SuperAPI AI Directory

cold-email-copywriter

Office & productivity

View repository

Updated 10-15 20:25

About this skill

Cold Email Copywriter

Good cold emails are < 90 words, reference something only that prospect would recognize, and ask for one small thing.

Workflow

  1. Gather inputs. Refuse to write without:
    • Prospect's role + company + one specific recent signal (post, hire, launch, news)
    • What the user sells in one concrete sentence
    • Proof point (customer + quantified outcome)
    • The ask — reply, 15 min, a resource
  2. Draft 3 variants with different angles:
    • A: Observation → hypothesis → ask (default)
    • B: Question-led (opens with a question the prospect cares about)
    • C: Peer proof (leads with similar customer's result)
  3. Score each against the checklist (see below) before presenting.
  4. Return the 3 emails + a one-line rationale for each.

The 90-word template (Variant A)

Subject: <4-6 words, lowercase, specific>

Hi <Name>,

<Observation: one specific thing they did/posted/shipped — 1 sentence>.

<Hypothesis: what that likely means for their work — 1 sentence that shows you get it>.

<Proof: "We helped <peer> <specific metric>." — 1 sentence>.

<Ask: open question, not a demo request — "Curious how you're handling <X>?">.

— <You>

Checklist (every email must pass)

  • < 90 words in body
  • Subject ≤ 6 words, lowercase, no emoji
  • First line is about them, not you/your company
  • One specific reference only that prospect's team would recognize
  • Proof point has a number
  • Ask is a question, not a calendar link (unless a clear warm lead)
  • No "hope you're well", "circling back", "quick question", "touch base"

Anti-patterns (reject)

  • "I know you're busy so I'll keep this brief" — then take 200 words
  • Feature lists
  • Generic vertical claims ("helps SaaS companies grow")
  • Multiple CTAs
  • More than one link

Guardrails

  • If the user can't supply a specific signal about the prospect, stop and research first — don't ship generic.
  • For technical buyers, replace the proof-point-metric with an architectural detail.
  • If the user insists on a generic template for 1000 prospects, warn about deliverability and reply rate (< 1%).