counterintuitively high reply rate
Connection request + 1-line message if no email reply
Three touches maximum per prospect per signal. If no reply after 3, remove them until a new signal triggers re-engagement.
Deliverability Basics
Send from a secondary domain (company-email.com), never your main domain
Warm up sending domain for 30 days before sequences — use Instantly or Lemlist warmup
Cap at 30–50 new emails/day per domain until reputation is established
SPF, DKIM, DMARC records configured — non-negotiable for inbox delivery
Keep bounce rate below 3% — verify emails with Hunter or ZeroBounce before sending
Clawdbot Configuration
| Setting | Value |
|---|---|
| Signal input | Prospect list with signal field (funding, job post, news, tech change) |
| Personalization | Unique line 1 per prospect generated from signal data |
| Output | Full 3-email sequence per prospect + LinkedIn message |
| Review | Human review of first 20 outputs before batch launch |
| Volume | 50–100 personalized emails/day across 2–3 sending domains |
| Tracking | Reply rate, meeting rate per signal type — optimize toward highest |
Foundation
Haven't set up Clawdbot yet?
OpenClaw + Telegram + Claude. Takes ~20 minutes.
