You don't always need to find a winner. I've found that advertisers are often unnecessarily obsessed with finding the perfect combination of primary text, headline, and creative. Even if Meta can find a winner, the sample size is often statistically insignificant. There's a time and place for split testing, but in most cases you should take a much simpler approach. For more on creative testing, check out my blog post at https://jonloomer.com/testing.