This week has been light on my posting about the subject directly but many of you have discovered one of the main facts on your own. Evaluation is figuring out if what you’re doing is working. It’s supposed to start with a definition of how you’ll know before you begin. It’s followed up by a periodic progress check of to see if what you’re building looks like it’s going to do what you intend. And after each implementation there’s a post mortem analysis of the project to see if it worked. This final evaluation step is geared at informing changes for the next iteration, or — more commonly — to justify the expense of what you’ve done.
Once you get beyond this simple outline, there’s just not that much to it. Oh, sure, there’s the potential for a lot of troublesome instrumentation to evaluate user response, or system capacity and the like. There are whole university courses based on the subject, but when it comes down to it, the idea is very simple.
Like a lot of other activities, it’s the implementation that’s hard.