The organization of a talk pretty much follows the organization of a
paper, but it needs to be even more focused. You can only get across one simple
point in 10 minutes. The talk will proceed as follows: I will introduce you and
your title. Before you put up your first viewgraph, you should give a brief
introduction to what you are about to say. The first seven seconds of your talk
are the most important, so think carefully about this. Try to catch people's
interest--why should they listen to you? After your introduction, put up a slide
with your title and affiliation. Then put up a slide with an outline. After
that, you should discuss your experimental methods (show a diagram of the
apparatus), results, analysis and conclusion. After the outline, the maximum
number of slides you should have is absolutely no more than 7, and "less is
more."
Each slide should do a job and be self contained. At the top of each
slide, put a title explaining what the slide is about (for example, "Apparatus").
Slides should emphasize graphics --sketches of equipment, plots of data, or
sketches of physical models for what you are trying to talk about--rather than
text or equations. The whole talk should avoid all but simple equations. Assume
that, if someone has a detailed question on an equation, they will ask you after
the talk.