Let's get honest, most scientific presentations suck mostly because people don't have the time to design them. Some do know that knowing their material is just not enough. In fact, some professors tend to expend a considerable amount of time preparing a course's material the first time. If time is the issue, the best way to solve that problem, is to present less times. It is also true that for giving a lecture a stack of visuals is not always necessary. In fact I wonder, how were the lectures our professors listened to?
The abuse of PowerPoint led to the idea that preparing a presentation was almost the same as to create the visuals stack. That's when the commonsense disaster was released. In a broad sense, a presentation that incorporates a PowerPoint/Keynote/OpenOffice/Beamer stack is a Multimedia Presentation. But instead of using the Multimedia Possibility, most of us use or have used on-canvas visuals as presenter notes, or Slideuments, as Garr Reynolds calls them. That is, slides that pretend to be documents, but that ultimately fail. The idea of Slideuments is a noble one: Kill two birds with one stone and save the presenter time. Noble in the eyes of the presenter, but a very powerful soporific for the audience. So, my advice is this: if a presenter is unwilling and/or unable to create meaning visuals that enhance his/her message, present without them, or even consider not presenting.
Allow me to be more precise on the commonsense disaster. Add to Slideuments disaster the commonsense of how the human information processing system. According to psychology professor Richard Mayer the design of some multimedia material is based on the false assumption that the human information processing system is singled-channel, has unlimited capacity and it is designed to process isolated pieces of information. In fact, the processing system looks more like this:
In scientific conferences, the most commonly attacked part of this system is the working memory. Working memory -that place in memory where information just needed to solve a problem or solve a task, such as remember a phone number, not to have to keep looking at the phone book- has limited capacity. A 90 minute PowerPoint driven lecture / presentation / pitch is most will overload the system if it is not carefully crafted.
Another commonsense disaster is the introduction given by a speaker. Common practice at the beginning of a conference is to explain the change of affiliation, involvement in a joint project or company. This is information that can be very well be given by the moderator, actually it should, for it will leverage your Ethos. The problem with this useless information is that it competes with the information / time you have to grab the audience's attention. The rule of thumb used by professional speakers is that if you don't grab the audience by the end your first 30 - 40 seconds, it will be very difficult to do that later on. To make matters worse, if you successfully grab their attention it will steady decrease over time, and after 10 minutes it is gone and you must relaunch.
In her book Better Beginnings, professional speaker Dr. Carmen Taran explains seven ideas to begin a presentation. This are
Very related to the beginning is the ending. Professional speakers say "Begin strong, end stronger". Ending in a high note is very important, even if the talk was bad. Deleting the 'thank you' slide, and instead saluting the audience at the beginning is sound advice from MIT Computer Science professor Patrick Henry Winston. Saying "Thank you for coming" at the beginning is enough. A good way to achieve a strong ending is by practicing your talk.
Consider avoiding the commonsense disaster in your next presentation!
Notes and references
 Cowan, N. (1997). The development of working memory. In N. Cowan (Ed.), The development of memory in childhood. Hove, East Sussex, UK: Psychology Press