In an online environment, procrastination can happen more frequently, just for the sheer fact that students are not expected to show up at class at a certain time and they don't necessarily feel the pressure of knowing their instructor and fellow students face to face

Asynchronus discussions are best used when involving all or multiple students. Synchronus discussions should be conducted with one student. Both are largely driven and motivated by the instructor.

Students need to have a sense that their instructor is engaged and accessible. 

Familiarization with my institution's CMS will help my students feel more comfortable, as well as myself.

Adherence to institutional policy is key. Keeping lines of communication between yourself and the students is crucial. Different types of students have different ways of learning and interacting, some of which can be disruptful if not handled appropriately.

Asynchronous discussions are more effective in conveying important and essential information. Synchronous discussions can deter participation by learners who are slow readers and/or typers. I'm sure I'd use asynchronous discussion boards at least twice weekly with a required number of characters and points of fact. I'd also require they reply to at least two other students' posts constructively

Asynchronous discussions are best for more complex and important course material/concepts. Synchronous discussions discourage participation from shy learners, slow readers and typers, etc. You're liklier to experience greater participation with asynchronous discussions. I'd probably utilize discussion boards with a required number of characters and several points of fact to be covered a couple to a few times weekly.

The importance of personalization and relating to students as individuals, not "just" as a class.

The importance of technical support and guidance in addition to teaching responsibilities. If students can sense that you know and are comfortable with the platform, they'll be more comfortable with distance learning.