The pricing for the click might kill off any benefit over less targeted advertising though. I poked around in keyword planner, and one click for some rolex related phrases costs $3-$10. I imagine you would have to do quite a lot of expensive experimentation to figure out how to make that profitable.
One of my favorite anecdotes from my time in comparison shopping (where north of 80% of our revenue was CPA deals) was looking at the variation of margins by product (electronics paid 1%; clothing more like 8-10% as a rev share) and seeing one outlier where we were paid a commission of 50%.
The product was a fake rolex, as in, the name of the item was literally something like "Fake Rolex", and while it was only something like $100, instead of $10,000, it was apparently still such a high margin that the seller would pay us a 50% sales commission.
The question is how many people search for Rolexes vs how many people actually end up buying one now (or at the very least, later but from the same seller, to offset the ad's cost)?
I think that for luxury goods, you can have a lot of people (more than for other expensive but otherwise boring goods) looking at them (maybe they are curious about the price) but very few that would actually follow through with their purchase.
