Cold email reply rates: how to measure them properly (measurement guide)

Revision note (29 September 2026): an earlier version of this article was titled "real data" and published reply-rate ranges by email type and industry, with unlinked references and a row attributed to "OpenSells data". We did not have a documented sample to support them, so we removed them. Instead, this guide explains how to measure your own data.

Why we don't publish an "average rate"

Reply rates depend on who you write to, whether they expected your message, the offer, the industry, the country and even the day. A general number mixes campaigns that have nothing in common. What helps is comparing your own sends, with a stable method.

A legal note: in Spain, art. 21 LSSI prohibits unsolicited or unauthorised commercial emails, including to businesses. Apply this method to sends that have a legal basis.

1. Define the metrics before sending

• Sent and bounced: delivered = sent − bounced.
• Reply rate: human replies ÷ delivered. Exclude auto-replies.
• Positive reply rate: replies asking for information or agreeing to talk ÷ delivered. Define "positive" in advance.
• Meetings and deals: what actually matters, per campaign.
• Opt-outs and complaints: a warning sign, not a detail.

Don't use open rate as your main indicator: many email clients preload or block images, so opens over- or under-count.

2. Decide the sample and time window

• Group comparable sends: same industry, same offer, same type of contact.
• Don't draw conclusions from a few dozen sends: one or two extra replies swing the percentage a lot.
• Fix a reply window (for example 14 days after the last message) and apply it to every group.
• Track follow-ups separately: note which message in the sequence each reply answers.

3. Change one variable at a time

If you change the subject, the copy and the industry together, you won't know what worked.

4. Log it in a spreadsheet

Suggested columns: campaign/group, industry and city, variant tested, send date, delivered (yes/no), reply (none/positive/negative/opt-out), message replied to, meeting/deal.

5. Know the limits

• Replies through other channels (they call you, message you) don't appear on their own: log them.
• Small groups give unstable percentages.
• What works in one industry or country doesn't transfer as-is.
• Sender reputation matters: Google requires domain authentication and a spam rate below 0.3% to deliver to Gmail.

If you call instead

The same method works: calls made, answered, useful conversations, interested, meetings. In OpenSells you can change each lead's status, add notes and tasks, and export your leads to CSV on paid plans to analyse them.

Try OpenSells free: first month of the Professional plan free, no card; then Professional €39/mo and Agency €89/mo including Spanish VAT (29 USD and 69 USD when paying from Latin America). See pricing and terms. Note that the app interface is currently in Spanish.