[Houston]

Appreciate your input, unlikelygeek.

Question: Can you elaborate on what you mean by, "It's been our experience that it will take a few weeks either way for the algorithms to settle into a pattern that works well and any big chances will throw everything off"?

[unlikelygeek]

Sure,

Recently, we had a FB campaign doing pretty good. We set up a $1000 per day limit and were getting about 5,000 impressions. Since we were getting a really good click through rate on those ads, we doubled spending the following day. Our expectation was that we would get around 10,000 impressions. What happened though, is we still only got 5,000 impressions. Our campaign (unchanged) went from serving at $.20 to $.40 that day.

FB does an excellent job of putting your ads in front of the right people, but it does better with fine tuning as opposed to big changes at one time.

[Houston]

Facebook is a tricky nut to crack.

I have a few questions (and these will go off track of the topic at hand, but can possibly help you): Are you split-testing ad images and ad copies?

What happens is that once you put up an initial batch of ads, Facebook will begin sending impressions. However, it doesn't spread the impressions equally. 1 to 2 of your ads will receive all the impressions.

They don't really respond well to fucking with your campaign once you set it up and the initial run of impressions have been sent.

It seems that this is what you're experiencing.