Avoid the engineer's and economist's fallacy: don't reason your way to a solution -- observe real people. We have to take human behavior the way it is, not the way we would wish it to be.
The author is Donald A. Norman (he of The Design of Everyday Things), who has written extensively about and done tons of research on human-computer interaction. He knows his material the way Linus knows the Linux kernel.
It's tempting to take an observation like that and apply it to the way other creative people work -- not software or product engineers, but authors, artists, filmmakers, etc. I get the impression those folks make the mistake of wishing for human behavior as it doesn't actually exist across a couple of different realms.
The first of these is the real killer. Whether or not we realize it upfront, it's always bad news for a story when the author tries to palm off on us behavior that simply doesn't make sense by any standard, even fantastical ones. I once said we have less of a problem with a story about a man who can levitate than we would about a proud father who caves for no discernible reason and lets his children be sold off into slavery. Unless the story is about why he caves, you see, and not simply because such perverse behavior is demanded by the plot.
Being a good observer of human behavior requires something that a lot of people find it difficult to cultivate: a degree of selflessness. You need to take yourself out of the picture and let what's in front of you speak entirely for itself. That's not something that can be cultivated in a day, though -- it's something that demands constant self-appraisal, and it's far too easy to get wrong.
New York City
Other Lives Of The Mind