But before lecturing the rest of us, you might want to consider that not everyone's work works that way. In my own work, it is not at all unusual to go weeks without having anything that could be usefully shown to a customer; and by the same token, the customer's desires are usually simple and unchanging.
Let me give you a concrete example: importing the geometry and topology from a Parasolid XT file. Now, the good news is that the file format is publicly available. The bad news is that it more or less amounts to a direct serialization of their internal data structures, in a format where there is absolutely no room for error: a single misread byte will make the rest of the file unreadable garbage. And of course, every version of Parasolids (there are 75 or so now) used a slightly different format, and a number of crucial details were left vague or thoughtfully undocumented.
As you might suspect from that, it took well over a month to get the basic file parser done. What was I going to show the customers in that? Me: "Well, here's the dump of this file as I parse it now. As you can see, it gets completely screwed up about three-quarters of the way through." Customer: "That's great progress! But we've been talking it over, and we think your debugging dump should be colored chartreuse."
That would be the norm for my work. On shorter things (adding a simple new feature or fixing a simple bug), it usually takes less than a week, and we get as much feedback from the customer as possible. On longer things, usually the customer's only feedback is "We want to read these files" and there is little meaningful I can show them in the middle of the project.