Judge sequences, not single emails. Across a 3-5 email sequence: Open rate 70%+ amazing, 50-70% great, under 30% is a deliverability problem Reply rate 10%+ amazing, 5-10% great, 3-5% fine, under 3% rebuild

Write all four emails before you send any. Email 1: why them, the problem, one question Email 2: one proof point Email 3: new angle, same person Email 4: the close-out Writing them together is what stops email 3 being just following up.

Your first sequence needs 4 emails, not 15. One ask. Three follow ups that each add something new. Then stop, politely. If you cannot fill four emails with new information, your offer is too thin.

Personalization is not a compliment about their recent post. It is proving you know which problem they have this quarter. If your first line could go to 500 people, it is not personal. It is padding.

Client stuck at 1.2% replies with a 38% open rate. Clean list, good setup. We turned off open tracking. Reply rate went to 2.6%. Tracking pixels get noticed by filters. You lose one number and gain the one that pays you.

Copilot: the AI drafts, you approve. Autopilot: it sends. Run copilot two weeks first. Whatever you keep rewriting is what your AI still does not know about your business.

Never send to an unverified list. Run it through a verifier. Delete hard bounces. Put catch-all addresses in their own group and send them last, in small batches. They bounce at 5-15%. Ten minutes. Highest return hour in outbound.