A/B test email subject lines

I need the A/B test set up for a subject-line trial to the current leads list. Use the two variants we agreed: “Quarterly insights inside” and “Your team’s Q2 results — quick look.” Send 10% to each variant and the remainder to the winning subject after three days by open-rate. Exclude anyone who opened a campaign in the last seven days. Schedule the test to start Thursday at 09:00 and confirm the test audience counts before sending.

Before I launch this subject-line A/B test, make it obvious for a reviewer what each subject promises, where the preview text differs, and how wins will be decided. Pull the last three subject-line winners for this list, flag any subject that reused the same preview text, and surface the expected audience sizes and spam-score flag if any. Suggest one alternative subject that’s more urgent than the two we picked.

I scheduled two subject lines to the product-updates list, with Friday as our send. I worry sending 50/50 will waste opens if one subject is clearly stronger. The buyer is a small startup CS team who ignore fluff; I'm afraid the test won’t produce a decisive winner in three days. Should I run a 20/80 probe instead, or keep 50/50 to avoid bias? What is the most defensible split and why?

Over the last six months we’ve launched weekly subject-line A/B tests and kept changing wording, but open-rate gains are temporary and we repeat the same mistakes. I suspect we lack consistent hypotheses, reuse weak preview text, and ignore segment behavior. What three habits should I change in how we choose subjects, define success, and analyze winners so tests produce repeatable lift?