I think a big problem with their logic which I don’t see discussed very much is the way they determine if someone “wanted” something. With metrics-driven marketing (or metrics-driven anything really, but the current topic is marketing) there is an assumption that actions are only taken because that action is wanted. There are plenty of alcoholics who don’t want to pick up another bottle, but they do anyway. The metrics-driven person would see the action of picking up the bottle and say “that person wanted to pick up that bottle”. Not everything is chemically addictive like alcohol, but the principle remains the same: simply selecting material that causes a certain response within a significant percentage of the population is not the same thing as selecting material that people want to see or that has a positive impact on our lives.
This is why I appreciated this section:
This is an example of looking at a metric, then also thinking about why that metric exists and making a change that is helpful to the intended purpose of the people driving that metric. It’s not “this ad made sales go up”, it’s “this ad caused people to go to this page, here is a plausible explanation for the need driving that action, so let’s find a way to better fill that need”. In this case the result still ends up being “make it easier to take that action”, but it has some degree of nuanced thought behind it.