It’s insane to me how much mileage a poll with 1200 participants will get in the media. I wouldn’t trust an a/b test with 12,000 responses. You’re telling me your 10% +/- polling number is accurate with 1200 randos that answered your phone call? GTFO of here.