The only subject-line test worth running

The only subject-line test worth running

Popular advice: "A/B test subject lines to lift open rates." Myth. You're optimizing the wrong number and most tools call an apple-loading pixel an "open."

Run this instead:
— Send version A to 30 contacts, version B to 30, same list, same hour.
— Ignore opens entirely. Count replies only.
— Kill any test under 50 sends per arm — the math is noise below that.
— Keep the winner for 200 sends before testing again.

A subject line's job isn't to get opened, it's to not lie about the body. Test for honesty, not curiosity.