You are right. The author should plug the numbers in here: http://www.usereffect.com/split-test-calculator http://www.usereffect.com/split-test-calculator I'd

You are right. The author should plug the numbers in here: http://www.usereffect.com/split-test-calculator http://www.usereffect.com/split-test-calculator I'd go as far as to consider the data to be worthless without the confidence value