"Hindsight is 20/20" is a cliché because, like most clichés, it's true. It's easy to look back weeks, months, or years later at a given action and see what you should have done differently.

When I work with my clients on any project, but especially AMS implementation projects, I tell them at the outset, "We're going to make the best choices given the information we have at the time. This likely means we will make decisions that we ultimately have to change later on."

Like so many other things related to projects, this is part of managing expectations. Far too often I will encounter groups who have implemented an AMS in the recent past, and they lament all the incorrect decisions they made. Were some of them avoidable? Probably. But the reality is that many of the decisions were made with the best information at the time. What they found out later was not information they had when the decision was made, so there's no way they could have known to make a different decision.

Project post-mortems are fine and can be very helpful learning experiences. But the key is to learn from them, not punish yourself for "mistakes" that were made.

Don't play "Woulda, shoulda, coulda." Take what you know now and adjust as needed.