You've been running the same cold email for three weeks. Opens are decent. Reply rate is... fine. Not great, but not terrible. So the question that keeps hitting you: should I change this or keep going?
Most people get this wrong. They either change copy too early (before they have enough data) or stick with something broken for way too long (because they're scared to mess with what's working). Both cost you deals.
Here's how to actually know when it's time to change your cold email copy - and when to hold steady.
The Data Threshold: Don't Change Until You Have This
First rule: you need a minimum sample size before any numbers mean anything. If you've sent 20 emails and gotten 2 replies, that's not a trend - that's noise.
The practical benchmark is 50-100 emails sent through a single variation. At that volume, you start seeing real patterns. Below that, you're just guessing.
Why 50-100? Because cold email reply rates naturally vary based on list quality, day of week, recipient's inbox state, and a dozen other factors outside your control. You need enough volume to see signal through that noise.
So if you've sent 25 emails with a piece of copy and you're not loving the results - wait. Send another 50. Then look at the numbers.
The Metrics That Actually Matter
Once you hit 50+ emails, there are three numbers worth watching:
1. Reply rate (the real one, not including auto-replies) - This is responses divided by emails delivered. Your baseline benchmark is 5-8% for most B2B cold email. Below 3%? That's a signal something's wrong.
2. Positive reply rate - Not all replies are good. Some are