PostEaseChange one variable at a time
Testing a new hook style and a new posting time in the same post makes it impossible to tell which change mattered. Isolate a single variable across otherwise similar posts: the opening line, the format, the posting time. That is the difference between an informal test and a guess dressed up as data.
Compare genuinely similar posts, not a random sample
Comparing a test post against a completely different, unrelated post from months ago introduces too many other variables to draw a real conclusion. Comparing it against a recent post covering a similar topic, in a similar format, isolates the actual variable being tested far more reliably.
Give a test enough posts to show a pattern
A single post performing better proves very little on its own. Performance is noisy. Repeating a variable across three or four posts before drawing a conclusion gives a far more reliable read than judging an entire approach off one post.