Cold email bounce rate: hello@ and contact@ bounced 7%

We run GrowFast on our own products first. In the week to 23 October 2026 it sent 160 cold emails for seven of them.

| Result | Count | | --- | --- | | Delivered | 151 | | Bounced | 5 (3.1%) | | No delivery status yet | 4 | | Replies | 0 |

Zero replies is a real result, and we are not hiding it. The bounce number is the one we could fix straight away.

Where the bounces came from

Four of the five bounces were generic greeting addresses: hello@ and contact@. Those two local parts accounted for 57 sends and 4 bounces, about 7%. The other 103 addresses had 1 bounce, about 1%. That is a small sample, so read the gap as a signal to act on, not a measured rule.

The addresses were published on each company's own website, so they looked safe. A published address is not the same as a mailbox somebody reads, and some of them point nowhere.

Why it matters more than the percentage

Bounces hurt the sender, not the recipient. Mail providers score a sending domain by its bounce and complaint rates, and a domain that sits near 4% risks suspension. Our sending infrastructure is shared with paying customers, so one noisy campaign is everybody's problem.

What we changed

1. Stopped sending to hello@, hi@ and contact@. GrowFast now writes only to a named person's own address, or to a role address that has not shown a bounce problem.
2. Check that the domain can receive mail. Before sending, GrowFast looks up the recipient's mail records. A domain that does not exist is dropped. A flaky DNS answer is not treated as a failure, so a slow resolver cannot cost you a good lead.
3. Paused sending until the rate is measured again. A rate that is high on 160 emails is a reason to stop and fix, not to send more and hope.

What to take from it

• Measure bounce rate per address type, not just in total. The total hid a 7% problem.
• Prefer a named person's published address over a generic inbox, even if it means fewer emails. We have not yet measured whether this lowers bounces; sending is paused until we do.
• Treat zero replies as a signal about who you email and what you say, which we are working on next.

We will update this post when the bounce rate has been measured again after the changes.