Jeremy Fox offers some advice about how to decide where to submit your paper. Also, read the response by Ethan White. There are possibly as many perspectives on this are there are researchers, which I'll definitely take as an excuse to present mine.
Aim as high as you reasonably can. Then submit a step higher than that. I've had really good surprizes with papers being accepted in journals I would have considered too high (speaking purely in term of impact factor). The worth you risk is being rejected (which is short, so you don't lose anything), or going into review. Which means receving comments, which will better your paper.
Don’t just go by journal prestige; consider “fit”. I don't really believe in this one. As I've mentionned before, I believe much more in papers than in journals. Being published in a good journal does not makes your paper good. Being published in a bad journal does not decrease the worth of your paper. I also think that the way we survey the literature changed a lot, with social networks and RSS; perhaps I'm wrong on this point, but I think that (at least for the younger generation) our view of the litterature focuses much more on papers than journals. In this perspective, the "fit" of the journal do not really matter. Elaborating a little bit more, initiatives like F1000Research or eLife are designed to be paper repositories. Where fit do not matter. Then, it's your job to advertize your paper, and make sure people are aware of it.
How much will it cost? I'm sure that most of the time, people are happy to pay the costs associated to a publication. And waivers are relatively easy to obtain.
How likely is the journal to send your paper out for external review? I have really no opinion about this. I used to think it matters, but I've had weird experiences this past year. I think it boils downs to "How good are you at writing cover letters?".
Is the journal open access? This is a false dilemna, in my opinion. Journals can not be open access, yet have open access mandates for some articles. Ecology Letters does, most of the Springer journals do as well. I think the question should be: "How open is the journal to new practices?". This includes ArXiV, putting data online in the open, and so forth. My feeling is that the younger people are really aware of these issues. My opinion is that open access is the tree (badly) hiding the forest, and we should do more than fighting for the access to the papers.
Does the journal evaluate papers only on technical soundness? Surprizingly, I don't care much about this. I thought PLoS One worked this way, and in my experience, it is not the case. Probably because the academic editors don't follow the rule. Also my research is wonderful and innovative, of course.
Rohr & Martin (2012) Tr Ecol Evol 27(4)
Is the journal part of a review cascade? This can be important, but again, this is not ambitious enough. Why not systematizing the process of review transfer? First, it will reduce the feeling that there are sinks and sources in journals (not saying it's not true, but there is a stigma on some journals because they are essentially perceived as Not good enough for [other journal]).
Is it a society journal? Frankly my dear, I don't give a damn.
Have you had good experiences with the journal in the past? Sure. But I've also had terrible experiences with some journals, and I'm still considering submitting something to them in the future. Ironically enough, because they fit. And I've had really good experiences with some journals, even though they keep rejecting everything I submit to them; but they do it with class? I honestly don't know.
Is there anyone on the editorial board who’d be a good person to handle your paper? I try to see which editors can be a good fit. But it's not really a decision-making or decision-breaking factor.
So in a few words, I think we should focus less on the journals. Papers will end somewhere eventually, even if it's on FigShare. But I think the exerice is interesting, and actually I'd like to see other people share their opinions. There is much to learn about how we make our decisions, and probably a lot of potential to see which areas need improving as well.