If you were to start out with one step in a job, then add one step at a time, each time the job would run for longer.
I guess you're not meaning that.
Is one step going to be affected by the fact that there are seven others in the job? No.
If all the data for all the steps were on the same volume, maybe letting them rip all at once in seperate jobs would introduce some delay, with read-heads flying about everywhere. Maybe.
You may want to split your job for scheduling purposes. Some other job may be able to kick-off when step 2 finishes, but has to wait until the end of the job, for instance.
OK, what is the real question?