In any control system that is functioning properly, the methods used to control a signal won’t be correlated with the signal they’re controlling.
Worse, there will be several variables that DO show relationships, and may give the wrong impression. You’re looking at variables A, B, C, and D. You see that when A goes up, so does B. When A goes down, C goes up. D never changes and isn’t related to anything else — must not be important, certainly not related to the rest of the system. But of course, A is the angle of the road, B is the gas pedal, C is the brake pedal, and D is the speed of the car.
In the past, GDP and resources use have always been tightly correlated. But this is just drawing a line through some data — it’s not based on any deep theory. And in fact, these correlations can change very quickly. Just as one example, here’s energy use versus GDP since 1949.
If you were sitting in 1970, you could look at this curve and claim, very confidently, that economic growth requires concomitant increases in energy use. And you’d be wrong. Because the trend is your friend til the bend at the end.
Perceptual: "They couldn't figure out what to do next", "they couldn't find the feature", "they didn't know they could click that button..." etc.
Domain-specific: "We need a way to jump back here because in their workflow this happens..."
In general, usability testing only catches type 1 perceptual problems. Because in those tests you take people out of the real world and assign them tasks. Usability testing doesn't catch domain-specific problems because they only come up in real life use.