If you have already clicked on Choice Hotels, you are more likely to click on them again than to click on another hotel chain. Based only on your Choice Hotels click, they could reasonably assume that you are more likely to click on another hotel, car rental, etc, than to click on an ad for Joe’s Crab Shack, but Choice is still the most likely than anything else and clicks are what they get paid for.
These people are PhDs and experts at data mining, with immense data sets and the hardware to mine them. They know what works.
It's probably more accurate to say they think they know what works; the problem with some companies/cultures/people is that when they get some authority/credibility they cease to consider other, differing points of view. (How many PhDs believe in anthropogenic global "climate change"?)
If you're going with a "know what works"-type mentality then you're likely going to be a lot more like Niklaus Wirth and his metric for compiler quality (self-compilation speed) embraced to the degree that he had an entire subsystem ripped out because it violated the metric. The problem though is that a lot of people get invested in the designs of what they've already done and thus consider only micro-optimization rather than re-design; but this also applies politically too, people get invested in a party and some go off the deep-end, fanatically supporting a candidate that they otherwise would consider crap [*cough romney *cough].