I am preparing to give a conference presentation from scratch.
What is an effective ratio of introduction / methods / results / conclusion slides?
How can I balance the details of research without loosing the audience on key points?
In my experience, 90% of conference talks are dull and there may be one or two at any given conference that are really inspiring. What makes a "great" talk?
Some things to think about (from the point of view of a mathematician, but I hope some of this might be relevant in other fields too): Does your main result have a special case or two that would be easier for your audience to understand than the general result (without being trivial)? If so, you might just present that (or those) special case(s), and add at the end one sentence to the effect that your full result is more general.
Does your result require terminology that people might not know? If so, does it really require that terminology, or could you perhaps get by without it? If you really need the terminology, budget enough time to explain it and, if possible, relate it to something your audience already knows.
The background information (previous results, open questions) that motivate your work is likely to be too much to present in full in the limited time available. Select just enough of that information to be understandable and to provide some (not necessarily all) motivation for your work. Having presented some motivation, you can add one sentence saying that there is additional motivation, which you don't have time to explain in the talk.
In two places, I've suggested adding a single sentence saying "there's more"; you should indicate your willingness (or even eagerness) to discuss the "more" later with anyone who is interested.