The first guy who trained me as a software consultant said that estimates are generally laughably incorrect, because you just never know what you will run into.

The first guy who trained me as a software consultant said that estimates are generally laughably incorrect, because you just never know what you will run into. Software (and technology in general) is about hiding complexity. As soon as you start to dig into complexity, you have no idea what you're going to get. But that's not really a pleasant concept to a businessperson who is trying (perfectly reasonably) to predict cleanly and clearly what something will cost so that they can gauge their ROI. So what happens is that everybody bids on a job, some people will massivley lowball their estimate to win work, and then a year or two later another company will get called in to clean up whatever mess the lowballer made. So it goes.