This is part of the risk with building IP based attractions - at the level of investment it takes to build in a theme park, you're looking at a 20-50 year intended lifespan. Rush too quickly on a property that ends up being a flash in the pan and you're still stuck with it for a good while yet. The main reason Stitch's Great Escape lasted as long as it did was that they spent a pretty penny on the AE to SGE conversion and didn't want to spend more money redoing it so quickly. Lilo & Stitch wasn't a flash in the pan movie, but the attraction was poorly recieved and far outstayed its welcome.
Now, the flip side of this is that a sufficiently excellent attraction has the ability to succeed despite its IP or lackthereof - Splash Mountain is generally the most famous (if also notorious) example of this, where most people riding have never seen the movie it's based on and that hasn't lessened its popularity one bit. And of course Pirates, Big Thunder, Haunted Mansion, Small World, Space Mountain . . . plenty of enduring attractions succeed for decades without any IP basis.
Generally speaking, the "cheat code" here is to base your attraction or anything you want or nothing at all, and just put the pedal to the metal and commit to making it an amazing attraction and chances are statistically good that it will carry its weight.