A spend of $4.58 is not statistically significant, and FB is trying to give your ads enough spend that it doesn't make a mistake.
Let's say you have an ad set that has 5 separate ads, that get shown a total of 10 times each, this is 50 impressions in total.
If one ad gets 2 sales, and the other ads get 0 sales, you're going to assume that the ad that got 2 sales is the winner, and the other 4 are the losers - the problem with this is that you don't have enough data yet.
If you ran those same ads to 1,000 impressions, it's entirely possible that the one ad that got 2 sales, ends up generating 0 more sales. The other 4 ads may end up generating, more, or less, than 0 sales - but you can't know for sure until you've reached a statistically significant number of impressions to a wide enough audience.
Over time, the ads will continue to adjust themselves in an attempt to slowly favor the better performers, while slowly eliminating the ones that perform worse. That said, you should also prune your ads based on your own knowledge of the size of your audience.
If you're selling a very niche product, you should give your ads less leeway, if you're selling a product that appeals to a much larger audience, you should give the system more time as it's going to take longer to reach a portion of your audience that's going to be statistically significant.
I personally won't adjust an ad until I've seen at least 5,000 impressions, per variation. Yes, this can get expensive, but it has also shown me that what you think you know on day one, may be totally different than what you actually know by the time you reach day 7.