Check the book: Design is a job by Mike Monteiro and/or watch this: https://vimeo.com/121082134 both highly recommend for your case.

1 - Identify the risks at the start of the project. I always ask the client to prioritize: scope/quality, budget and time(delivery).

If budget is the least of their concern and they're focused on scope/quality I can just quote more if they want additional stuff or if I have additional ideas. If the deadline is important to them I can use that as an argument against further changes (we'll miss the deadline if you want me to do this) Something can't be cheap, high quality and delivered fast.

2 - Create a clear roadmap for the project so the client knows when the concept phase ends and the refining begins (when can they give feedback and when is the design locked down?).

3 - Never do changes for free or outside of what you've put down in your roadmap. Instead give them options such as:

It'll cost more (you will have to re-do work which was already done and accepted) It'll take longer to deliver the product (again, you will have to put in extra work) The quality will take a hit (sure, we can have a second go at the splash screen but that means no time to work on the social media integration)

Based on the priorities you've asked at the start you can make a pretty good guess on how to best control the client.

Usually clients don't understand the creative process or the order in which things are made and you sort of have to guide them through it. These late changes usually don't come from the clients themselves but friends/co-workers/aunts/cousins etc. .

You give the reasons behind the design, focusing on the business and user objectives they are meeting, not on personal preference -"I chose purple for the buttons because it fits the branding guidelines, and will give your audience a cohesive feeling throughout the site" "If I make the text green it will affect legibility and users will not be able to read the text comfortably".