My though is that if something impacts the schedule, you should track it.
We have to submit change orders (CO) with milestone dates -- the problem is approval for the CO can vary drastically (from days to months) and you never know what it will be. Therefore we include tasks like the following (All dependent on their predecessor):
- Jan-1 "Submit CO [1-Day]" (Assigned to me)
- Jan-2 "Approve CO [30-days]" (Assigned to Client)
- Feb-1 "Do real work part 1 [x-days]". (Assigned to me)
- Mar-1 "Do real work part 2 [x-days]". (Assigned to me)
- April "End User Approval [x-days]" (Assigned to client)
- May-1 Milestone: "Project Complete"
In this example, we show work that occurs before the contract is approved because we want to show real work cannot start until the client gets approval.
If the approval takes longer than 30-days, that pushes out the other deliverables. If the client misses their responsibility (Approve CO) that affects the rest of the agreed on schedule and requires a change management process address the schedule change. The change management process addresses changes in scope/money/schedule and is a negotiation between you and the client.
The change management process is an important step because you may have agreed to finishing up by May-1 because in June you have another project starting or maybe your developer is going on vacation. If approval ends up taking an extra 15-days (45D vs 30D), it could push your final deliverable out 60-days (Because the developer is taking time off in May). By adding the "Approval" step, you make it clear to the client that the date is important.
NOTE: You can also cover this under the risks/assumptions section of a project plan instead of the project schedule.
SIDE NOTE for complex CO/SOW: Most times, creating a CO/SOW only requires a meeting or two and we swallow that cost. But, if preparation for the CO/SOW is long, you may need to treat creating the CO/SOW as a separate project with a deliverable of the CO/SOW for the second project (actually doing the work). Sometimes creating the CO/SOW requires a bunch of work from our BA/Architect; that is when we do a CO/SOW to create the final CH/SOW. Due to budgets, we have done a $50k project to create a CO/SOW and then waited a year before executing that CO/SOW.