Skip to content
EMPRAOUTBOUND
EngineMethodDeliverabilityIntentReply agentAutomationsMonitored markets
Apply
EngineMethodDeliverabilityIntentReply agentAutomationsMonitored marketsApply

EMPRA

EMPRA

Outbound

RUN BY HUGO DUPONT & GEORGE LEVESON
LONDON

WHAT WE RUN

  • The engine
  • Deliverability
  • Intent signalling
  • Reply agent
  • Automations
  • Method
  • Monitored markets

FREE TOOLS

  • Grade my email
  • Deliverability test
  • SPF checker
  • DKIM checker
  • DMARC checker
  • Subject line tester
  • Spam checker

THE PLAYBOOK

  • Playbook overview
  • Subject lines
  • The ask
  • The sequence
  • Follow-ups
  • Identify your ICP
  • Infrastructure

COMPANY

  • Cold email agency
  • Agency shortlist
  • Agency vs DIY
  • Blog
  • About
  • Start a build

Every figure on this site is a count of work we have done. Pipeline and closed revenue are our clients’ results, and results vary. Replies are shown as they arrived, with senders blurred at source.

The sequence

Four touches, built on evidence.

A cold email sequence is not eleven emails wearing the reader down. It is four, each carrying something the last one did not: a fact, a new fact, proof, and the honest walk-away. Here is the shape, the cadence, and the measured rule behind each step.

  1. Touch 1Day 0

    The opener

    Under 40 words. Opens with a fact about the reader, checkable from their own site. One small ask, and it is not a meeting. This is the email the data is strictest about: shorter first touches meaningfully outperform longer ones across the accounts we measure.

  2. Touch 2Day 2–3

    The new fact

    A follow-up earns its place with information the opener did not carry. Not “just checking in” — a competitor moved, a number changed, a peer did the thing. Same length discipline, same single ask.

  3. Touch 3Day 6–8

    The proof

    One named, checkable outcome from a peer. The ask drops to a single word to reply with. This is where a real share of replies actually land, because now there is evidence on the table.

  4. Touch 4Day 12–14

    The walk-away

    State the standing offer once, wish them well on something real, and leave. The walk-away converts because it is true — and then you stop. Four touches, not eleven.

EVIDENCE LABELLED AS EVIDENCE; CADENCE IS CRAFT JUDGEMENT. MEASURED FINDINGS SPAN 44 CLIENT ACCOUNTS.

Test it

Run your sequence through the machine.

The grader checks each email against these rules: length, the five-sentence ceiling, the meeting-ask penalty, the researched-fact requirement. Free, instant, nothing stored.

Grade an emailSee the follow-up templates

Questions

The details.

How many emails should a cold email sequence have?

Four: an opener and three follow-ups, then stop. Each one must carry something new — a fact, then a new fact, then proof, then the honest walk-away. Past four you are not persisting, you are loitering, and loitering is how sending domains get marked as spam.

How long should a cold email be?

The first touch should be under 40 words. In our corpus, first touches under 40 words meaningfully outperform emails in the 40-to-69-word band, measured across 44 client accounts — and it is the one copy rule that survives scrutiny across the whole dataset. Keep every email under five sentences; past five, reply rates drop about 30%.

What is the right cadence between cold emails?

Roughly: two to three working days to the first follow-up, four to five to the proof email, about a week to the walk-away. This spacing is craft judgement, not a measured law. The real rule is that a follow-up sends when you have something new to say, never because a timer fired.

Should every email be in the same thread?

Keep the opener and first follow-up on one thread so context travels. Break the thread for the proof email if its subject line is doing new work. The walk-away goes back on the thread, because it ends a conversation rather than starting one.

Should I A/B test my sequence?

Not at low volume. Cold email positive rates are low enough that splitting a list of a few thousand three ways gives you about one positive per variant, which answers nothing. Below roughly 30,000 sends, write one good sequence and put the effort into research per account instead.

Part of the cold email playbook. Go deeper on the subject line, the ask, or hand the whole thing to a cold email agency.