I prefer to start a step back from the nuts and bolts of what sounds to put in. It's more important to start with understanding what the needs of the scene(s) are. What's the overarching feel? What are the characters feeling? What should the viewer be feeling? What message is the piece conveying, and how can the sound support that? Those are more important questions than, "how many car doors and footsteps do I need?" because the answers to those questions should dictate what sounds you use (actual vs. commentative, tonal qualities, duration, etc.).
Then I'll go through and do my spotting. I always start out with more than I'll need, because I'd rather remove elements from a mix than have to add. I always keep those other questions and their answers in my mind at this stage, because it helps me determine what the key focus of the audio track should be; and that will be indicated in my notes.
When I'm done, i'll have two sets of notes: a set that will pertain to actual design and key elements that I think will enhance the picture's tone, and a massive list by scene of everything that I think is needed.
Of course, working on television doesn't always provide the time for detailed hand written notes, so I often have to adapt. Usually, that means my spotting notes actually end up being markers in my session timeline.