They invest the time and the learning to master a workflow. They expect a payoff from this investment in their ability to use these workflows to achieve other ends. When you mess with a workflow, you negate that investment. They have to spend time learning and mastering a workflow all over again before they can apply it toward their actual goals.
Nobody uses software "to be using software" or "for a good experience." They use it to get things done. If they have to spend two weeks mastering a new workflow then your improvements had better deliver a multiple of that value in return, or they're going to come back with "that's cool, but it would trip me up for all of my muscle and click memory to be invalidated."
People aren't averse to improvements. They're averse to evolutionary improvements that cost more to the user in practice (time invested and mistakes avoided) than they deliver on the other end. "Small tweaks" often fall into this category. Some dev moves a button to a more "logical" placement and for the next two weeks, the users lose five or ten seconds every single time they need to use it because their absent minded clicking—absent-minded because they're focusing on what they're really trying to accomplish, not on 'using the software'—keeps ending up in the wrong place vs. what they're accustomed to.
Dev says "BUT IT'S BETTER." User experience is actually that of being irritated and not getting things done as efficiently as usual, so their response is "IN PRACTICE, IN THE CURRENT CONTEXT OF MY LIFE, NO IT'S NOT."