There is also a structural argument. A provider whose revenue depends on the relationship continuing has a quiet incentive to keep you dependent, and that incentive works on people who have no intention of acting on it. Building the engagement to end removes it entirely. Everything I recommend is what I would recommend if I were never going to hear from you again, because that is the actual arrangement.
That said, ending is not disappearing, and a few things genuinely continue. Questions before a decision, which I am happy to answer and do not bill for. A second opinion on a proposal you have received from somebody else, which is often the most useful hour I can give a former client. A defined project later, if something specific comes up that sits inside what I do rather than inside the scaling work. And introductions, which is the part that matters most, because knowing who is good is worth more than any of the rest.
When the time comes to hire your next provider, the criteria shift. In year one you needed somebody who could decide things with you. In year two you need somebody who can execute at volume and measure it, which is a different profile. Ask any candidate what they would need to know before recommending a budget, whether they will work inside the accounts you already own rather than creating their own, what they measure and how it connects to revenue, and under what circumstances they would tell you to spend less. Hesitation on the last one is the signal.
The honest measure of whether year one worked is whether you can evaluate that next provider yourself. If you can read their proposal, tell whether their answers are substantive, and recognize when something does not add up, the engagement did its job. That capability is the actual deliverable, and it is more durable than anything built during the year.